(dissenting). Reversal and new trial is required because (1) the matter having been submitted to the jury only on manufacturing defect and not on design defect, evidence of subsequent design change was irrelevant, (2) evidence of subsequent design change is so prejudicial that it should be admitted, even if admissible in strict liability cases, only when, because other proof is not available, its probative value outweighs its prejudicial effect; in the instant case other evidence was available, and (3) the Trial Judge erred in denying defendant’s request to charge that future lost earnings should be discounted to present value.
I
The theory upon which the action was submitted to the jury, as the majority concedes (at p 120), was of a defect in manufacture and assembly. Straining to sustain the admission of evidence of subsequent design change in a case not submitted to the jury on a design defect theory, the majority fails to recognize the essential difference between the two, a distinction re-emphasized only last year in Robinson *128v Reed-Prentice Div. of Package Mach. Co. (49 NY2d 471), and muddies not only the law of evidence but the rules governing product liability as well.
Strict products liability permits recovery by the plaintiff upon a showing that the product which caused the injury was defective when it was put on the market (Codling v Paglia, 32 NY2d 330, 342). As the law of products liability has developed in this State, a defect in a product may arise in one of three ways: a defect in construction, through mistake in manufacturing or assembly (Victorson v Bock Laundry Mach. Co., 37 NY2d 395; Codling v Paglia, 32 NY2d 330, supra), a defect in design (Robinson v Reed-Prentice Div. of Package Mach. Co., 49 NY2d 471, supra; Micallef v Miehle Co., Div. of Miehle-Goss Dexter, 39 NY2d 376; Bolm v Triumph Corp., 33 NY2d 151), or failure to give adequate warning or instruction in the use of the product (Wolfgruber v Upjohn Co., 72 AD2d 59, affd 52 NY2d 768; Torrogrossa v Towmotor Co., 44 NY2d 709). Though this appeal properly involves only the first, the majority’s confusion of the first two requires that we differentiate defective manufacture from improper design.
Although sometimes difficult to delineate (see Barker v Lull Eng. Co., 20 Cal 3d 413; Wade, On the Nature of Strict Tort Liability for Products, 44 Miss LJ 825; see, also, Prosser, Law of Torts [4th ed], § 99), a real distinction exists, both in law and in fact, between manufacturing defects and design defects. Manufacturing defects, by definition, are “imperfections that inevitably occur in a typically small percentage of products of a given design as a result of the fallibility of the manufacturing process. A [defectively manufactured] product does not conform in some significant aspect to the intended design, nor does it conform to the great majority of products manufactured in accordance with that design.” (Henderson, Judicial Review of Manufacturers’ Conscious Design Choices: The Limits of Adjudication, 73 Col L Rev 1531, 1543.) Stated differently, a defectively manufactured product is flawed because it is misconstructed without regard to whether the intended design of the manufacturer was safe or not. Such *129defects result from some mishap in the manufacturing process itself, improper workmanship, or because defective materials were used in construction. Common examples of manufacturing defects are the proverbial hairline fracture in the exploding bottle or the flawed piece of metal in the automobile steering column. (See Victorson v Bock Laundry Mach. Co., 37 NY2d 395, supra; Codling v Paglia, 32 NY2d 330, supra.)
In contrast, a design defect is one which “presents an unreasonable risk of harm, notwithstanding that it was meticulously made according to [the] detailed plans and specifications” of the manufacturer (Robinson v Reed-Prentice Div. of Package Mach. Co., 49 NY2d 471, 479, supra.) Thus, unlike manufacturing defects, design defects involve products which are made in precise conformity with the manufacturer’s design but nevertheless result in injury to the user because the design itself was improper. Moreover, unlike manufacturing defect cases where the decisive issue is the existence of the defect without regard to the care exercised by the manufacturer, a defendant in a design defect case can avoid liability by demonstrating that he used that “degree of care in his plan or design so as to avoid any unreasonable risk of harm to anyone who is likely to be exposed to the danger when the product is used in the manner for which the product was intended” (Micallef v Miehle Co., Div. of Miehle-Goss Dexter, 39 NY2d 376, 385, supra.) Indeed, it has been said that “the standards for imposing liability for such unreasonably dangerous design defects are *• * * general negligence principles”. (Bolm v Triumph Corp., 33 NY2d 151, 157-158, supra.) Instances of design defects are the motorcycle luggage rack situated such that it aggravates injuries to the driver when involved in a collision (Bolm v Triumph Corp., 33 NY2d 151, supra), or the printing press without hand guards to protect its user (Micallef v Miehle Co., Div. of Miehle-Goss Dexter, 39 NY2d 376, supra).
In this case, though plaintiff’s complaint alleged, inter alia, that the ball joint was improperly “manufactured and assembled” and “defectively designed”, the evidence initially introduced advanced only the theory of defective *130manufacture and assembly. Ten days into the trial, however, plaintiff’s theory shifted to the concept that Chrysler’s ball joint was defectively designed. Only then was evidence of the subsequent design change introduced and were Doran and Burrill called to testify as to the addition of the plastic insert in the Chrysler ball joint. The matter was, however, submitted to the jury solely on the issue of manufacturing defect, apparently because the Trial Judge, having decided to strike Burrill’s design change testimony, concluded that there was insufficient evidence to present a jury question on the issue of design defect.1 Thus, the only issue before the jury was whether the particular ball joint in plaintiff’s automobile had been defectively manufactured because of faulty crimping or misassembly. On the question whether a mishap in the manufacturing process caused the ball joint in plaintiff’s automobile to be defective, evidence that four years later the design of the joint was changed was in no sense probative. The majority’s suggestions that such evidence, was bound to help the jury understand that Chrysler itself “eventually formed the opinion that the ball joints in Caprara’s car had a potential for movement when installed” (at p 125) and “added to the probability that it was this defect rather than other courses that produced the accident” (id.) are nonsequiturs in light of the uncontroverted evidence that there was no potential for movement when the car was on the ground, that the only “end play” in the ball joint reduced by the plastic insert was the “artificial end play” created when the automobile was lifted off the ground, and that the weight of the automobile when it is on the ground causes the ball joint to be pressed into its housing, locking it into place. The evidence of design change by Chrysler should have been excluded because it had no relevance to the only issue before the jury: whether there had been a manufacturing defect (see Hoenig, Products Liability: Views on Proposed Code of Evidence, NYLJ, Dec. 15,1980, atp 1, col 1; atp 3, col 1).
*131II
Even if we accept the majority’s view that the evidence had some probative effect, it should have been excluded because it was cumulative and its probative value did not substantially outweigh its prejudical effect.
Both briefs state that the Trial Judge admitted the post-accident design evidence on the basis of the holding in Barry v Manglass (55 AD2d 1, 7) that “the rule excluding evidence of postoccurrence repairs in negligence cases should not be applied to products liability cases.” The ruling was, thus, made on the law and not in the exercise of discretion. Accepting for purposes of argument that there may be cases in which such evidence should be admitted, not only under the usual exceptions to show feasibility, ownership or control if controverted, or to impeach, but also to establish defect in design, we conclude that it was reversible error to permit introduction of the evidence in the present case.
Barry v Manglass (supra) is distinguishable because it involved a recall letter sent under a statute (US Code, tit 15, § 1402, subd [a]) requiring a manufacturer to notify a purchaser of “any defect * * * which [the manufacturer] determines, in good faith, relates to motor vehicle safety” and which thus constituted a clear admission, albeit under force of the statute, of defect, while postoccurrence design change to the extent that it constitutes an admission at all is much less probative of defect. Moreover, both Barry and Ault v International Harvester Co. (13 Cal 3d 113), on which Barry heavily relies, proceed from the premise that the rationale for the exclusionary rule is deterrence of the manufacturer from effecting repairs. Since the manufacturer subject to strict products liability has other incentives for marketing a nondefective product, it is reasoned, it is not necessary to exclude evidence of subsequent repair or design change as an encouragement for him to make such repair or change.
While the deterrence rationale is one of the reasons usually given for exclusion, it is neither the sole nor should it be the primary reason. To reason as do Ault and Barry is to accord deterrence undue weight and to grossly under*132play the prejudicial effect of the evidence. Neither history nor logic requires or warrants so doing.
The early cases emphasized lack of probativeness and the prejudice resulting from admission as the primary reasons for exclusion, and referred to deterrence only incidentally if at all. Thus, in Morse v Minneapolis & St. Louis Ry. Co. (30 Minn 465, 468-469), decided in 1883, the basis for the exclusion was stated as follows: “[S]uch acts afford no legitimate basis for construing such an act as an admission of previous neglect of duty. A person may have exercised all the care which the law required, and yet, in the light, of his new experience, after an unexpected accident has occurred, and as a measure of extreme caution, he may adopt additional safeguards. The more careful a person is, the more regard he has for the lives of others, the more likely he would be to do so, and it would seem unjust that he could not do so without being liable to have such acts construed as an admission of prior negligence. We think such a rule puts an unfair interpretation upon human conduct, and virtually holds out an inducement for continued negligence.”
The United States Supreme Court in Columbia R. R. Co. v Hawthorne (144 US 202, 207), after quoting the Morse statement, phrased the rule in the following language: “[T]he evidence is incompetent, because the taking of such precautions against the future is not be construed as an admission of responsibility for the past, has no legitimate tendency to prove that the defendant had been negligent before the accident happened, and is calculated to distract the minds of the jury from the real issue, and to create a prejudice against the defendant” and in Smyth v Upjohn Co. (529 F2d 803, 804), the Second Circuit Court of Appeals, concluding that New York State courts did not accept the Ault rule, stated that: “The rationale behind the doctrine is that the evidence of remedial action, being based on hindsight, does not tend to show that the defendant had failed to act with reasonable care at an earlier period of time. The evidence, on the other hand, could be highly prejudicial and might discourage a person from making repairs to remedy a potentially dangerous situation.” Without unduly expanding this opinion by quoting from other *133similar out-of-State case authorities,2 we note that the instant case is the first in which this court has departed from this basic rationale, and that in a carefully reasoned opinion by Justice Richard D. Simons, the Fourth Department, in Bolm v Triumph Corp. (71 AD2d 429), distinguished both Ault and Barry and refused to permit evidence of post-accident design change in a strict liability case.3
Secondary authorities likewise fail to accord to deterrence an overriding role. Wigmore (Evidence [Chadbourn rev, 1979], vol 2, § 283, subd [4]) states the ground for exclusion as “that the supposed inference [of negligence or fault] from the act is not the plain and most probable one,” then notes that such theoretical relevance as the evidence has is overcome on policy grounds that the evidence “would be liable to overemphasis by the jury, and that it would discourage all owners, even those who had genuinely been careful, from improving the place or thing that had caused the injury” (at p 175). Except as part of the quotation from the Morse case, neither Richardson ([10th ed — Prince], §§ 168, 221) nor Fisch (New York Evidence, § 798) mentions deterrence. McCormick ([2d ed], at pp 666-669) notes that in most cases courts have not felt called upon to state the basis of the rule (at p 666, n 9). Though he writes that, with those that do, the predominant reason is the policy against discouraging the taking of safety measures rather than the lack of probative significance, he, nevertheless, concludes (atpp 668-669) that before admitting evidence of subsequent change for any purpose “the court should be satisfied that the issue on which it is offered is of substantial importance and is actually, and not merely formally in dispute, that the plaintiff cannot establish the *134fact to be inferred conveniently by other proof, and consequently that the need for the evidence outweighs the danger of its misuse” (emphasis supplied).
It is, of course, true that even the traditional exceptions to the subsequent repair rule in negligence cases involve the danger that the evidence will be misused by the jury, and that the law has in such cases relied upon jury instructions limiting the use to which the evidence can be put by the jury as a sufficient protection against misuse. That reliance has, however, been born of necessity by reason of the inability to prove an element of the cause of action, such as control, controverted by defendant, without the subsequent repair evidence. It recognizes the unfairness of permitting defendant to both deny such an element and object to evidence which strongly tends to prove that element, and leaves to defendant the choice between admitting the element (e.g., control) and thus excluding the evidence, or denying the element and thus opening the door to admission of subsequent repair or design. Such instructions have, moreover, been relied upon in areas of fault with which jurors, who in final analysis establish for each case what reasonable care means, are much more likely to be conversant than are they with respect to the technically sophisticated question of design or manufacturing defect (cf. Henderson, Manufacturer’s Liability For Defective Product Design, 56NCL Rev 625, 626). Accepting the somewhat dubious concept that such instructions can be understood and will be applied in such cases,4 we suggest that in strict product liability cases, where, as the majority points out, “it is not even essential to a prima facie strict products case that the defect be isolated” (at p 128) subsequent design change evidence, like the lure of the Lorelei’s siren song, will in almost all cases determine the result.5 Human nature *135being what it is, jurors who need only determine whether plaintiff has reasonably excluded other causes and whether the product was defective can hardly be expected to look beyond the admission which they will believe to be implicit in defendant’s having made a design change, the more particularly so when the field involved is a technical one with which the average juror has no familiarity except through the battle of experts presented before him. Thus, the very element of strict liability law on which the majority relies is the element which, in our view, according realistic significance to the probability of prejudice, requires exclusion of the evidence unless plaintiff satisfies the Trial Judge that despite real effort to uncover other evidence in proof of defect he has been unable to do so. Only in such a case can the need for the evidence justify exposing defendant to the danger of its prejudicial misuse.
To admit evidence of subsequent change on a less limited basis in strict liability cases generally is to make every manufacturer the insurer of the safety of his product. By the conclusion they reach, the majority, while verbalizing the concept that a manufacturer is not an insurer, in effect makes him just that. As recently as Robinson v Reed-Prentice Div of Package Mach. Co. (49 NY2d 471, 481 supra), we have refused to expand the scope of the manufacturer’s duty so far as to “be tantamount to imposing absolute liability on manufacturers for all product-related injuries,” yet the majority broadly declares (at p 124) that “Those who market products must stand behind them.” Such reasoning is not far from that of the Pennsylvania Supreme Court which in Azzarello v Black Bros. Co. (480 Pa 547, 553), while denying that it was imposing liability as an insurer, cast the supplier “in the role of a guarantor of his product’s safety”. Yet as Professor Wade correctly notes:6 “What is the distinction between an insurer and a guarantor?” The economics involved may ultimately justify imposing such a burden on the manufacturer in favor of *136the consumer, but if that is to be done it should be done by the Legislature not by the courts.7 The absolute rule adopted by the majority, admitting subsequent design evidence without requiring that its prejudicial effect be balanced against plaintiff’s ability to prove a case without such evidence cannot be logically justified.8
Were the balancing rule applied to the present case reversal would be required, for the evidence of change was “cumulative at best” (Knight v Otis Elevator Co., 596 F2d 84, 91). There was other evidence available to plaintiff to prove the defect for there was evidence that Chrysler knew that more than eight years prior to the sale to plaintiff of the car in which he was injured, its leading competitor had changed its ball joint design to include a plastic insert. Nor does the cautionary instruction given obviate the error, for there being other evidence available there was no reason to subject Chrysler to the possibility of prejudice which admission of the evidence necessarily involved.9
Ill
With respect to damages, Chrysler’s attorney requested an instruction “as contained in PJI 2:290 commencing at *137page 647.” Mistaken are the majority’s suggestions (at p 126) that defendant’s attorney should have asked for a charge concerning inflation (that being the function of plaintiff’s attorney, not defendant’s) and that it was necessary to ask the Trial Judge separately to notice judicially the facts that enter into the present value calculation, they being clearly spelled out in the charge requested, and the Trial Judge having in any event denied the request not for any such reason, but out of hand.
The charge referred to by Chrysler’s attorney relates to future earnings only, and the request was preceded by counsel’s statement that he assumed the court would “charge PJI 2:290 regarding loss of earnings.” Chrysler now argues that not only future lost earnings, but future medical expenses and future pain and suffering should have been discounted. As to the latter two issues, the error10 was not preserved. It was, however, preserved as to future earnings and since plaintiff had a life expectancy at the time of trial of 35 years, the amount awarded for future earnings, if discounted, could have been but a small fraction of the aggregate of annual earnings on an undiscounted basis over the full life expectancy.* 11 Since the jury’s award was a lump sum $3,600,000, we should not assume that the reduction made by the Appellate Division necessarily compensates for the failure to give the charge requested.' Indeed, in our view, since defendant presses the failure to charge as requested as reversible error, we should reverse, rather than assume that the Appellate Division’s reduction removed any prejudice from the error (especially when, as here, the Appellate Division’s reduction is made without any indication of the method by which it is arrived at). By allowing such reduced verdicts to stand we permit the *138Appellate Division to function as the trier of fact in fixing damages notwithstanding that defendant is entitled to the jury’s verdict on that as well as all other factual issues, under proper instructions on the law. Moreover, we stultify the development of the law in relation to the items that are discountable (see note 10 above), the way in which the matter should be presented to the jury (Chiarello v Domenico Bus Serv., 542 F2d 883, supra; Lawless, Computation of Future Damages; A View From the Bench, 54 Georgetown LJ 1131), the effect of inflation (cf. Spadaccini v Dolan, 63 AD2d 110), what, if anything, the jury is to be told about income taxes in relation to discounted value (cf. Coleman v New York City Tr. Auth., 37 NY2d 137; Norfolk & Western Ry. Co. v Liepelt, 444 US 490, rehearing den 445 US 972; and see Encyclopedia New York Law, Damages, §§ 1031, 1142; 1 NY PJI, p 644), areas of New York’s law of damages in need of clarification if damage awards are to retain logical relationship to the items of injury compensated for.
It would serve little purpose in a dissent to expound at length on each of those points. Suffice it to say that it was reversible error for the Trial Judge not to have charged as requested and that even if there had been no error in the admission of evidence, the matter should be remanded for retrial on the question of damages.
Chief Judge Cooke and Judges Gabrielli and Wachtler concur with Judge Fuchsberg; Judges Jasen, Jones and Meyer dissent and vote to reverse in a separate opinion.
Order affirmed, with costs.
. The fact that plaintiff presented no evidence that Chrysler ever received a complaint of failure of the ball joint plaintiff claims to have been defectively designed in the 13 million automobiles using the joint manufactured between 1962 and 1972 supports the conclusion of the Trial Judge not to submit the issue to the jury. .
. (See, e.g., Ortho Pharm. Corp. v Chapman, 388 NE2d 541, 561 [Ind]; Moldovan v Allis Chalmers Mfg. Co., 83 Mich App 373; La Monica v Outboard Mar. Corp., 48 Ohio App 2d 43; Hayson v Coleman Lantern Co., 89 Wn 2d 474. Generally, see, Costello and Weinberger, The Subsequent Repair Doctrine and Products Liability, 51 NYSBJ 463; Comment: 13 San Diego L Rev 208; Admissibility of Evidence of Subsequent Repairs or Other Remedial Measures in Product Liability Cases, Ann., 74 ALR3d 1001.)
. The majority takes the position that subsequent change evidence is admissible in any product liability case even though based upon defect in manufacturing rather than design. We disagree and would exclude such evidence in. design defect cases, essentially for the reasons stated by Justice Simons in Bolm (supra).
. (1 Weinstein’s Evidence, par 105 [05], at p 105-36; vol 2, par 407[02], at p 407-9; vol 2, par 407 [04], at p 407-23, n 26; Note, Evidence of Subsequent Repairs: Yesterday, Today And Tomorrow, 9 U Cal Davis L Rev 421, 428, 438, 440.)
. Unfortunately so far as law review literature reveals, there is no empirical data establishing the extent to which admission of such evidence results in a verdict for plaintiff. It has been characterized as “extremely damaging” (Bauman v Volkswagenwerk AG., 621 F2d 230, 233; see, also, Werner v Upjohn Co., 628 F2d 848, 853).
. (Wade, On Product “Design Defects” and Their Actionability, 33 Vanderbilt L Rev 551, 567, n 81.) As Professor Wade recognizes (ibid,., at p 570), “coherent analysis in design defect cases requires a balancing process. An absolute test for liability is not feasible unless one seeks to impose an insurer’s liability.”
. The legislative trend is toward restriction rather than expansion of liability and exclusion rather than admission of subsequent repair evidence. See the Model Uniform Product Liability Act, proposed by the United States Department of Commerce (44 Fed Reg 62714 [Oct. 31, 1979]). That act prohibits use of subsequent repair or design change to prove defect (id., at p 62728), as have a number of State statutes (Twerski, Rebuilding the Citadel, The Legislative Assault on the Common Law, 15 Trial No. 11, at pp 55, 58). But what Professor Twerski decries is not that they exclude subsequent design change evidence but that they “go well beyond the narrow issue of subsequent design change and exclude a broad range of newly developed information from the litigation scene” (id.).
. To say (at p 126) that there is no good reason to limit admission to evidence necessary to plaintiff’s case is simply to ignore entirely the prejudice inherent in such evidence. Even the Note, Evidence of Subsequent Repairs: Yesterday, Today And Tomorrow (9 U Cal Davis L Rev 421) cited by the majority, which argues for the discard of the exclusionary rule, recognizes (at p 440) that “the defendant is entitled to have such evidence excluded if irrelevant or if the trial court agrees that admission would be unfairly prejudicial to defendant.”
. Evidence that General Motors used such an insert, though no more probative than the later Chrysler design change that incorporated an insert in the joint, is far less prejudicial because it involves no element of admission by Chrysler.
. Future expenses, like future earnings, should be discounted (Wolfe v General Mills, 35 Misc 2d 996; Leasure, How to Prove Reduction to Present Worth, 21 Ohio St LJ 204, 205; 1 NY PJI, at p 640). Pain and suffering likewise should be discounted (Chiarello v Domenico Bus Serv., 542 F2d 883, 887; Leasure, op. cit., at p 207; and see, generally, Restatement, Torts 2d, § 913A).
. Defendant calculates the figures, based on salary at time of injury of $150 per week increased for wage increases by 10% for each of the remaining 35 years, at $2,113,990 before taxes, and those earnings discounted on the basis of a 6% return compounded annually at $275,041.76.