Concept
The Rule in Product Cases
Remedial measures
The Rule in Product Cases
Rule 407 now lists four things that a later safety measure may not be used to prove: negligence, culpable conduct, a defect in a product or its design, and a need for a warning or instruction.1 The first two describe fault in the ordinary sense, and the last two describe what a products plaintiff tries to show. Commentators say the amendment that added them made clear that the rule applies in product liability cases,2 because the original wording left the point in doubt.3
The doubt had a textual source. The original wording excluded a later measure only when it was offered to prove negligence or culpable conduct, yet a strict liability plaintiff recovers by proving that the product was defective and never has to show a breach of due care.4 A court reading those words literally could conclude that a redesign offered to prove a defect fell outside them. Plaintiffs have good reason to press that reading. When a manufacturer eliminates the very danger that injured the plaintiff, the change can look like powerful evidence that the safety measure was practicable and reasonable, and a manufacturer that improves a design or adds a warning implicitly acknowledges that the benefits of the improvement exceeded its costs.5 Whether a jury should be allowed to draw that inference in a strict liability case is the question on which courts split.
The answer in any case therefore depends on the forum. The federal text settles the question for federal courts. Many states did not conform their counterparts to it,6 and a minority of states, including California, decline to extend the exclusion to product liability cases, while the majority apply it across the board whenever the evidence is offered for a forbidden purpose.7 The first step in any product case is to identify the court and read its version of the rule. The sections below take the federal majority view, then the California minority view, then the ways a claim's theory and a state's text can change the outcome, and finally the diversity case, where a federal court must choose between its own rule and a state rule that would admit the evidence.
The Federal Majority View
The federal rule answers the product question in its text: a later measure may not be offered to prove “a defect in a product or its design” or “a need for a warning or instruction,” in strict liability and negligence claims alike.8 That text ended a split. Under the original wording, the Eighth and Tenth Circuits held that Rule 407 does not generally apply in strict liability cases,9 and Uptain v. Huntington Laboratory, Inc. frames the dispute in the same terms: the first question is whether Rule 407 applies to strict liability cases at all.10 Other courts applied the rule to strict liability claims, and the opinions below show why.
Cann v. Ford Motor Co. met the textual argument directly. The appellants noted that a negligence action puts in issue whether the defendant's conduct was reasonable, so the jury looks at the defendant, while a strict liability action asks whether the product was defective, so the jury looks only at the product. The court did not dispute that description; it denied that the difference mattered. The defendant must pay the judgment under either theory, wherever the jury's attention was focused when it found against him, and the policy of not discouraging remedial measures matters to defendants sued under either theory.11
The Advisory Committee's explanation of the rule supports that result. The Notes describe a twofold purpose. The first objective is to avoid unfairly prejudicing the defendant, because jurors would too readily equate a later design modification with an admission of a prior defective design and would lose sight of the relevant time frame, which is whether the product was defective at the time of manufacture and sale.12 The second is the policy of not discouraging people from taking remedial measures, which Cann treated as relevant to every defendant.13
Flaminio v. Honda Motor Co. applied the rule to a motorcycle. The rider's theory was that wobble caused his accident, and he offered two blueprints that, he said, showed Honda had thickened the front forks by two millimeters afterward to reduce wobble.14 The Seventh Circuit dismissed the distinction between negligence and strict liability as “purely semantic” for purposes of Rule 407,15 and it held the rule's policy applicable because the offer of proof assumed that a defect caused the accident and that a minor design change had eliminated it.16 On the court's analysis, strict liability cases fall into two groups. If the accident could not have been avoided at reasonable cost, the defendant has no reason to take a later measure, so there is nothing for the rule to exclude. If it could have been avoided, by removing a defect or warning of a danger, the strict liability defendant stands where a negligent one does, and admitting the later change would deter it just as much.17
The court also saw a practical problem in splitting the claims by theory. When a case goes to the jury on both negligence and strict liability, evidence of a later repair would carry a great prejudicial effect even if the court instructed the jury to consider it only on strict liability, although the court did not treat that consideration as decisive.18 Product plaintiffs routinely plead both theories, as the plaintiffs in Cann did,19 so a rule that turned on the label of the claim would break down in the cases where both labels appear in one complaint.
The majority view bars one use of the later change, and a plaintiff with a different use is not shut out. The rule's closing sentence allows evidence of later measures to prove the feasibility of precautions when that point is disputed.20 A manufacturer disputes feasibility when it denies the technological possibility of a safety measure it later adopted, even implicitly, as by asserting that the product that injured the plaintiff was the safest possible. If it explicitly admits feasibility, proof of the later measure normally should be excluded.21
Flaminio itself falls on the other side of that line. Honda argued only that there is a tradeoff between wobble and weave, and that it had judged weave the greater danger for a motorcycle built for high speeds. It never denied that reducing wobble was feasible, so feasibility, as distinct from net advantages, was not in issue, and the blueprints stayed out. The court added that if Honda's witnesses had testified that they would never have thickened the struts, the blueprints would have been impeaching.22 A motel-security case shows impeachment doing its work. The defendants testified that they had done everything necessary for a secure motel and that chain locks and peepholes would not succeed, yet they installed those devices after a guest was raped. The court held that the plaintiffs were entitled to show feasibility and to impeach, with a limiting instruction warning the jury that the evidence could not prove negligence, and that refusing the evidence was an abuse of discretion.23
Across the states the picture is mixed. Commentators count 13 states whose counterparts were amended to apply in product cases, among them Minnesota, Pennsylvania, Kentucky, and Texas, and they note that Washington's decision in Hyjek v. Anthony Industries applies its counterpart in product suits.24 Kallio v. Ford Motor Co. belongs on the same side, because it applied Minnesota's exclusion to a design defect claim. A minority of states, including California, instead decline to extend the exclusion to product liability cases, while the majority apply it across the board whenever the evidence is offered for a forbidden purpose.25 The California position is the subject of the next section.
The Ault Minority View
California reached the opposite answer under the original, narrower wording. The statute in Ault v. International Harvester Co., Evidence Code section 1151, was the model for the original federal rule,28 and it made later remedial measures inadmissible “to prove negligence or culpable conduct in connection with the event.”29 The California Supreme Court held that this wording leaves a strict liability plaintiff free to prove a later change. The decision still matters, because a minority of states follow it and its arguments are the ones every court must answer. It cannot govern a federal case under the current text, which names product and design defects.30
The facts show what was at stake. A Scout vehicle plunged 500 feet to the bottom of a canyon, allegedly because a gearbox failed from metal fatigue, and the manufacturer afterward changed the gearbox metal from aluminum to malleable iron. The court held that the plaintiff should have been permitted to prove that change.31 It reached that holding in two steps, one about the statute's words and one about its purpose.
The textual step began with what a strict liability plaintiff must prove. Section 1151 excludes the evidence only when it is offered to prove negligence or culpable conduct,32 and a plaintiff in strict liability recovers by establishing that the product was defective, without showing that the defendant breached a duty of due care.33 The manufacturer answered that “culpable conduct” was broad enough to reach strict liability, since otherwise the phrase would merely duplicate “negligence.” The court replied that culpable conduct covers faulty conduct beyond negligence, such as wanton and reckless misconduct, so the phrase keeps a meaning of its own.34
The policy step took up the other support for the exclusion, the fear that a party will hold back improvements if they can be used against it. For mass producers, the court reasoned, the exclusionary rule “does not affect the primary conduct of the mass producer of goods, but serves merely as a shield against potential liability.”35 A footnote supplied the mechanism. A distributor with thousands of goods on the market probably faces greater total liability by leaving defective products in circulation than by losing one suit in which a repair is admitted, and consumer organizations, federal agencies, and media exposure press for repair in any event.36 In effect, Ault argues that the exclusion exists to encourage responsible conduct outside court and will not do so in product cases.37 Once the statute was confined to negligence and the deterrence premise failed for mass producers, nothing remained to justify stretching the exclusion to strict liability.
Courts that decline to follow Ault answer on three fronts. The Pennsylvania Supreme Court's decision in Duchess v. Langston Corp., which responds to a Superior Court that had adopted Ault's economic-motivation rationale, collects most of the answers.38
The first front is relevance. Courts applying Ault accept that a later measure does not bear directly on the defendant's conduct, but they treat its relevance as enhanced once the question shifts from fault to whether the product is defective.39 Their critics respond that manufacturers change designs for many reasons other than remedying a defect. They add that evidence of later changes threatens to confuse the jury, by diverting its attention from whether the product was defective at the relevant time to what was done later, and by inviting jurors to equate a modification with an admission of an earlier defect.40
The second front is the mass-producer premise. That logic addresses only a segment of strict liability claims, and if it controlled, it would equally support removing the exclusion from negligence cases against mass manufacturers.41 It also looks only at large producers, although a rule of admissibility governs defendants of every size, including small manufacturers whose long-term prospects can turn on a single products liability suit.42 Even for large manufacturers, scale cuts both ways. A larger operation gains more from preventing future accidents, but it also risks more in damages for earlier accidents if the repair is admitted, so in Flaminio's words, “the effects of scale are symmetrical.”43
The third front attacks the line Ault drew between negligence and strict liability. Design defect claims use risk-utility balancing like the balancing in negligence cases, so the worries about misleading evidence and about deterring safety measures apply equally to both.44 The Pennsylvania court could not meaningfully distinguish negligent design from design defect in their effect on manufacturers' decisions to improve products, since the potency of the evidence as a weapon against the manufacturer is the same in both settings.45 It conceded that Rule 403 balancing by trial judges might be the preferable course if social policy were left out, but it found that the policy considerations tip the balance toward a general rule of exclusion.46 Plaintiffs keep a route. They remain free to present expert testimony that a design change was necessary to make the product safe, and the general rule only bars them from relying on the defendant's later change to establish defectiveness, a rule tempered by exceptions that maintain fundamental fairness.47
Justice Clark's dissent in Ault attacked the textual step. Section 1151 excludes later modifications offered to prove culpable conduct, he wrote, and culpable conduct includes conduct that breaches a legal duty. A plaintiff on a product liability theory must prove that the defendant breached his legal duty not to place a defective product in the stream of commerce, so section 1151 applies.48 The dissent added that juries in negligence or product liability trials who learn of a single change may conclude that it reflects an admission and give it decisive weight, and that this danger is at least as great in product cases as in negligence cases.49
Several states still follow Ault, by different routes. Counterparts to Rule 407 in Connecticut, Iowa, and Hawaii exempt product cases by their terms, and decisions in at least ten more states do the same: Alaska, California, Colorado, Louisiana, Missouri, Nevada, New York, Ohio, Wisconsin, and Wyoming. Other states apply the exclusion in product suits.50 Ohio's Minton v. Honda of America shows the Ault approach at work. The court held that proof that Honda equipped its 1992 Accords with air bags and positioned the shoulder belts differently than in the 1990 Accords bore directly on whether alternative designs were feasible and whether the 1990 Accord was defective, and was probative of the quality of the 1990 Accord.51 Texas reached a similar place through its rule's text. The version applied in Uniroyal Goodrich Tire Co. v. Martinez ended with a sentence preserving admissibility in strict liability product cases, and the Texas Supreme Court held that the rule therefore did not bar the evidence. Texas's counterpart has since been amended to apply in product cases,52 which shows that the list of Ault states shifts and that the forum's current rule controls.
Design claims such as the one in Barker v. Lull Engineering Co., where a worker injured at a construction site while operating a high-lift loader alleged that its defective design proximately caused his injuries,53 are the claims in which a plaintiff would offer a later redesign, so the forum's rule decides whether the jury hears it.
The forum therefore decides the outcome before any question of persuasiveness arises. A federal court has no room for Ault's reading of a negligence-only text, because the federal rule bars a later measure offered to prove a defect in a product or its design.54 A state court must be asked which wording its rule uses and whether its courts have carved out strict liability, and Ault is the usual reason they have.
Claim Type and State Variation
Federal Rule 407 settles the product question by naming product and design defects, so a federal court never has to decide whether the plaintiff's claim is strict liability.56 A state rule that keeps the original wording, or that exempts strict liability in so many words, leaves the question to be answered another way, because many states did not conform their counterparts to the amended federal text.57 The answer then depends on whether the claim before the court is a strict liability claim at all. The carve-out follows the theory the claim actually uses, not the label on the complaint.
The carve-out makes most sense for a claim that tests the product without regard to care. A manufacturing defect claim asks whether a unit left the factory in a flawed condition because it did not conform to its original design specifications, and it assumes that the design is safe.58 A design defect claim asks a different question. It uses risk-utility balancing much like the balancing in a negligence claim, which is why courts that apply the exclusion say its concerns about misleading evidence and deterring safety measures reach design claims just as they reach negligence claims.59 A state exemption for strict liability therefore reaches a design or warning claim only if the state's own products law treats that claim as strict liability.
Iowa supplies the clearest illustration. Its counterpart exempts product cases from the exclusionary principle,60 but Scott v. Dutton-Lainson Co. shows how narrow that exemption is. Stephen Scott was injured while using a boat trailer jack and sued the jack's manufacturer, alleging that the jack was defectively designed and lacked proper warnings. He offered evidence that the manufacturer modified the jack's design after his injury, and the district court excluded it under Iowa Rule of Evidence 5.407. The Iowa Supreme Court upheld the exclusion, holding that design defect and failure to warn claims are not strict liability claims and that Scott sought to introduce the evidence to prove negligence or culpable conduct.61 That conclusion rests on how Iowa's products law tests those claims, by reasonableness, including whether a reasonable alternative design was available, so the claims carry negligence principles with them.62
The label on the claim does not control. A warranty claim, for example, requires proof of the standard for a manufacturing defect, a design defect, or a failure to warn, and whether the Iowa rule applies depends on which of the three tests the plaintiff chooses to proceed under.63 By elimination, the exemption has work left to do mainly for manufacturing defect claims, which test the product against its intended design without regard to care.
Even where an exemption applies, the evidence must still bear on the theory the plaintiff chose. In Rix v. General Motors Corp. the plaintiff tried to prove a manufacturing defect by showing that a brake line contained a bad flare that let the line come out of the nut where it was fastened to the Hydrovac.64 A manufacturing defect theory assumes a safe design and asks only whether the product was built to it, so a later change in the design says nothing about whether this unit departed from the design. An exemption from the exclusion is therefore not a promise that the jury will hear the evidence. Relevance and the rule on unfair prejudice still apply, and trial judges can apply them case by case.65 A federal court's treatment of a Borough's later changes at a checkpoint shows that balancing at work. The Borough was already a defendant, the evidence of the changes would have been cumulative, and its probative value on deliberate indifference was slight because the changes came about a year after the accident and nothing showed that the Borough knew of any danger beforehand, so the court did not abuse its discretion in excluding the evidence under Rule 403.66
Exclusion in a design case does not leave the plaintiff without a route. The rule's other-purpose exceptions remain available, and feasibility is the one a design plaintiff reaches for. In Scott, the plaintiff had argued in the district court that the evidence showed the feasibility of an alternative design with a longer pin, and the manufacturer replied that it did not take the position that the pin could not be lengthened. The supreme court did not reach the question because Scott did not raise it on appeal.67 The feasibility exception receives its own treatment later in this chapter. A tally that lists Iowa among the states exempting product cases is accurate only as shorthand, since Scott shows that the exemption does not carry design and warning claims with it.
Not every state reaches its carve-out through the courts. The opinion in Burchett v. Commonwealth recounts that the Kentucky court amended proposed KRE 407 to ensure that the rule would not apply to criminal or products liability cases.68 Commentators, however, list Kentucky among the states whose counterparts were amended to apply in product cases.69 The two reports are easiest to reconcile if the rule's text changed over time, and they show why a tally of exempting and non-exempting states is only a starting point: the forum's current text controls.
Statutes can add their own layer. Arizona's Revised Statutes section 12-686(2) makes changes the manufacturer made after the product was first sold inadmissible “as direct evidence of a defect.” In Seisinger v. Siebel, discussing Readenour v. Marion Power Shovel, the Arizona Supreme Court restated how such a statute fits with an evidence rule: the legislature and the court both have rulemaking power, but when a procedural statute and a rule irreconcilably conflict, the rule prevails.70 A statutory limit on post-sale changes therefore survives only if it can be read to coexist with the rule.
The same attention to wording applies outside tort law. A court deciding a disparate impact challenge by officers seeking promotion to sergeant in Chicago observed that Rule 407's plain language does not readily apply to such claims, which are not naturally described as “an injury or harm allegedly caused by an event.” Even if the rule applied, the court reasoned, the analysis required it to address the availability of an alternative promotional method, and a later method bears on whether the alternative was available earlier, which places the evidence within the exception for another purpose.71 Admission on that ground did not decide the case, because the officers still failed to show that Chicago had an opportunity to adopt an alternative method before the 1997 promotions.72
Diversity Cases and Erie
The sections above assume a federal court applying its own evidence rules. A federal court that hears a state-law claim because the parties are citizens of different states raises a further question: does the Erie doctrine require the court to follow state practice on subsequent measures? Most modern authority says no, though the matter is controversial.74 Flaminio shows why the question matters. Wisconsin's statute was worded identically to Rule 407, yet the Wisconsin Supreme Court had held that it does not apply to a mass-produced product, reasoning that economic realities will set the course because the sooner remedial measures are taken, the less costly the defect will be to the manufacturer. A later decision extended that holding to every product liability case in which at least one of the plaintiff's theories is strict liability.75 The Seventh Circuit applied the federal rule anyway, and the rider's blueprints stayed out.
The court's reasoning begins with Congress's power over the federal courts. Article III extends the federal judicial power to diversity cases, and Article I empowers Congress to make all necessary and proper laws for executing the power to create the lower federal courts. If Congress may prescribe the qualifications of jurors in diversity cases, it may also decide what evidence should be excluded because its impact on those jurors is believed to be unduly prejudicial. Congress's judgment that juries are apt to give too much weight to such evidence is a procedural judgment, one concerning procedures designed to enhance accuracy or reduce expense in the adjudicative process.76 The court did not claim that the rule is purely procedural. Rule 407 is not based on substantive considerations only, and distrust of juries was an important though not the primary reason for it. That distrust was mild, as the exceptions built into the rule show, but it is enough to establish the rule's constitutionality in diversity cases.77
The court also took the Wisconsin decisions on their own terms and found them wanting. The benefits of a repair to the defendant, in avoiding future accidents and their costs, grow with the scale of the defendant's activity, but so do the costs of the repair, in making it more likely that the defendant will pay damages for earlier accidents. The effects of scale are symmetrical. The later Wisconsin decision dropped the mass-production rationale without proposing a substitute, and its skepticism that defendants even know about the rule does not describe large manufacturers such as Honda, which are defendants in hundreds or thousands of product liability suits and must know about it.78
A federal court applying Pennsylvania products liability law drew the same line more briefly. State law controls what is relevant, because relevance depends on the facts that matter to the claim, but Rule 407's assessment of the dangers of unfair prejudice and confusion of the issues is procedural and governs in federal court notwithstanding a state policy to the contrary.79 State law supplies the elements of the claim, and the federal rule supplies the exclusion.
The consequence is that the identity of the deciding court can settle the evidentiary question before any argument about persuasiveness begins. Had the rider in Flaminio filed in Wisconsin's own courts, the state's rule for product cases would have governed, and the redesign would not have been barred by the exclusion, although relevance and unfair prejudice would still have been open.80 The same injury, claim, and design change thus produce different rulings in the two forums. A student who sees a state carve-out for strict liability or mass producers and applies it to a federal diversity case has skipped the first question, which asks which court is deciding.