In this case we face the apparently novel question of who bears the burden of proving what future tax rates will be. That question along with other more technical claims of error comprise this appeal from a judgment holding Ayco Corporation, an investment counseling firm, liable to its client, Raymond Oddi, for errors of calculation it made in its presentation of investment alternatives. Both parties appeal, and we affirm.

Oddi served in various capacities as an executive of Baxter Travenol Laboratories, Inc. (Baxter) and a subsidiary. As part of his compensation package, he participated in Baxter's profit-sharing plan. Baxter hired Ayco to provide certain executives, including Oddi, with financial planning assistance. In 1985, Cynthia Garrett at Ayco was assigned to counsel Oddi on his retirement plan. Oddi chose early retirement from Baxter, effective February 1, 1986. This decision required him either to take lump-sum distribution of his profit-sharing funds or to roll these funds over into an individual retirement account (IRA) for continued, tax-deferred savings.

Congress caused a stir among financial planners in 1985 and 1986 when it considered and passed the Tax Reform Act of 1986 (TRA '86). Apparently concerned with the great accumulation of wealth in qualified retirement plans, Congress introduced for the first time in TRA '86 a tax on excess distributions from these plans. 26 U.S.C. § 4981A (Supp. V 1987) (subsequently renumbered as amended, 26 U.S.C. § 4980A (1988)). This tax applied to distributions whether they occurred during the investor's life or at death. Congress recognized the rethinking that such a measure would require of taxpayers, and it therefore allowed withdrawals from the affected plans through March 15, 1987 to be treated as if they had occurred in 1986 (before TRA '86 was effective). This provision resulted not only in taxpayers' avoiding the excise tax but also in their receiving the benefit of 10-year averaging and a 20% capital gains rate, both of which were no longer available after TRA '86 became effective. Tr. at 417-18. The combination of the low income tax rates then in effect (28% rate for any amount over $29,750) and the impending effectiveness of the excise tax on distributions moved Garrett to advise Oddi that he should make a lump-sum withdrawal. In January of 1986 Garrett recommended that Oddi consider taking lump-sum distribution of his retirement plan funds before the excise tax became effective, placing that amount instead in tax-free investments. Plaintiff's Ex. 514. (We refer to the proposed plan as the lump-sum plan, as distinguished from the tax-deferred plan.)

Oddi was not persuaded. Throughout 1986, Oddi and Garrett discussed the legislative developments, but Oddi continued to express a desire to roll the amount into an IRA, taking minimum distributions until both he and his wife died, at which point the funds would be distributed and become part of the survivor's estate.1 This course provided the benefit of maximum tax deferral together with minimum oversight on Oddi's part. Garrett persisted. By using various income tax rates--all of them greater than the unusually low 28% rate--which one might hypothesize as effective after 1998 (the year Oddi would be forced to begin taking minimum distributions from his IRA), she tried to show him that his estate upon death would benefit from her lump-sum recommendation. Her calculations included the assumption that the retirement plan would earn a 9% return on investment while alternative nontaxable securities would bring a 7% return. None of the 1986 calculations convinced Oddi. In February of 1987, with only a month to go before the lump-sum option would be lost to Oddi, Garrett tried again. This time, without prompting from Oddi, she ran the comparison of the alternative plans using a constant 28% income tax rate and intending to assume a 9% and 6% return on taxable and nontaxable securities, respectively. Even if the current low tax rate continued, her calculations showed more than a three million dollar advantage for the lump-sum plan. The last effort succeeded, and Oddi agreed to a full distribution of the balance of his plan funds on February 9, 1987.

Garrett's too-good-to-be-true advice turned out to be just that, however. Within the calculations she presented in the February meeting, Garrett made the important error of reversing the 6% and 9% figures, attributing the higher return to the nontaxable investments. Oddi caught the mistake, but not until May of that year. He immediately contacted Garrett, who apologized for her earlier miscalculation in a May 7 letter, but assured him that she still believed he had chosen the right course. As it turns out, there were at that point nearly two weeks remaining, during which Oddi could still roll the withdrawn amount into an IRA (presumably without tax consequences for the delay), but he did not know of this possibility, nor did anyone at Ayco inform him of it. Opinion Tr. at 3-4 (Oct. 5, 1989).

Garrett's apology was small consolation to Oddi, who had steadfastly asserted a desire to stay with the deferred-income plan until Garrett's February projection made that course appear foolhardy. Neither was her error of little consequence: while the February proposal depicted a benefit from the lump-sum plan of more than three million dollars by the year 2017, a letter written by Garrett in May with the "corrected" figures showed that the deferred-income plan would exceed the lump-sum plan by two million dollars within the same time frame if the then-current 28% tax rate was continued. (Ayco has since lowered its estimate of the deferred-income plan's advantage given these same tax assumptions.)

Oddi filed suit in both contract and tort for the difference between the projected return on his deferred-income investment and the same under the lump-sum plan. After wading through the mire of calculations from experts on both sides, the district court found in favor of Oddi, awarding him $483,088 as the present value of damages sustained plus income tax on the award. To do so, the court made several findings with respect to the variables involved in computing damages. It adopted the then-current 28% income tax rate as continuing for purposes of calculating the after-tax return on the deferred-income plan. It also adopted a 2.7% "spread"--that is, the difference between the rates of return on taxable and nontaxable investments. It further determined that the joint life expectancy of Oddi and his wife extended only until 2006, not until 2017 as Oddi's evidence contended. (The shorter time frame apparently works to the advantage of the lump-sum plan.)

Ayco does not contest the finding that a mistake in calculations was made, nor does it dispute (on appeal) that Oddi relied on its advice in withdrawing the money from the profit-sharing plan. Nevertheless, it presses two arguments in hopes of winning reversal of the district court's finding of liability. The liability finding can be broken down into findings of law and findings of fact. From this nonjury trial, we review legal conclusions de novo while we look only for clear error in fact findings. Nohcra Communications, Inc. v. AM Communications, Inc., 909 F.2d 1007, 1010 (7th Cir.1990).

Ayco's most interesting challenge questions whether a harm occurring only if tax rates remain at an arguably low level can lead to liability when future tax rates may well be higher.2 Liability thus involves a narrow but extremely difficult issue. When the future tax rate will determine whether the defendant's acts had a positive or negative impact on the plaintiff, on which party does the burden fall to prove that future rate? The question is not a simple one, but it is made even more complex in this case by a lower marginal income tax rate for the top bracket than the country has ever experienced, at least since the Twenties. Many experts, including defendant's expert in this case, believed that the 28% tax rate would not last long.

A court must adopt a rule that it can apply in times of historic high tax rates as well as historic low ones (and any in between). At trial, the district court interpreted plaintiff's position to be that the 28% rate would continue, and it understood the defendant's position broadly to be that the rate would rise sufficiently before 1998 to expunge any liability. Given this framework, the legal question we review de novo is whether the district court's placement of the burden of proving the tax rate was appropriate.

The district court's opinion with respect to the future tax rate makes clear that, ultimately, it considered the burden to be on the defendant to prove that the current rate would move upward before 1998. "Defendant urges that income tax rates will inevitably go up from their present 28 percent.... The conclusion that rates will go up, however justified as a matter of policy, is for now speculative." Opinion Tr. at 9. The importance of this determination cannot be overstated. Where the burden falls with regard to an issue as elusive as proving a future tax rate is critical to whether the trier will find liability. Indeed in many cases it may be dispositive. Parading experts before the trier might in some circumstances win the day for one party or the other in determining what the current government or some future one will do to taxes. But when reasonable minds can differ widely, the trier will quite likely be left in equipoise on the fate of future tax rates, uncertain whether the plaintiff or defendant has constructed the more persuasive analysis of the relevant economic and political facts. When the trier is thus undecided, defeat will befall the party who shoulders the burden on this question. In the case before us, placing the burden on the defendant to show that tax rates would change was appropriate.

We do not relieve Oddi of his burden of proof to show that he has been (or will be) damaged. Rather we hold that Oddi met that burden by showing that he will be damaged if tax rates remain the same. In other words, we give plaintiffs in cases like this the benefit of a presumption that the status quo will continue.

This presumption may be too human or too simplistic when it comes to predicting a future federal income tax rate given the regularity with which Congress has adjusted that rate over the years. Nonetheless, on occasions when predictions of the future rate are truly speculative, to deny plaintiffs the benefit of a presumption would absolve financial advisors of any liability for their erroneous advise.

By "error" in this setting, we mean more than a mere failure to foresee some future development or trend: that advice is a prediction upon which the reasonable investor may not generally rely. (Oddi apparently never relied on Ayco's predictive analysis in this case--witness his steadfast adherence to the deferred-income plan.) The error in this case was one of calculation, of simple number-crunching. It produced a defect in the objective comparison of the two plans given certain understood parameters. This comparison was the type of information on which Oddi could and did reasonably rely. Ayco was paid to perform this service, competent to do so and best able to prevent the sort of error it made.

Imposing liability on Ayco is not the only reason for applying a presumption. After all, the presumption we establish here will offer little comfort to plaintiffs who have relied on similarly erroneous advice but will not be damaged unless the tax rate changes. There is a strong temptation to place the burden on Ayco in every case, as the party whose error has made the exercise of prediction necessary. But we are unwilling to accept any tax rate a plaintiff proposes should a financial advisor fail to meet its burden of proof. Plaintiffs with unreasonable expectations about the future should not have those expectations guaranteed by a financial advisor's mistake.

Because a presumption makes it more likely that courts will use the current tax rate to calculate the existence and amount of the plaintiff's damages, the course of future litigation should become more predictable. Predictability should discourage frivolous suits and improve the odds that meritorious claims will be settled. Moreover, if Ayco can predict the potential cost of its mistakes more accurately, it can estimate how much it should invest in precautionary measures and how much it should pay to insure against the errors it cannot prevent.

Of course we could increase predictability by imposing any arbitrary rule. But a presumption that the tax rate will continue is not arbitrary. Ordinary legislative inertia ensures that any statutory status quo will continue for at least a few years. This natural tendency to let sleeping dogs lie may be even more pronounced when changes in the tax code are proposed, since any change implicates significant financial interests and is therefore intensely controversial.

A presumption of continuity is also appropriate for courts to use because they are courts. As judges, we are institutionally reluctant to put the legislative branch on trial, even figuratively. Moreover, we are inclined to regard the laws of the United States as immutable, even when we know, as citizens, that they are not.

By giving Oddi the benefit of the presumption, we have neither improperly abbreviated the plaintiff's prima facie burden nor unfairly tipped the scale in his favor. Plaintiffs in similar cases must still prove that a wrong has occurred and that they will incur damage under the existing tax rate. Defendants' burden will not always be a difficult one. For example, legislation might be pending which makes the case for a change in the tax rate convincing to the trier of fact. Alternatively, defendants may offer a probability analysis for a number of different tax rates from which a court could strike a statistically supportable middle ground, different from the status quo.3 These same options are also available to plaintiffs who wish to prove that the rate will change.4

Having thus upheld the district court's placement of the burden on the defendant to prove a tax rate different from the status quo, we now review its finding that the defendant failed to meet this burden. As a finding of fact, this decision is owed very substantial deference, especially in light of the important credibility evaluations with respect to either party's expert. Fed.R.Civ.P. 52(a) (1991). "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).

Given the evidence before it at trial, the district court's finding that the defendant had failed to prove a more likely rate than 28% from which to calculate damages was not clearly erroneous. On the contrary, much of the testimony pointed toward such a conclusion. Plaintiff's expert testified that he felt at the time of trial that 28% is "the best predictor of what rates will be for the foreseeable future." Tr. at 346 (Aug. 15, 1989). In response to defendant's attempt to impeach the expert with his prediction several years earlier of a rise in the tax rate, the expert noted intervening events that brought him to his new outlook: President Bush's successful lipreading campaign oratory; Director Richard Darman's statements to Congress that the administration would seek revenues in other ways besides raising the income tax; a proposal in Congress to lower the capital gains tax. Id. at 446. Moreover, defendant's expert based his belief that a "40-percent rate is very likely to happen" on the mere fact that the 28% rate was historically low, and on the chance that the State of Illinois might begin taxing distributions from retirement accounts. Neither of these comments on the 28% current rate is so telling that it leaves us with the clear and definite conviction that a mistake has been made. As appellate judges, we may have strong personal opinions on this universally debated subject, but we, as well as the trial judge, must confine our respective analyses to the evidence at trial.5

Ayco presses a second argument which, if successful, would require reversal of the finding of liability in this case. After Oddi discovered the mistake and reported it to Garrett, he still had some two weeks to roll the lump sum into an IRA, thereby reverting to the course he had originally chosen. Ayco argues that by not rolling the money into an IRA, Oddi failed to mitigate damages, and thus can recover nothing.

It is well settled in Illinois that defendant has the burden of proving plaintiff's failure to mitigate damages. Toushin v. Gonsky, 77 Ill.App.3d 508, 517, 32 Ill. Dec. 743, 395 N.E.2d 1124 (1979). The district court found that Ayco failed to meet this burden. In announcing its opinion the court stated:

The plaintiff was, in various ways, advised of the 60-day window, but that was but one item among a mass of information presented to him.... Perhaps he noted it at the time, perhaps not. It did not become material until the need to rethink the election came up, and plaintiff had long since forgotten [the rollover window], if he had ever consciously had it in mind. Plaintiff was entitled to rely upon Ayco to bring that option to mind.

Opinion Tr. at 4 (emphasis added). It is clear from this response that the court found Oddi did not know of his rollover options as of May 2, 1987. So long as the plaintiff exercises due diligence, he has a duty to mitigate damages only when he has the opportunity to do so, and opportunity requires knowledge as well as ability. Lovejoy Elec., Inc. v. O'Berto, 873 F.2d 1001, 1006 (7th Cir.1989); Fairchild v. Keene, 93 Ill.App.3d 23, 25-26, 48 Ill. Dec. 475, 477, 416 N.E.2d 748, 750 (1981). It is true that Oddi was a financial executive at Baxter, and therefore presumably was sophisticated in financial matters. Nevertheless, deciding exactly what Oddi knew in May 1987 requires a credibility determination, a task for which the district court is peculiarly suited. We cannot say that the district court's holding that Oddi lacked the appropriate information was clearly erroneous. Whether Ayco itself understood the rollover option or not--and the district court assumed that, as the expert, it did--it cannot charge Oddi with the duty to mitigate by means of an alternative about which he knew nothing.

Ayco also requests that we reduce the amount of damages awarded plaintiff below. As the most difficult question, we address first its argument that the district court's determination of the spread--the difference between the rates of return on taxable and nontaxable investments--was clearly erroneous. The spread is important in calculating damages because the smaller it is, the greater the advantage to a lump-sum distribution; if a nontaxable investment's return is close enough to a taxable investment's return, taxes will make the nontaxable investment a better alternative. In her February memo to Oddi, Garrett intended to presume a 3% spread, with the taxable deferred-income plan bringing a 9% return and the nontaxable lump-sum investment producing a 6% return. Oddi asked the court to adopt that spread figure for damage computations.

Notwithstanding its agent's assumptions, Ayco attempted to show at trial that the spread would in reality be much smaller than 3%. It did this by comparing the present and historical rates available on certain taxable and nontaxable instruments. For purposes of comparison, Ayco chose short term guaranteed interest contracts (GICs) as the likely taxable investment vehicle for the Baxter plan assets (these were the instruments Oddi believed the Baxter plan would purchase) and thirty-year AAA guaranteed municipal bonds as its choice for the lump-sum nontaxable investment. This comparison allegedly produced a spread on average of 1.42%. The district court never ruled on the validity of Ayco's calculations in this regard, because it found the comparison inapposite, and we agree.

Any comparison of investment instruments will immediately raise the question whether they involve comparable risks. Recently we have been made aware that municipal bonds may involve risk. (And even more recently, GICs offered even by major insurance companies might provide reasons for caution.) Computing the spread between two such unlike instruments must involve a degree of unreliability from the outset. Of particular significance may be Ayco's choice of nontaxable investment instruments. Thirty-year guaranteed municipal bonds may produce relatively high rates of untaxed return, but they provide no protection against inflation beyond that discounted in their yield. Thus, while the Ernst & Whinney study utilized by the court to calculate damages in this case appears to have assumed annual inflation of 4.3%, Ayco's Post-Trial Memorandum with Respect to Damage Calculations, Ex. A at 1, any significant period of higher inflation during the thirty years in question would substantially affect the relative total return on the bonds. In contrast the shorter term investments projected for the deferred-income plan, while perhaps not generating the greatest current return, would provide more flexibility in dealing with inflation.

These variables were not lost on the district court. "Defendant wants us to recognize plaintiff's stated intention to have plan assets in GICs and ignore his preference for short-term municipals with their lower rates ..., but also with greater flexibility to respond to changes in interest rates." Opinion Tr. at 11. The court found that both parties recognized that the spread would increase as the tax rate increased; equivalent taxable and nontaxable investments must relate in this way given the tendency toward equalizing after-tax income. The court then adopted a 30% tax rate (28% federal tax and 2% state tax) and, using the 9% return on taxable investments supported by the evidence, computed the spread to be 2.7% (30% of a 9% return on investment). Before us, defendant simply presses the same arguments rejected by the trial judge. We agree that the comparison Ayco proposes stacks the deck in its favor. We have little doubt that a fairer comparison might have been found, but Ayco did not present it. With the evidence that was before the district court, we do not find clear error in its decision to adopt a 2.7% spread.

Ayco's next argument is a variation on its earlier future tax rate hypothesis. Here Ayco argues that damages must be reduced because of the district court's error in refusing to acknowledge that tax rates will rise. We reject this argument for reasons similar to our analysis with respect to liability. Given the extremes presented to the district court on the question of future tax rates, we do not believe its decision to accept the 28% rate amounts to clear error. Ayco simply did not persuade the district court that tax rates of 38-40% were a sure enough bet to drive the calculations. We also note in passing that once we accept the district court's conclusion that the spread can be linked directly to the tax rate for purposes of computing damages, changes in the tax rate will cause mitigating fluctuations in the spread. Although a larger spread would not wholly offset the effect of a higher tax rate, Oddi's damages would not be reduced as dramatically as might be hypothesized.

As noted, the district court found from the actuarial evidence that the joint life expectancy of the Oddis extended until 2006. It made no such determination with regard to Raymond Oddi individually. Assuming that his individual life expectancy would be even shorter (due to his more advanced age), Ayco complains that it was error to award Oddi damages for deficiencies in any earnings his estate would have received after his death. The defendant argues that a plaintiff cannot recover for damages that may only accrue after his death. We find this argument to be without merit.

There are actually two components to Ayco's argument. First is the question of standing. Ayco invokes one of the prudential limits on a plaintiff's standing--the requirement that he assert his own legal rights and interests, not the rights or interests of third parties. Warth v. Seldin, 422 U.S. 490, 499, 95 S. Ct. 2197, 2205, 45 L. Ed. 2d 343 (1975). Defendant's argument does not withstand close scrutiny. The Supreme Court itself has suggested that third-party standing will be recognized when the rights of the real party in interest are inextricably bound up with the rights that the litigating party seeks to pursue. Singleton v. Wulff, 428 U.S. 106, 114-15, 96 S. Ct. 2868, 2874-75, 49 L. Ed. 2d 826 (1976) (plurality opinion). Our precedents make clear that this prudential limitation is flexible, and designed merely to ensure that the plaintiff is the best suited to assert the claim in question. Gorski v. Troy, 929 F.2d 1183, 1186 (7th Cir.1991). In the case before us Oddi asserts rights that are at least in part his alone; his right to be made whole up until his expected death is peculiarly his own interest. Even if Oddi's survivors have some interest separable from his own, their rights certainly are tightly bound to his. Even if the effect of Ayco's wrong is in part to injure those survivors--a proposition whose implications we will reject shortly--Oddi is still the best suited to recover for that injury. He is in fact the only person able to recover for the wrong, since the heirs to his estate will not be determined until his death, and therefore no one else currently has standing to sue.

But Ayco's argument raises a secondary issue whether it is appropriate to award a party damages that would not accrue until after his death. This claim is a seductive one, given the quantum of evidence already introduced concerning the Oddis' life expectancy. Nevertheless, the life expectancy calculation was used to determine the likely impact of the defendant's wrong; only after a joint life expectancy was determined could the court calculate the difference in projected values for the two alternative plans. Once the harm is quantified, however, this case becomes like any other action that requires a current award for damages that extend into the future. When courts award damages for future lost profits or for breach of a long-term contract, they do not insist on an actuarial assessment of a plaintiff's life expectancy to determine whether she would survive to enjoy the entire amount. Amounts owed a person at her death generally pass with her estate to her heirs, and we therefore determine present values for the entire amount of damage expected, not merely that part that is incurred during a plaintiff's life. The present case is no different. Having determined the value of the wrong to Oddi and his estate, the court must then award the entire amount (reduced to present value) to the plaintiff.

Linked to this argument in Ayco's brief is the related contention that the district court erred by assuming that after Oddi's death his wife would continue to take only the minimum withdrawals from the projected IRA account. His intent, Ayco argues, cannot be imposed on Mrs. Oddi, who might very well have other plans for the money. Any amount beyond the minimum withdrawals required would reduce the tax advantages of the Oddis' investment and therefore reduce the damages.

This argument is in essence a dispute over the district court's finding that Oddi's survivors would have carried on his effort to defer taxes as long as possible. The district court put it thus:

We assume for damages calculations that plaintiff would have used his other assets to maximize his tax benefits from a rollover, and that his wife and children, if they survived him, would have continued that program. That means that he would have taken minimum distributions and paid income and excise taxes on those distributions to the extent required, and that his wife and children would have postponed the inevitable day as long as possible.

We do not read the "assume" in the first sentence of this passage to mean that the district court did not consider Mrs. Oddi's prospective decision regarding distributions. That word was chosen because the court was addressing how the parties should proceed during the damages phase of the trial. The court recognized that the Oddis were not without other substantial means of support, even after Mr. Oddi's death, and therefore concluded that Mrs. Oddi would carry out her husband's plan of maximizing the tax deferral. This determination is not clearly erroneous.

Ayco also alleges error in the district court's decision to increase the damage award by the amount it anticipated Oddi would pay in income tax on the award. The defendant argues that such an increase is error as a matter of law, that the award is unsupported by the evidence and that the district court failed to offset the increase by the tax deductions the plaintiff would be able to take for litigation expenses. Generally courts do not increase damages to compensate for expected tax liability on the damage award. When damages place a plaintiff in the position he would have occupied had the defendant's obligation been fulfilled, the amount recovered would (but for the breach) have been income, and thus taxable. Jones v. Corbyn, 186 F.2d 450, 453 (10th Cir.1950). Since the plaintiff would have paid taxes even absent the breach, he should not be compensated for the taxes he will have to pay on the damage award he receives as a result of the breach.

Courts have adopted a pragmatic approach to the question of accounting for income taxes on damage awards, however. In wrongful death actions under the Federal Employers' Liability Act (FELA), for example, plaintiffs can recover amounts they would have received as support from a decedent. In such cases, the Supreme Court has approved the practice of reducing the award by the amount the party would have paid in income taxes on the relevant earnings. Norfolk & Western R.R. Co. v. Liepelt, 444 U.S. 490, 493-94, 100 S. Ct. 755, 757-58, 62 L. Ed. 2d 689 (1980). Even more to the point, the Court in that same case hinted that FELA damage awards might also be increased by some amount to account for the income tax the recipient would have to pay on a lump-sum award. Id. at 495, 100 S.Ct. at 758. At least two federal courts have followed the Supreme Court's lead and allowed such an increase. Sosa v. M/V Lago Izabal, 736 F.2d 1028, 1033-34 (5th Cir.1984); De Lucca v. United States, 670 F.2d 843, 845 (9th Cir.1982).6

Similarly, the case before us calls for a practical approach. Although occurring in a contract setting, defendant's error did not deprive plaintiff of taxable income. The entire purpose of comparing the deferred-income plan and the lump-sum plan was to predict which one would produce the greatest after-tax fund of capital for Oddi's estate. Therefore, damages in the case were computed as if taxes had already been paid on either basis--this being the appropriate way to compare the differential between the two plans. Oddi will now incur a new tax on the damages awarded to reflect the differential, one never required had he followed the deferred-income plan from the beginning. This new tax is solely the result of Ayco's mistake, and the company must therefore compensate him for the taxes thus imposed.

Of the cases cited by Ayco, only one is apparently on point. Ehly v. Cady, 212 Mont. 82, 687 P.2d 687 (1984), concerns a similar theory of recovery. There, Ehly arranged to purchase Montana real estate from the Cadys, informing them that the investment was in part for tax purposes. The couple eventually backed out of the deal. The court accepted that the tax advantages lost by Ehly were a foreseeable harm from the Cadys' breach, and so awarded the purchaser the lost tax credit. But the court denied Ehly's request for an additional amount to compensate for the taxes he would have to pay on the tax credit award. 212 Mont. at 98, 687 P.2d at 69 5.

With all due respect to the Montana court, we believe Illinois would follow a different rule. The scheme in Ehly would have secured for the plaintiff tax benefits based on the purchase of the property. The court found that the lost deduction should be recoverable since it flowed directly from the defendants' failure to perform. By this finding the court must have determined that one mutually understood purpose of the transaction was to provide plaintiff a tax shelter. To award a plaintiff damages for what would have been a nontaxable return on his investment yet to award nothing for a tax he would not have incurred but for the defendant's breach strikes us as unprincipled. One distinction may explain Ehly's result: there the defendants were unsophisticated, and arguably unable to foresee the tax implications of their act. Ayco on the other hand is a major national financial consultant servicing the country's wealthiest corporate clients. The tax consequences of Ayco's mistake were held to be foreseeable, Mem. and Order at 3 (January 8, 1990), making the tax award appropriate.

After taking evidence about the Oddis' current financial situation, the district court determined that the plaintiff would incur a combined state and federal income tax of 30.16%. The court then calculated the award based on damages as enhanced by income taxes at this rate and reduced it to present value (assuming that the tax would not be assessed until the year following the award of damages). Ayco, in its briefs before us, now calls this tax "hypothetical" and argues that Oddi did not prove that he would incur such an expense. Yet Ayco has not specified any consideration, either within the facts of the case or within the tax code itself, which would allow Oddi to escape this tax. Having proven his own tax rate for the relevant year, Oddi satisfied his burden of presenting evidence that shows he will be liable for the prescribed amount of taxes. If Ayco believes that something in the facts or in the tax law would allow Oddi to avoid tax liability, it is the defendant's burden to argue it. Failing to do so, Ayco cannot be heard to complain that Oddi did not support this award for taxes.

Ayco finally argues that the tax award must be reduced by any deduction that Oddi may make for legal expenses incurred in the conservation of property held for the production of income. 26 U.S.C. § 212 (1988). This argument attempts to draw an insupportable connection between the tax adjustment and any deduction for legal expenses. Plaintiff has had to prosecute this lawsuit in order to preserve his source of expected income, and we assume arguendo that he will be entitled to deduct the cost of that preservation of income. The right to deduct legal expenses in this situation would exist irrespective of whether the award included compensation for anticipated tax liability.

Ayco's argument seems premised on the following logic: the company's mistake forced Oddi to incur an unnecessary tax assessment, which is bad; but the mistake also was the occasion for Oddi to take a deduction for legal expenses, which is good; therefore, the court must offset the compensation awarded for the harm Ayco caused, which arises in part on account of taxes, by the amount of tax benefit Oddi received at the same time. But the tax deduction for legal expenses is not a "benefit" in Ayco's sense. It merely recognizes that Oddi should not be taxed for income spent to conserve income. The litigation would have been undertaken regardless whether there was an additional recovery for taxes. Hence, this element of the recovery is not a reason for making an offset for the deduction.

We can treat the remaining two contentions in Ayco's appeal, as well as Oddi's cross-appeal, jointly. All three arguments concern the appropriate treatment of the anticipated estate tax in computing Oddi's damages. The district court, facing arguments from both sides regarding the appropriate handling of expected estate taxes upon the Oddis' deaths, opted for a streamlined solution. It determined that it would compare the impact of the alternative plans on the Oddis' estate, that is, before the assessment of estate taxes.

If the plaintiff elected [the deferred income plan], he would have elected a package which brought with it tax advantages but led ultimately to certain tax consequences, which were fixed as part of that package. We do not include [estate] taxes. Those are the result of having assets at death and are not a part of the package necessarily elected. They result only from the extent of the assets then held.

Opinion Tr. at 9. This approach also paralleled the method Garrett used in making the February 1987 projection to Oddi. Appellant's Br. at 9.

This treatment of estate taxes achieved a symmetry in the computation of the two alternatives, while keeping in mind that it is Oddi's estate, not his heirs, that requires compensation. The objective for both Oddi and Ayco was to maximize the value of his estate. As such, Garrett and Oddi compared the effect of each plan measured before payment of estate taxes. Estate taxes would be assessed ultimately on the accumulation from either plan, and therefore it was not clear error to ignore their impact.

Ayco argues here as below that, if (contrary to its own arguments) the deferred-income plan leaves Oddi's estate with a larger net result than the lump-sum plan, estate taxes on that larger sum would be greater (thus marginally diminishing the deferred-income plan's advantage). First, as stated above, the goal of both parties was to maximize Oddi's estate, not his heirs' inheritance. Second, estate taxes will in fact be assessed when the Oddis eventually die, regardless whether we reduce the award now by the amount of anticipated taxes. Reducing the award now by these anticipated taxes, as Ayco requests, would result, in effect, in a double estate tax on the Oddis at the time the tax is actually levied after they die. Therefore, the district court correctly chose not to deduct expected estate taxes now.

Likewise, plaintiff's cross-appeal to the effect that the court erred in failing to increase the award by the amount that his anticipated estate taxes would have been reduced (through a deduction allowed for paying the excise tax) must fail.

Nevertheless, the court did not wholly ignore estate taxes. Ayco concedes that 26 U.S.C. § 691(c) allows the estate to deduct from its excise tax payment the sum it expects to pay in estate taxes on the amount of deferred income. This deduction worked to the substantial advantage of the deferred-income alternative--for those like Oddi who planned to defer distributions for as long as possible--by lowering considerably the 15% excise tax otherwise assessed upon death and final distribution. The 691(c) deduction affects the amount remaining in the estate before devolution to the heirs, and the district court was correct in considering it. Ayco's arguments notwithstanding, the treatment of estate taxes by the district court was consistent and therefore not clear error.

All of the alleged errors having thus been considered and rejected, the determination of the district court is therefore AFFIRMED.