Gloria Felder, Individually and As Personal Representativeof the Estate of Harry Felder, Jr., Deceased, Etal., Plaintiffs-appellees, v. the United States of America, Defendant-appellant.gloria Felder, Etc., et al., Betty Ann Henschen, Etc., Etal., and Lynda M. Burns, Etc., et al.,plaintiffs-cross-appellants, v. the United States of America, Defendant-cross-appellee, 543 F.2d 657 (9th Cir. 1976)

Eloise Davies, Asst. Atty. Gen. (argued), App. Section, Civil Div., Dept. of Justice, Washington, D.C., for appellant in No. 75-1267 and appellee in No. 75-1455.

Charles M. Brewer (argued), of Lewis & Roca, Phoenix, Ariz., for appellant in No. 74-1455 and appellees in No. 75-1267.

Before BARNES and ELY, Circuit Judges, and PALMIERI,^*  District Judge.

PALMIERI, District Judge.

The Government has appealed from judgments for over two and a quarter million dollars awarded under the Federal Tort Claims Act (28 U.S.C. §§ 1346(b) and 2674) as a result of three wrongful death claims in the fatal crash of a Piper Comanche airplane at Tucson International Airport (the airport) on May 12, 1969. The pilot and two passengers in the plane were killed, and the plane (for which $9,000.00 was awarded) totally destroyed. These appeals present issues both with respect to liability and damages. Plaintiffs have cross-appealed, contending that the district judge erred as a matter of law in subtracting from their awards estimated income taxes that would have been paid on lost income.

The trial court was amply supported by the record in concluding that the actions of the control tower personnel were the proximate cause of the crash, the consequent destruction of the airplane, and the death of the three occupants. Viewing the evidence in the light most favorable to the plaintiffs, the facts which follow appear to have been sufficiently established. The accident took place in daylight and in clear weather, with a visibility of 50 miles, and with visual flight rules in effect. The control tower was utilized for the control of air traffic at the airport and was being operated by employees of the Federal Aeronautics Administration (FAA). The primary concern of this traffic control was to insure the orderly movement of arriving and departing aircraft in the zone of control. One of the primary duties of the air traffic controllers was to advise pilots of any unsafe conditions that might affect the operation of their aircraft.

On May 12, 1969, Lufthansa German Airlines was engaged in performing "touch and go" landings for training purposes at the airport, utilizing a Boeing 707 aircraft. These landings involved a touch down by the aircraft followed by a take-off before the aircraft came to a complete stop. This type of training activity had been conducted at the airport on numerous prior occasions and for a number of years. The tower personnel at the airport were aware that touch and go landings by a plane of the size, speed, and weight of a Boeing 707 would expose a following light plane to wake turbulence caused by wing tip vorti- ces. It is common knowledge that the passage of the wing of a heavy aircraft like a Boeing 707 through the air causes a rapid swirl of air at the tip of each wing in substantially the same manner as a tornado, except that its position in the air is horizontal. Such air turbulence when encountered by a light aircraft of the Piper Comanche type could subject it to forces exceeding its control capabilities.

Harry Felder, the deceased pilot, was competent and experienced. He was inbound from Nogales, Mexico, landed at the airport and taxied to the Customs area for clearance. He intended to resume his flight to Phoenix, Arizona, after clearing Customs. Captain Emil Kuhnl of the Boeing 707 was an experienced pilot and bears no responsibility for the accident which was shortly to occur. The Boeing 707 had just taken off on its touch and go landing maneuver when the Piper Comanche was cleared for take-off. Felder was not given any cautionary warning concerning the take-off of the Boeing 707 and it was reasonable to infer from the evidence that he did not know that this had occurred. The Customs area was located behind the southeast corner of the Tucson terminal, and by the time the Piper Comanche had reached the point at which it was to initiate its take-off, the Boeing 707 was already airborne and sufficiently advanced in its flight pattern so that it was not readily visible to the pilot of the small plane. When Felder requested and received clearance for his departure it is clear that the FAA controller knew of the touch and go maneuver of the large plane and was chargeable with knowledge of the imminent exposure of the smaller plane to the danger of wake turbulence. He did not, however, issue any cautionary warning nor did he advise Felder of the take-off of the Boeing 707 which had occurred shortly before. The controller could have delayed the departure of the Pipe Comanche for a sufficient length of time so as to avoid the risk of any encounter with the wing tip vortices of the larger plane, but he did not do so. There is persuasive evidence that when the Piper Comanche reached a point approximately 150 to 250 feet above the runway, pilot Felder lost control of his Piper Comanche because it encountered wake turbulence created by the recently departed Boeing 707. Felder's plane flipped on its back abruptly, spiralled nose-down, and crashed without any possibility of pilot control. The district court was justified in rejecting the contention of the Government that the Piper Comanche had crashed because of a stall. The encounter with wake turbulence was the proximate cause of the crash, the ensuing deaths of the pilot, Harry Felder, and the two passengers, Wilbur G. Henschen and Peter P. Burns, and the destruction of the aircraft.

The Government also challenges on this appeal the damages awarded by the district court. It objects to numerous aspects of the trial judge's calculations and attacks the total amount as so excessive as to be punitive in contravention of the applicable statute. The plaintiffs have cross-appealed arguing that the trial judge committed reversible error by deducting estimated income taxes each decedent would have paid on his future earnings. Wishing further illumination of the damages issues, we requested supplemental briefs on all aspects of damages.

We approach the issue of damages raised in this case cautiously because it was substantially uncontested at the trial level. The Government offered no affirmative evidence of its own by way of expert testimony or otherwise. The Government made no motion to amend the findings of the district court or the judgment entered by that court, nor did it move for a new trial or otherwise raise the issue of damages before that court after its decision.

We are faced at the outset with the proper standard of review of a trial court's damage award under the Federal Tort Claims Act (FTCA, or the Act). The district judge gave careful and conscientious attention to the evidence before him and deference must be given to his judgment. On questions of law, such as the proper elements of an award of damages, we, of course, have a free hand in scrutinizing the district judge's decisions. United States v. Mississippi Valley Generating Co., 364 U.S. 520, 526, 81 S. Ct. 294, 5 L. Ed. 2d 268 (1961). But there is some confusion in the briefs as to the standard of review we should apply to the district court's findings of fact on the issue of damages.

After a non-jury federal trial the standard of federal appellate review of a finding of fact such as the calculation of damages is the clearly erroneous standard of Rule 52(a). Fed. R. Civ. P. Rule 52(a) states in relevant part:

Findings by the Court

(a) Effect. * * * Findings of fact shall not be set aside unless clearly erroneous, and due regard shall be given to the opportunity of the trial court to judge of the credibility of the witnesses.

The Supreme Court has supplied us with a definition of this standard of review:

A finding is "clearly erroneous" when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.

United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S. Ct. 525, 542, 92 L. Ed. 746 (1948). Accord, Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 123, 89 S. Ct. 1562, 23 L. Ed. 2d 129 (1969); Guzman v. Ruiz Pichirilo, 369 U.S. 698, 702, 82 S. Ct. 1095, 8 L. Ed. 2d 205 (1962).

The case we review today, however, is complicated by the fact that it is not concerned exclusively with federal law. Although it involves a non-jury case tried in a federal district court, the measure of damages under the Federal Tort Claims Act is based on state law. 28 U.S.C. § 2674. The question thus arises whether we should apply the state or the federal standard in reviewing the district court's calculation of damages.

The plaintiffs urge us to adopt the Arizona test as stated in Young Candy & Tobacco Co. v. Montoya, 91 Ariz. 363, 370, 372 P.2d 703, 707 (1962):

The damages, therefore, must be so excessive as to strike mankind, at first blush, as being beyond all measure, unreasonable, and outrageous, and such as manifestly show the jury to have been actuated by passion, partiality, prejudice, or corruption. In short, the damages must be flagrantly outrageous and extravagant . . . .

The Government argues that the damages are excessive by any standard.¹ 

In three cases, we have applied the clearly erroneous standard in reviewing damages set by a trial judge under the FTCA. Mills v. Tucker, 499 F.2d 866, 868 (9th Cir. 1974); McCauley v. United States, 470 F.2d 137, 139 (9th Cir. 1972); Layne v. United States, 460 F.2d 409, 412 (9th Cir. 1972). See United States v. Furumizo, 381 F.2d 965, 970 (9th Cir. 1967). In two other cases we applied the state standard of review in the same situation. Early v. United States, 474 F.2d 756, 759 (9th Cir. 1973) (Alaska standard); United States v. Becker, 378 F.2d 319 (9th Cir. 1967) (Arizona standard). The question of which standard was proper was not fully considered in any of these cases.

We believe that the federal standard should apply in the review of a finding of fact made in a non-jury trial by a federal court applying state law.²  In general, the standard of appellate review is a procedural matter in which the forum will utilize its own standards even when it is applying the substantive law of another jurisdiction.³  The question we face on reviewing the quantum of damages in a case like this is whether or not the district court correctly applied the substantive state law. The substantive state law includes the state statute as well as the state court's interpretation of that statute. The calculation of damages, on the other hand, is a question of fact, and review of that question of fact is not an interpretation of the statute, and not a part of the substantive state law. Rather, the standard of appellate review of a finding of fact more closely resembles a matter of procedure and entitles this Court to apply its own standard. In this context the state standard is not binding upon a federal court.⁴ 

The only Arizona standard cited to us is particularly inappropriate for use in this case. This Arizona standard on its face applies to the review of a jury verdict awarding damages. Since appellate review of an award by the court sitting without a jury is not limited by considerations, federal or state, of deference to the province of the jury,⁵  it can be less exacting and still give proper deference to the trier of fact. Cases concerning review of jury verdicts and abuse of the trial court's discretion in adjudicating motions for new trials following jury verdicts do not set forth the standard for review of a judgment rendered by a court sitting without a jury.⁶  The deference we show to a trial judge's first hand experience with the evidence and the due regard to be given his opportunity to judge the credibility of the witnesses may be referred to as discretion, but it should not be confused with his discretion in dealing with motions for new trials following jury verdicts.

In arriving at a measure of damages the district judge deducted some federal and Arizona income taxes that he estimated the decedents would have paid on their projected incomes. The plaintiffs cite this as reversible error. They argue both that Arizona law must be applied and does not permit such a deduction and that, even if Arizona law is not followed, the better rule of law is that taxes should not be considered in calculating damages.

This is an issue on which there has been much discussion and little agreement. We recently had occasion to clarify our views on the matter in Burlington Northern, Inc. v. Boxberger, 529 F.2d 284 (9th Cir. 1975). Upon a thorough analysis and synthesis of the cases and commentaries, we found that most of the commentators who had carefully analyzed the issues favored consideration of the effect of income taxes. We firmly supported this rule, concluding

that, as a matter of fairness and logic, the just approach would require a rule providing for the admissibility of evidence of, and corresponding deduction to account for, future income taxes in all cases. (Emphasis in original)

529 F.2d at 294.

On the facts of that case, and after due consideration of the leading case, McWeeney v. New York, N.H. & H. R.R. Co., 282 F.2d 34 (2d Cir.) (en banc), cert. denied, 364 U.S. 870, 81 S. Ct. 115, 5 L. Ed. 2d 93 (1960), we held that

in cases wherein the gross earnings in question are beyond "the lower or middle reach of the income scale," and consequently "the impact of income tax has a significant and substantial effect in the computation of probable future contributions" to the beneficiaries, both parties should be permitted to introduce evidence of the extent to which future earnings would have been taxed.

529 F.2d at 295.

We here reaffirm the views we expressed in Boxberger. However, Boxberger was a Federal Employers' Liability Act case in which federal law controlled the issue of damages. The Federal Tort Claims Act requires a somewhat different analysis. We note in passing that most FTCA cases are in one sense more appropriate cases for the deduction of taxes than FELA cases. Since there is no jury in FTCA cases (28 U.S.C. § 2402), one of the main reasons for disfavoring deduction of taxes is avoided, namely, that calculation of the tax would be too confusing for the jury. See Burlington Northern, Inc. v. Boxberger, supra, 529 F.2d at 289 and 293.

The FTCA gives a prominent role to state law. The liability of the federal government for the conduct of its employees is the same as that " under circumstances where . . . a private person . . . would be liable to the claimant in accordance with the law of the place where the act or omission occurred." 28 U.S.C. § 1346(b). The measure of damages that can be recovered against the United States is also based on state law. 28 U.S.C. § 2674 provides:

The United States shall be liable, respecting the provisions of this title relating to tort claims, in the same manner and to the same extent as a private individual under like circumstances, but shall not be liable for interest prior to judgment or for punitive damages.

If, however, in any case wherein death was caused, the law of the place where the act or omission complained of occurred provides, or has been construed to provide, for damages only punitive in nature, the United States shall be liable for actual or compensatory damages, measured by the pecuniary injuries resulting from such death to the persons respectively, for whose benefit the action was brought, in lieu thereof. (Emphasis supplied).⁷ 

The law of Arizona on the incidence of income taxes in tort damage awards is not as clear as it might be. The only Arizona case cited to us or that we have been able to find which discusses the subject is Mitchell v. Emblade, 80 Ariz. 398, 405, 298 P.2d 1034 (1956). In that case, Arizona's highest court reversed a trial court on a matter not here relevant and then went on to "pass upon the other claimed errors for the guidance of the trial court." 298 P.2d at 1036. One such claimed error was the denial of a requested jury instruction:

The defendants requested an instruction that damages for medical expenses and pain and suffering were not subject to income taxes and that, in the event the jury found for the plaintiff, it could take into consideration this fact in fixing the amount of plaintiff's damages.

298 P.2d at 1037.

The court reviewed a number of decisions in other states and aligned itself with those that concluded that such an instruction would be based on an impermissible assumption that the jury would not confine itself to the evidence and to the court's charge on the elements of damage to be considered but would consider matters not mentioned in the charge:

. . . we are of the view that Illinois, Texas and Ohio are correct in requiring that the case be tried on the issues and presented to the jury with a correct measure of damages, of which the incident of income tax has no part. We would prefer to assume that the jurors to the best of their ability will follow the instructions given and will not depart from the issues and the law as announced.⁸ 

298 P.2d at 1038. The court also expressed its concern that the charge might become too complex for the jury.

The plaintiffs cite Mitchell v. Emblade, supra, for the proposition that, under Arizona law, "the incident of income tax has no part" at all in determining the measure of damages in any tort action. The Government attempts to distinguish Mitchell by pointing out that it deals only with the fact that damage awards in personal injury and wrongful death cases are not taxable, while this case deals with the question of whether or not to deduct projected income taxes where lost future income is an element of damages an analytically distinct issue. Boxberger, supra, 529 F.2d at 296; United States v. English, 521 F.2d 63, 72 n. 8 (9th Cir. 1975). The Government might also have suggested that Arizona would perhaps not follow its 20 year old opinion today because of recent trends in the law. And it might have sought to bolster this argument by noting that the relevant Arizona statute has changed since Mitchell.⁹  The Government might also have argued that Arizona would allow a court, as opposed to a jury, to consider taxes in arriving at a just measure of damages. Fortunately, we are spared on the facts of this case the difficult and disconcerting task of second guessing the Arizona court.

The general purpose of damage awards in tort actions has been to compensate plaintiffs for losses incurred.

The primary aim in measuring damages has been compensation, and this contemplates that the damages for a tort should place the injured person as nearly as possible in the condition he would have occupied had the wrong not occurred. . . .

C. McCormick, Law of Damages § 137 at 560 (1935). Under such a measure of damages, a survivor could not expect to receive that part of a decedent's lost income that would have been paid in taxes because neither the survivor nor the decedent would have had use of that money if the decedent had lived. But the language of the damages section of the Act is somewhat less clear. The Act contains a built-in tension between the section 2674 language making the United States liable "in the same manner and to the same extent as a private individual under like circumstances" and its numerous safeguards to limit the liability of the Government,¹⁰  and impliedly, to prevent any undue largesse in fixing awards under the statute. One of these safeguards is nonliability for punitive damages. Was this meant to limit the Government's liability to compensation for actual losses? The long and complex legislative history of the Act does not speak directly to this issue.¹¹  We have been unable to find therein a definition of "punitive damages" or an unequivocal statement that the Government was making itself liable only for losses actually suffered by plaintiffs.

However, there are passages that suggest such an intent. In H.R.Rep.No.1112, 77th Cong., 1st Sess. 2 (1941), the Committee on Claims, reporting on a predecessor of the bill that was eventually enacted, noted that a government agency had been working "so that private citizens who have suffered injury without fault of their own may be reimbursed promptly" (emphasis added). The bill was recommended for passage because "(y)our committee feels that it is the essence of justice and good administration that such injuries . . . should be promptly recompensed " (emphasis added). Id. A spokesman for the Attorney General stated that the purpose of such a bill was "to redress tortious wrongs arising out of Government activity," and he clarified that by saying "we think it is enough to satisfy the actual claim rather than impose punitive damages on the United States."¹² 

The second paragraph of 28 U.S.C. § 2674, although not applicable to this case (see footnote 7 above), gives an indication of the damages Congress had in mind. This language was added by a 1947 amendment to remedy the fact that the Government was maintaining that it was liable for no damages under the Act under the laws of Alabama and Massachusetts which awarded only punitive damages for wrongful death. In explaining the approach of the bill, the Committee Reports, as well as a spokesman for the Comptroller General, implied that Congress, in passing the Act originally, had intended to make itself liable only for compensatory damages:

Since in those two states compensatory damages are not allowed, all that is required is to amend the Federal Tort Claims Act to say that in such states compensatory damages shall be allowed. . . . It is believed that that suggestion would eliminate the discrepancy and would make the settlement of claims in those two states to be exactly in accord with the general rules followed in the other 46 states. . . .¹³ 

This passage has obvious limitations as an authoritative statement. In effect, we have Congress telling us in 1947 what it meant in 1946 when it passed the Act. Furthermore, the bill as enacted into the second paragraph of section 2674 restricted compensation to pecuniary losses, a much more narrow measure of damages than that allowed by the first paragraph. Despite these limitations, we accept this passage as some evidence of congressional intent to expose the Government to liability only for compensatory damages.

The Supreme Court has stated of the Act that

(t)he broad and just purpose which the statute was designed to effect was to compensate the victims of negligence in the conduct of governmental activities in circumstances like unto those in which a private person would be liable and not to leave just treatment to the caprice and legislative burden of individual private laws (emphasis supplied).¹⁴ 

In United States v. English, 521 F.2d 63 (9th Cir. 1975), we formulated the rule that damages under the Act are to be limited to compensation for actual losses:

If the local law provides for punitive damages, or permits application of standards which result in plaintiffs getting more than compensatory damages, only compensatory damages may be awarded.

521 F.2d at 70. Under this rule we held it was error not to discount projected earnings to present value.¹⁵  521 F.2d at 72, citing Chesapeake & Ohio R. Co. v. Kelly, 241 U.S. 485, 491, 36 S. Ct. 630, 632, 60 L. Ed. 1117 (1916) (FELA) ("the principle of limiting the recovery to compensation requires that adequate allowance be made . . . for the earning power of money").

The First Circuit reached a similar result in D'Ambra v. United States, 481 F.2d 14, cert. denied, 414 U.S. 1075, 94 S. Ct. 592, 38 L. Ed. 2d 482 (1973). There, the plaintiffs were parents of a four year old decedent. The Rhode Island statute measured damages on the projected accumulated value of decedent's estate without regard to whether or not the plaintiffs would have had any expectation of receiving any of the benefits of that estate in the normal course of events. The court posed the question whether Congress intended to do more than compensate survivors when the state statute required or permitted more. Their answer was that Congress refused to allow anything more than compensation by its specific exclusion of punitive damages:

Punitive damages, whether viewed as "smart money" or as a deterrent, is that part of the award that is not compensatory; the terms are mutually exclusive.

481 F.2d at 17.

Similarly, in Hartz v. United States, 415 F.2d 259 (1969), the Fifth Circuit held that the Georgia statute was punitive to the extent that it did not allow deduction for decedent's living or other personal expenses and thus permitted recovery of more than the plaintiff's loss.

It appears settled, then, that the purpose of the FTCA is compensation, that is, it is intended to repay the amount of loss or injury sustained by a plaintiff as a proximate result of governmental misconduct which gave rise to the cause of action. Although the Act does not define punitive damages, they may be thought of generally as damages intended to punish and deter. However, we are not bound to a narrow definition. Since the interpretation and application of the Act is a matter of federal law,¹⁶  we look to the purpose of the Act for a definition of punitive. Likewise, in deciding if a state statute is punitive, we look not to its language nor to the state court's characterization of it. Rather, we look to its effect. D'Ambra, supra, 481 F.2d at 17-18.

The decedents' incomes were projected to range from $30,000 to $125,000 per year. The amount of income taxes, federal and Arizona, that decedents would have paid on these incomes would undoubtedly have been substantial. To award the full incomes to the survivors without deducting these taxes would be to award them monetary compensation that they could not logically or reasonably have expected to receive had decedents lived. United States v. Furumizo, 381 F.2d 965, 971 (9th Cir. 1967) ("It is reasonably certain that what would have been available would have been after-tax dollars, not pre-tax dollars.") United States v. English, 521 F.2d 63, 72, n. 8 (9th Cir. 1975); Cox v. Northwest Airlines, Inc., 379 F.2d 893, 896 (7th Cir. 1967), cert. denied, 389 U.S. 1044, 88 S. Ct. 788, 19 L. Ed. 2d 836 (1968) (FELA).

Since failure to deduct income taxes would result in plaintiffs receiving greater financial support than they would have in the normal course of events, we would consider the effect of such an award to be punitive.¹⁷ 

We hold that, in dealing with incomes as large as those involved here ($30,000 to $125,000 per year), failure to deduct income taxes in computing lost earnings would result in an award of punitive damages that is impermissible under the Federal Tort Claims Act.

Our decision today appears to be the first to apply this interpretation of punitive damages under the Act to the issue of income taxes. However, the possibility that failure to consider the incidence of income tax on large incomes could produce unacceptable results has been clearly recognized even by courts that generally disfavor or disallow consideration of the tax effect. McWeeney, supra, 282 F.2d at 38 ("There may be cases where failure to make some adjustment (for) income taxes would produce an improper result . . . ."); Johnson v. Penrod Drilling Co., 510 F.2d 234, 237 (5th Cir., 1975) (en banc) (Tax question not to be presented to the jury "in a case such as those (sic) at bar in which the annual wage rate . . . is subject to an income tax rate which would not distort projections or result in a windfall recovery . . . ."). See also United States v. Sommers, 351 F.2d 354, 360 (10th Cir. 1965) (Tax question left to the trial judge, but "(u)ndoubtedly situations may arise in which the failure to take into account income tax liability would produce an unconscionable result . . . .").¹⁸ 

We do not consider our holdings in United States v. Becker, 378 F.2d 319 (1967), and McCauley v. United States, 470 F.2d 137 (1972) to be wholly inconsistent with this opinion. In both of those cases we affirmed FTCA awards based on Arizona law although the trial court had not deducted projected income taxes. However, in McCauley, the punitive damages issue apparently was not raised. Moreover, the court stated that no compelling reasons had been cited for adopting a fixed rule that would restrict the trial court's evaluation of what damages would be reasonable. In Becker the lost income was $15,000 per year much lower than the incomes which concern us here. The review of cases in Boxberger, supra, 529 F.2d at 290, indicates that this $15,000 income would fall near the top of the McWeeney "lower or middle reach of the income scale," 282 F.2d at 39, where income taxes are not taken into account. Furthermore, in both of these cases we avoided establishing a hard and fast rule that income taxes could not be deducted. In McCauley we said:

We assume that in the case at bar the experienced trial judge was able to evaluate, as required by Arizona law, the whole of the evidence and reach a reasonable decision.

470 F.2d at 139. In Becker we stressed that the fixing of "fair and just" damages was primarily for the trier of fact under Arizona law. 378 F.2d at 324. In both cases the facts and circumstances made the issue of income tax a close enough question that we felt it proper to defer to the discretion of the trial court. The situation we face today is far different and compels a different result.

We now turn to the specific awards made by the district court. We have concluded that it is necessary to reduce some of these awards and to increase one of them. However, we have calculated these changes ourselves rather than remand for recalculation. In the circumstances of this case, we think we not only may but should take this procedural route. The authority for doing so is 28 U.S.C. § 2106 which reads:

The Supreme Court or any other court of appellate jurisdiction may . . . modify . . . any judgment, decree, or order of a court lawfully brought before it for review, and may remand the cause and direct the entry of such appropriate judgment . . . as may be just under the circumstances.

This broad grant of power is, of course, delimited by the "clearly erroneous" standard of Rule 52(a), Fed. R. Civ. P., discussed above. Thus we must and have given due regard to the opportunity of the trial court to judge the credibility of the witnesses, and have changed the awards only to the extent necessary to bring them within an acceptable range. Chris-Craft Industries, Inc. v. Piper Aircraft Corp., 516 F.2d 172, 187 (2d Cir. 1975); cert. granted, 425 U.S. 910, 96 S. Ct. 1505, 47 L. Ed. 2d 760 (1976); Petition of United States Steel Corp., 479 F.2d 489, 501 (6th Cir.), cert. denied, 414 U.S. 859, 94 S. Ct. 71, 38 L. Ed. 2d 110 (1973); Mitchell v. Evelyn C. Brown, Inc., 310 F.2d 420, 426 (1st Cir. 1962). Wherever possible we have utilized findings by the trial court in recalculating the awards. Most of the changes we have made involved arithmetical calculations that we could perform as easily as the trial court and a remand would necessarily have involved a waste of judicial resources. Dale Benz, Inc. v. American Casualty Co., 303 F.2d 80, 82 (9th Cir. 1962); Fuchstadt v. United States, 442 F.2d 400 (2d Cir. 1971); Simpson v. United States, 322 F.2d 688, 693 (5th Cir. 1963).

Moreover, consideration of the passage of time since the commencement of this litigation compels us to seek an early disposition of the matter. Due to the complexity of the suit, it is still in the courts seven years after the occurrence of the underlying accident. A remand could easily lead to a further appeal. In order for damages to fulfill their purpose of compensating survivors, they must be received as soon as possible after the loss. The interests of justice and the best interest of the parties require that we recalculate the damages on the basis of the record before us and order the entry of a modified judgment. Chris-Craft Industries, supra, 516 F.2d at 186-87; Georgia-Pacific Corp. v. U. S. Plywood-Champion Papers, Inc., 446 F.2d 295, 299 (2d Cir.), cert. denied, 404 U.S. 870, 92 S. Ct. 105, 30 L. Ed. 2d 114 (1971).

The first element of damages that we turn to is the Felder pension plan. The trial judge found that an amount equal to 8 percent of Felder's annual earnings (2 percent taken from those earnings; 6 percent contributed by Felder's employer) would be placed in the pension fund in Felder's name. He found that these funds would grow at 7.5 percent per annum, the fund's earnings rate, until 1986 when they would be transferred to a payout fund (earning 5 percent per annum according to the record), and a ten year payout to Felder would begin. The court found that the sum of these payments would be $365,000 and added this full amount to the recovery by Felder's survivors. The Government argues that it was error not to reduce this amount by deducting appropriate income taxes and personal expenses, and by reducing it to present value. We agree. United States v. English, 521 F.2d 63, 71-72 (9th Cir. 1975). The appellees do not dispute the fact that personal expenses and the time value of money should have been taken into consideration. Consequently, we have recalculated this element of recovery. Using the figures cited above, we find that the yearly payment to Felder from the pension fund would have been $34,025. We then calculated the tax on this amount assuming this was the gross income of a married couple filing jointly, claiming three exemptions,¹⁹  utilizing the standard deduction and 1974 tax rates. After allowing just over $13,000 for Felder's personal expenses,²⁰  we find that $13,000 per year would have been available to Felder's survivors. The value of this stream of income on January 1, 1975 (shortly after the November 1974 judgment entered in this case) is $43,380. This is the present value of the pension fund to Felder's survivors.

The Government argues that the amounts deducted as taxes from the projected earnings of the decedents are so low as to be erroneous. It urges that the taxes be calculated by direct application of the Internal Revenue Code. The plaintiffs stress the evidence in the record of the use of tax shelters, particularly by persons in income brackets as high as those projected for Felder and Burns. We decline to ordain any specific procedure for the trial court to follow in calculating taxes, particularly when there is reason to believe that decedents would have been sophisticated or wealthy enough to attempt minimization of taxes. In this case, however, we believe the incidence of taxes has been given insufficient weight.

Felders' income was projected to range from $75,000 in 1970 to $125,000 per year from 1977 to 1985. Henschen's income was projected at $60,000 per year from 1970 through 1978 (when he reached age 60) and thereafter at $30,000 per year through 1983. Burns' income was projected to range from approximately $30,000 in 1970 up to $111,000 in 1997.

To obtain a guidepost as to how much federal income tax is actually paid on incomes in these ranges, we have turned to the Internal Revenue Service's published figures based on the individual income tax returns filed for the years 1970, 1971 and 1972.²¹  They have a special pertinence to the cases before us and they are the most authoritative and accurate source of guidance we have been able to find. The IRS report for each of these years includes a table which lists total income tax by size of adjusted gross income.²²  From these figures one can calculate that approximately 20 percent of adjusted gross incomes between $30,000 and $50,000 is paid as federal income tax. The figure for adjusted gross incomes between $50,000 and $100,000 is 27 percent.²³  These percentages reflect any success these taxpayers may have had with the utilization of tax shelters and presumably would be much higher if there were no tax shelters. In 1973 a taxable income of $30,000 put one in the 39 percent marginal tax bracket; a taxable income of $50,000 put one in the 50 percent marginal tax bracket; and a taxable income of $100,000 put one in the 60 percent marginal tax bracket.

The trial court found that Felder would pay federal and Arizona income taxes of 25 percent of the amount of his income left after deduction of business expenses, pension plan payments and personal maintenance. This can be recalculated as 17 percent of his adjusted gross income, compared with a national average of 27 percent which, as mentioned, reflects the effects of income sheltering. Of course, some allowance must be made for individual variation. But there is nothing in the record to show that Felder's position was exceptional or that he would have had exceptional success in sheltering his income. He had four children but they reached 21 in 1969, 1970, 1972 and 1973. Neither he nor his wife would have reached 65 while his income was expected to be in a high bracket. His business expenses of 20 percent of his gross income are, of course, excluded from adjusted gross income. The testimony of his familiarity with and access to tax shelters was offered in an attempt to show that he would have utilized them. There was no evidence, however, that he had ever done so in the past. We are firmly convinced that on this record it would be a mistake to conclude that Felder would have incurred income taxes, federal and Arizona combined, of less than 23 percent of his adjusted gross income. To reflect this, we have raised his projected tax burden by amounts that have a value in 1974 dollars of $60,000.

Henschen's taxes were calculated at 25 percent of his gross earnings. The district judge did not find that he had deductible business expenses, so his gross and adjusted gross incomes are assumed to be the same. Since 25 percent is fairly close to the guidepost we have adopted, we think this amount is not erroneous.

Burns' income tax was calculated at between 10 and 15 percent when he was grossing $30,000 to $56,000, and between 15 and 20 percent when he was grossing between $59,000 and $111,000. As with Felder, the limits of these income ranges are somewhat higher than those employed in the IRS table, suggesting that our guideposts should be revised upward. However, the fact that these are gross income figures, compared to the adjusted gross income used in the IRS table, offsets this to a degree, although it is in turn offset by the fringe benefits found by the court, some of which would have had to be reported as income. In sum, we can make a rough comparison between 12.5 and 17.5 percent for Burns and the 20 and 27 percent of our guidepost. As with Felder, we are firmly convinced that an insufficient amount has been deducted for what Burns could realistically have expected to pay in federal and Arizona income taxes. Burns supported three children (two of them, born to Mrs. Burns by a previous marriage, received no award in the district court), but they would very likely cease being his dependents no later than their twenty-first birthdays, one in 1970 and two in 1978. His wife would reach 65 (thus affording him another exemption) only toward the end of his projected earnings period, in 1993, and he would reach 65 in 1998, after the period. Since his income was projected to be significantly lower than Felder's in the majority of the years projected, he would presumably have less interest in, and less means of, utilizing tax shelters. There is no evidence in the record directly linking Burns to the use of tax shelters. On this record we think that Burns would have paid in taxes at least 17 percent of the income figures noted above when they ranged between $30,000 and $56,000, and 24 percent when they ranged between $59,000 and $111,000. Consequently, to increase the tax rate to this level, we have increased his tax burden over the 28 years of his projected earnings by amounts that have a present value in 1974 dollars of $50,000.

The Government has vigorously attacked the awards of non-pecuniary damages as excessive. It argues that there is no evidence in the record, to justify extraordinary non-pecuniary awards, that the awards exceed the amounts awarded in similar cases decided under Arizona law and that they are so large that in combination with the other amounts awarded, the total must be considered to be punitive. The last of these three arguments is spurious. First, the Government quotes in support of it the second paragraph of 28 U.S.C. § 2674 which is not applicable to this suit as we have noted above. See Massachusetts Bonding & Ins. Co. v. United States, 352 U.S. 128, 77 S. Ct. 186, 1 L. Ed. 2d 189 (1956). Second, if this type of loss is properly compensable under Arizona and federal law but an excessive amount has been awarded therefor, this excess can be recognized as such and excised without the necessity of labeling it punitive. Punitive bears upon the notion of inclusion, exclusion or deduction of an element of damages. Excessiveness relates to the amount of damages. Finally, the reasonableness of an award should not be approached from the point of view of the total award granted in one case or the total award granted because of one person's death. Since the purpose of such an award is compensation of survivors for losses suffered, it follows that reasonableness or excessiveness must be viewed primarily in terms of the award granted each survivor.

On the question of non-pecuniary damages we are treading on delicate and difficult ground. Just as no human being can truly measure the sorrow of another, neither can he put a specific monetary price on it. No dollar amount could ever compensate for heartfelt sorrow and loneliness. And yet, since monetary compensation is the only type we can assess against defendants in cases such as this, we must attempt to arrive at some monetary measure.

Although in assessing the amount of damages in a wrongful death case, each case must stand on its own facts, United States v. English, 521 F.2d 63, 72 (9th Cir. 1975), Batchkowsky v. Penn Central Co., 525 F.2d 1121, 1125 (2d Cir. 1975), courts nonetheless seek to maintain some degree of uniformity in cases involving similar losses. 2 S. Speiser, Recovery for Wrongful Death § 9.3 at 11-14 (2d ed. 1975). The Government relies on Merritt-Chapman & Scott Corp. v. Frazier, 289 F.2d 849 (9th Cir. 1961), City of Phoenix v. Whiting, 10 Ariz.App. 189, 457 P.2d 729 (1969), and Fulton v. Johannson, 3 Ariz.App. 562, 416 P.2d 983 (1966). The plaintiffs rely on Southern Pacific Transportation Co. v. Lueck, 111 Ariz. 560, 535 P.2d 599 (1975), United States v. Cline, 410 F.2d 1337 (9th Cir. 1969), and Dickens v. United States, 378 F. Supp. 845 (S.D.Texas 1974). We have considered the facts in the cited cases and in this case. Using the cited cases for general guidance and relying primarily on the facts in this case, we are left with the definite and firm conviction that some of the non-pecuniary awards here are excessive and one is insufficient. United States v. United States Gypsum Co.,333 U.S. 364, 395, 68 S. Ct. 525, 92 L. Ed. 746 (1948). We start by agreeing with the Government that the facts do not justify the award of extraordinary non-pecuniary damages. There is evidence to show that all three decedents were devoted husbands and fathers. But none of the survivors has suffered any extraordinary emotional injuries requiring special treatment, hospitalization or psychiatric care.

The Felder children as a group were granted somewhat larger non-pecuniary awards than the other surviving children after due consideration is made for age differences. There is no evidence in the record that they were closer to their father or suffered more grief at his loss than the others. The only difference appears to be that Mr. Felder had a larger income. Contrary to what plaintiffs' counsel may have sought to establish at oral argument, non-pecuniary damages must be based on the relationship between the decedent and his survivors, not upon economic status.

Of the six combined Felder and Henschen children, three were 20 or older (21, 20, 20) and the other three were in their late teens (18, 18, 17). The three oldest Felder children were in college and the youngest was three weeks from graduation from high school. They were at the age when contact with the parental home could be expected to become increasingly attenuated. The fact that three of the children left the Felder home shortly after their father's death cannot be attributed solely to his death. It is very likely that this would have occurred in the normal course of events. It goes without saying that a child who receives the guidance and companionship of his parents while living with them will receive less of these benefits after he has left them. The oldest Felder son stayed in college after his father's death and went on to earn his master's degree. A son and a daughter dropped out of college and the youngest did not go to college, but it was not shown that they had planned to complete their college studies or that this situation was a result of the accident. Robert Henschen finished college after his father's death. Thomas, 18 at the time of his father's death and two years younger than Robert, was finishing college at the time of trial.

Mrs. Henschen was awarded $80,000 non-pecuniary damages compared to $50,000 for Mrs. Burns and $35,000 for Mrs. Felder. We can find no evidence in the record that Mrs. Henschen had a closer or more dependent relationship with her husband than the other widows had with their husbands. She was 45 at the time of the accident four years older than the other two widows but this could not account for the difference. She, like Mrs. Felder, was remarried for a short period between the time of the accident and the time of the trial. This is some indication that she is an attractive person who will not be isolated from social relationships. On the other hand, there is likewise no evidence that explains the lower non-pecuniary award to Mrs. Felder. She married at the age of 17 1/2 years, did not complete high school, and was dependent on her husband for the management of family finances. They spent a good deal of time together and there is testimony that her husband's death had a marked effect on her lifestyle. While the record might support a finding that Mrs. Felder's grief and loss of companionship was no more severe that that of Mrs. Henschen and Mrs. Burns, we are certain it cannot support a finding that it was less. We find no basis in the record for these discrepancies in the widows' awards for non-pecuniary losses. Thus we are left with the impression that they are predicated on the fact that Mrs. Henschen received the smallest award for pecuniary damages and Mrs. Felder received the largest such award. This, of course, is an impermissible criterion upon which to base damages intended to be compensatory.

In reducing the non-pecuniary awards of these plaintiffs, we have not reduced them to what we may have given had the matter originally been ours to decide, but only to the upper limit of what we think in justice we should allow to stand. The Government has suggested that, in this case where the pecuniary damages are large and the plaintiffs are "all normal healthy persons with no special needs requiring extraordinary care and attention," only nominal damages are required to compensate them for their mental anguish and loss of affection and companionship. But this argument would lead us to impinge seriously upon the architecture of the Act which provides for recovery according to the lex loci delictus. And the local law applicable here, that of Arizona, would require no such result. Noting our requirements of deduction for taxes and personal maintenance and reduction to present value, United States v. English, supra, 521 F.2d at 71-72, the Government continues its argument as follows:For courts to require strict adherence to legal principles in the award of pecuniary damages is a useless exercise if there are no legal principles to guide the district courts in the award of non-pecuniary damages. Indeed, to avoid having its (sic) awards overturned, district courts will be encouraged to minimize pecuniary damages and maximize non-pecuniary damages.

By cutting back on some of the non-pecuniary awards here we wish it to be understood that there must be some limit on these incalculable damages. At the same time, we are unwilling to go as far as the Government suggests. Under the Act, non-pecuniary damages, like pecuniary damages, should be substantially compensatory. The exercise of some supervisory control on one and not on the other would destroy the essential compensatory purpose of the statute. On the facts of this case where substantial grief and loss of affection and companionship have been demonstrated, we do not believe that nominal non-pecuniary damages would be compensatory.

After a careful review of each award, and in the light of what we have said above, we conclude that the amounts the plaintiffs shall recover are the following:

Mrs. Felder: pecuniary, $448,380; non-pecuniary, $50,000; total, $498,380. Harry Felder, III: pecuniary, $20,000; non-pecuniary, $15,000; total $35,000. Colette Diane Felder: pecuniary, $20,000; non-pecuniary, $25,000; total, $45,000. Richard Felder: pecuniary, $35,000; non-pecuniary, $30,000; total $65,000. Lauren Felder: pecuniary, $40,000; non-pecuniary, $35,000; total $75,000.

Mrs. Henschen: pecuniary $320,000; non-pecuniary, $50,000; total, $370,000. Robert Henschen: pecuniary, $20,000; non-pecuniary, $25,000; total $45,000. Thomas Henschen: pecuniary, $35,000; non-pecuniary, $30,000; total, $65,000.

Mrs. Burns: pecuniary, $450,000; non-pecuniary, $50,000; total, $500,000. Timothy Burns: pecuniary, $40,000; non-pecuniary, $50,000; total, $90,000.

The judgment of the district court is affirmed on the issue of liability and modified on the issue of damages. The cause is remanded with directions to enter a modified judgment in the amounts herein set forth. No costs will be allowed.