Before us are the appeal and cross-appeal from a substantial judgment in a diversity suit charging fraud and breach of contract. The plaintiff, AMPAT, a large window and curtainwall contractor operating nationwide, was awarded the subcontract to install the windows in the Onterie Center, a 60-story building in downtown Chicago. A key step in installation is to anchor the window frame to the wall, and for this one needs anchors. “Tapcon” is the trade name of a carbon steel screw designed and manufactured by defendant Illinois Tool Works for use as a masonry anchor. From Triangle Fastener Corporation, the other defendant, AMPAT bought 30,000 Tapcons at a total price of $4,700; Triangle had bought them from Illinois Tool Works for $3,000. The subcontract between AMPAT and the builder of the Onterie Center specified stainless steel— not carbon steel — anchors. Carbon steel is more susceptible than stainless steel to stress-corrosion cracking, which occurs when stress is placed on a metal object that has become corroded from contact with water (or some other corroding substance), or when a metal object that is already under stress becomes corroded. Cf.
McLaughlin v. Union Oil Co.,
The Tapcons that AMPAT bought were defectively manufactured, and the consequence, according to AMPAT, was that the cost of performing the subcontract skyrocketed. AMPAT had no recourse against the prime contractor — if only because it had failed to use stainless steel anchors, as its subcontract required. Instead it sued Triangle Fastener and Illinois Tool Works for breach of warranty, and Illinois Tool Works for fraud as well. The jury exonerated Triangle Fastener but found Illinois Tool Works liable for breach of warranty and for fraud and awarded AMPAT more than $900,000 in compensatory damages, plus $500,000 in punitive damages. The district judge set aside the
*1039
award of punitive damages on the ground that Illinois law (which the parties agree governs the substantive issues in the case) does not authorize an award of punitive damages without circumstances of aggravation not present here. AMPAT does not appeal the dismissal of Triangle Fastener (which therefore is no longer a party), but it does appeal the setting aside of the award of punitive damages. Illinois Tool Works does not appeal the finding of breach of warranty, but does appeal the finding of fraud and the award of compensatory damages. The issues of fraud and damages are connected because most of the damages awarded AMPAT were consequential damages, which are fully recoverable in a fraud case,
Tan v. Boyke,
As there is no contention that the district judge made errors in his evidentiary rulings on liability or in his instructions to the jury, we construe the facts relating to liability as favorably to AMPAT as is reasonable. AMPAT began installing windows in the Onterie Center in the fall of 1984. On January 2, 1985, it discovered broken Tap-cons among those it had used before Christmas to anchor sloped window frames. (Some of the windows in the building are not vertical, but instead slope inward.) The Tapcons had broken even though, since the glass had not yet been inserted in the frames, the frames were not bearing their full weight. An inspection two days later revealed that 30 to 40 percent of the Tap-cons in the sides of the sloped window frames were broken.
The broken Tapcons were collected and shipped off in three groups for inspection. One group was inspected by a metallurgist named Geiselman, whom AMPAT had retained for this purpose. The others were inspected by two employees of Illinois Tool Works — Lyu, also a metallurgist, and Jan-usz, an engineer who testified that he did not understand Lyu’s metallurgical reports. In addition, unused Tapcons were taken from the three boxes in which they had been shipped by Triangle Fastener and were given to Lyu and Geiselman to test. The test evaluations differed. Janusz at first found no problems. Later both he and Geiselman found defects in the Tap-cons they inspected. They were minor defects, and neither he nor Geiselman thought they had caused the Tapcons to break; nevertheless Geiselman recommended, as a precautionary measure, that AMPAT replace all the Tapcons that Triangle had shipped to it. Lyu’s findings were the most ominous: he found all six Tapcons in two of the three lots of unused Tapcons to be either seriously or marginally defective.
Enter, stage front, Patterson, the national sales manager for the division of Illinois Tool Works that manufactures Tapcons. Patterson told AMPAT’s project manager that AMPAT was installing the Tapcons incorrectly, although in a grudging and inaccurate nod to Lyu’s report he added that one of the three lots that had been inspected had been found to be marginal. The project manager wanted this in writing and Janusz drafted a letter for Patterson’s signature that described two of the three lots as marginal. In the letter as mailed, however (dated January 28), Patterson changed the number of marginal lots to one and described the other two lots as okay; the problem, he said, had been “isolated” to the one lot. He promised to ship replacement Tapcons forthwith, bypassing Triangle, and said the replacements would be “fully inspected for compliance with our product design parameters.” The letter attributed the problems AMPAT had encountered to the combination of the one lot’s marginality and improper installation.
*1040 Patterson’s letter was misleading. Not only had Lyu found defects in two lots rather than one, but he had found seriously defective Tapcons in those lots, not just marginally defective ones. Patterson’s statement that the problem had been “isolated” to one lot made it sound as if Illinois Tool Works had found one marginal production run, implying that so long as it replaced the defective Tapcons from a different run the problem would not recur— especially since all the new Tapcons would be “fully inspected.” In fact the “lots” to which the letter referred were simply the contents of three boxes in which Triangle Fastener had happened to ship Tapcons to the Onterie building site; the boxes were not identified with particular production runs. Lyu’s findings — never communicated to AMPAT — supported the hypothesis that two-thirds of all the Tapcons that Illinois Tool Works had sold to Triangle Fastener were defective. Of course Lyu had been working with a small sample. And it is uncertain that the Tapcons were so defective that they flunked the specifications in Illinois Tool Works’ sales literature. Tapcons just may not have been up to a job for which, after all, stainless steel anchors rather than carbon steel anchors had been specified. But after the initial failure and ensuing complaints by AMPAT, Illinois Tool Works knew what the Tapcons were being used for; and Patterson’s letter was an implied representation that the replacement Tapcons he was shipping would do the job for which AMPAT wanted them.
A week after Patterson’s letter was mailed, Janusz received a report of defective Tapcons at another building site. Tests run on them by Lyu revealed defects similar to those in the Tapcons at the On-terie Center. Illinois Tool Works did not tell AMPAT about this development.
In accordance with Patterson’s promise that the replacement Tapcons would be fully inspected before being shipped, Lyu tested one Tapcon in each of the 30 boxes of replacements. Twenty-six of the 30 Tap-cons that he tested flunked, but the replacements were shipped anyway and AM-PAT was not told about the test results. AMPAT received the replacements on February 12 and began installing them — and a month later discovered that a large percentage (perhaps as many as a third) of the installed replacement Tapcons had broken, just like their predecessors. At this point AMPAT stopped using Tapcons and added additional anchors (made by another manufacturer) to the windows that it had installed already. Expert evidence at trial established that the Tapcons, both the original and the replacement ones, had broken because of stress-corrosion cracking due in part to deficiencies in the manufacturing process at Illinois Tool Works’ plant.
A reasonable jury could conclude, to the degree of confidence implied by the “clear and convincing” standard applicable to a fraud case,
Beaton & Associates, Ltd. v. Joslyn Mfg. & Supply Co.,
For an omission to rise to the level of fraud, however, there must be a duty to disclose.
Latigo Ventures v. Laventhol & Horwath,
To be actionable, however, fraud must induce reliance — must in other words be both believed (if it is not believed, it can hardly be described as fraud) and acted on.
Soules v. General Motors Corp.,
This principle coexists uneasily, however, with a requirement that the victim of a fraud prove justifiable, or in other words reasonable, reliance, implying
some
duty of care on the part of the victim.
Chicago Title & Trust Co. v. First Arlington National Bank,
These attenuated duties of care that a fraud victim has are two, not one. The victim “cannot close his eyes to a known risk.”
Teamsters Local 282 Pension Trust Fund v. Angelos, supra,
The requirement of not being reckless serves in the law of fraud not only to define a duty, but also to disambiguate evidentiary ambiguities concerning the very existence of the fraud. If the victim acted recklessly in the face of the alleged fraud, it is difficult to believe that he was actually deceived; he may simply regret having assumed a risk that has turned out badly. The requirement of justifiable reliance backstops the jury’s determination of actual reliance. It is on this theory that a person who plays ostrich may be held to have acted deliberately: to know, and to want not to know because one suspects, may be, if not the same state of mind, the same degree of fault.
United States v. Ramsey,
“Degree” is the right word; we are speaking of differences of degree rather than of kind. The spectrum is continuous between involuntary conduct at the one end and deliberate action on full information at the other. There are no clean breaks marked “involuntary,” “unavoidable,” “negligent," “grossly negligent,” “reckless,” “deliberate.” If contributory negligence were the test, the victim of fraud might be barred from recovering damages if he had any reasonable opportunity to ascertain the truth of the representation. But with reasonable reliance the test, his opportunity must be “ample,” as the Supreme Court of Illinois said in
Schmidt v. Landfield,
There are reasons for the distinctions in the degrees of care that tort law requires of victims. The deliberate tortfeasor can prevent the harm at lower cost—for he need only refrain, in order to avert it—than the victim who must exert himself, even if only slightly, to avoid the harm. Moreover, transaction costs would rise if law induced buyers to regard the unequivocal assurances of their sellers with flinty-eyed skepticism—-to regard them indeed as lies necessitating potentially elaborate self-protective measures such as, in this case, AM-PAT’s directing Geiselman to test the replacement parts despite Patterson’s unequivocal written assurance that they would be fully tested by Illinois Tool Works, their designer and manufacturer. Trust is economical—a substitute for costly methods of self-protection.
Id.
at 528. The law of fraud fosters trust by protecting the trusting, though only up to a point.
Soules
speaks of “facts of which plaintiff ... ‘might have availed himself by the exercise of ordinary prudence,’ ”
An additional point may dissolve the debate over the difference between contrib
*1043
utory negligence and reasonable reliance: ordinary prudence—the test for contributory negligence as well as the term used in
Soules
in explicating (perhaps inadvertently) reasonable reliance—varies with the circumstances. It is not ordinary prudence to assume that your seller is a liar. And just because AMPAT knew that some of the Tapcons had proved defective—more perhaps than Patterson would acknowledge—it does not follow that AMPAT also knew, or was reckless in failing to discover, that the replacement Tapcons were defective too, despite Patterson’s assurances. Cf.
General Foods Corp. v. Valley Lea Dairies, Inc.,
Illinois Tool Works asks, rhetorically, why it would commit fraud exposing it to substantial liability on a sale that yielded it gross revenues of only $3,000. That is indeed a puzzle. But the puzzle, while perhaps undermining the jury’s finding of fraud somewhat, strengthens rather than weakens the case for justifiable reliance: AMPAT had no reason to think that Illinois Tool Works would behave irrationally, and to take protective measures accordingly-
The explanation for the puzzle may lie in a divergence of objectives between Patterson and his employer—the well-known problem of agency costs. However this may be, it is not a defense to fraud that the defendant erred in thinking fraud the profit-maximizing response to the problem in which it found itself. Litigation arises from the pathology of social interactions, so we should not be surprised that often parties to litigation appear to have acted from incomprehensible motives, behaved irrationally, misconceived their self-interest, exhibited bizarre deviations from rationality. It was for the jury to decide whether this was such a case, or whether, as Illinois Tool Works argued,.the defects in the Tap-cons were trivial and the main cause of their breaking was incompetent installation.
So there was liability; should punitive damages have been allowed as a sanction for it? Fraud not only is a deliberate tort but also involves concealment, and on both counts an award of punitive damages might appear to follow as a matter of course from a finding of liability. See
FDIC v. W.R. Grace & Co., supra,
The district judge found that AM-PAT had failed to make the showing required for an award of punitive damages under Illinois law. We cannot say that this finding is unreasonable, given the broad discretionary character of the Illinois standard, which invites deferential appellate re
*1044
view. “[W]here a plaintiffs factual allegations are sufficient to state a legally cognizable claim for punitive damages, the trial court has discretion to submit the issue to the jury, and the court’s determination will not be disturbed absent an abuse of that discretion.”
Motsch v. Pine Roofing Co.,
Perhaps, too, the judge thought he was doing rough justice by vacating the award of punitive damages, since the award of compensatory damages was so generous. Illinois Tool Works argues that it was too generous, and maybe this is right; but the standard is whether a reasonable jury could have awarded the damages it did, and the answer (as we are about to see) is yes, given the state of the record and the arguments by the parties. Generally courts are not supposed to do rough justice — they are supposed to do legal justice — but the Illinois standard for punitive damages gives the district judge a very broad discretion. This was a close case on liability for fraud, and it makes sense to scale down an award of damages in recognition of a distinct possibility that the judgment may be premised on a mistaken finding of liability.
Jordan v. Duff and Phelps, Inc.,
The last issue is whether the award of compensatory damages can be sustained. The principal items of compensatory damages that AMP AT sought and obtained were two. First was the cost of replacing the defective Tapcons with a different anchor. As to this item of damages, Illinois Tool Works’ only argument is that the underlying data were presented to the jury in the form of summaries that did not satisfy the requirements of Fed.R.Evid. 1006, because the underlying data had been gathered for use in litigation and therefore were not the records of a “regularly con
*1045
ducted business activity”; such records are admissible under the hearsay exception in Rule 803(6). Summaries of records prepared for litigation are indeed inadmissible.
Paddack v. Dave Christensen, Inc.,
Moreover, “if of a type reasonably relied on by experts in the particular field in forming opinions or inferences upon the subject, the facts or data need not be admissible in evidence.” Fed.R.Evid. 703. AMPAT’s damages witness was its project manager for the Onterie Center installation, and although not designated as an expert witness he would easily have qualified under the liberal standard of Fed.R. Evid. 702. Cf.
Tagatz v. Marquette University,
Illinois Tool Works mounts its main attack on what AMPAT calls its “impact damages,” consisting of the extra costs it incurred in installing windows after the broken Tapcons were discovered, and amounting to two-thirds of the total damages. The impact damages were the difference between AMPAT’s average cost of installation in the first three months of the job (October through December), before the broken Tapcons were discovered and replaced (eventually by a different anchor altogether), and AMPAT’s average cost of installation (not replacement) for the remainder of the job, multiplied by the number of windows installed after December 31. AMPAT argues that the cost of installation was higher after that date because of the disruption in the installation schedule that ensued from the discovery of the broken and cracked Tapcons. The pace of the installation slowed down, which raised the cost because AMPAT’s installers found themselves working behind rather than ahead of the crews doing other work in the building. Furthermore, AMPAT had to divert its most experienced crews to replacing the anchors in the windows that had already been installed, and the less experienced crews that were used to install the new windows worked more slowly. And the replacement anchors were of a different size, and the crews had to redrill the window frames in order to insert them.
Illinois Tool Works criticizes AMPAT’s failure to make an adjustment in its calculation to take care of other, nonculpable causes of the higher average cost of installation after December 31. In particular it stresses the colder weather in winter than in fall and notes that the Chicago winter of 1985 was a particularly harsh one. Moreover, the windows for the lower floors are installed before the windows for the upper ones and it takes longer for the crews to reach the upper floors, which were the floors on which windows were being installed after December 31. AMPAT ripostes that normally the average cost of installation falls with the progress of the job, because the installers are moving down the learning curve (learning by doing), and that the lower floors of the Onterie Center, being less uniform than the upper ones, required more complicated installation.
These various effects, although tugging in opposite directions, might not, of course, be equal and so cancel out. In a world of zero litigation costs, therefore, we would insist on the plaintiff's using modern
*1046
statistical techniques for identifying the partial effect of different explanatory variables, such as weather and broken Tap-cons. But in our real world of positive and steep litigation costs, and considering that few contract cases involve substantial stakes warranting heavy expenditures on expert testimony, we hold that the plaintiff satisfies his burden of production by presenting a simple before-and-after comparison accompanied by a plausible though not necessarily conclusive connection between the difference and the defendant’s wrongdoing. If the defendant is unconvinced, he can hire an expert to do a multivariate analysis designed to break the connection. Cf.
Allen v. Seidman,
It must stand even though there seems to be a large hole in the plaintiff’s damages case. There was no fraud until Patterson’s letter of January 28 (or, at the earliest, his telephone conversation with AMPAT’s project manager a few days earlier). Yet AMPAT’s impact damages (two-thirds of its total damages) included costs it had incurred in reanchoring the windows that it had installed before then. These damages were recoverable if at all only for breach of warranty; and to the (undetermined) extent that they were consequential damages, they may not have been recoverable on that basis. But as Illinois Tool Works has never made this argument, it is waived (perhaps because of a flaw in the argument that we have not detected). We conclude that the judgment, in its entirety, must be
Affirmed.
