Professor Page considers whether strict liability should be imposed for injuries caused by products that pose generic risks – risks that do not derive from flaws in the manufacturing process but from product design or from the very nature of the product. He reviews the ALL debate that preceded adoption of section 402A of the Restatement (Second) of Torts and finds the ambiguous meaning of comment k. which deals with “unavoidably unsafe” products, of little use in determining whether section 402A applies to generic product risks. After examining the policy justification for imposing strict liability in cases involving design defects and construction defects, Professor Page concludes that, at least in cases involving generic product risks that were unknown at the time of sale. strict liability should be imposed as a modest incentive to manufacturers, to improve product safety and as a means of satisfying justifiable consumer expectation.
Recent litigation involving asbestos and DES has attracted widespread interest, not only because of the staggering numbers of claimants alleging serious harm from these products and the filing of a bankruptcy petition by the nation’s largest asbestos manufacturer, but also because of the complexity of the issues that the cases involve For example, many DES claimants, daughters of women who took the drug during pregnancy, are unable to identify the maker of the particular pills consumed by their mothers. The courts have had to decide whether to depart from traditional causation rules that would require directed verdicts for defendants, and if so, what new rules to adopt. In the asbestos cases, courts have had to determine the obligations of successive insurers to indemnify asbestos manufacturers against claims made by persons who allegedly contracted respiratory diseases from continuous exposure to asbestos over many years. In addition to these problems, an array of legal theories asserted against an array of defendants who do not manufacture asbestos or DES has emerged in these cases.
Generic Product Risks: The Case Against Comment k and for Strict Tort Liability, Georgetown University Law Center, 1983.
The few courts reaching the merits of claims made by asbestos and DES victims have, for the most part, refused to venture beyond the familiar confines of negligence law. Giving dispositive weight to section 402A of the Restatement (Second) of Torts, which imposes strict liability for “any product in a defective condition unreasonably dangerous to the user, and to comment k of section 402A, which recognizes an exception to strict liability for products deemed “unavoidably unsafe, these courts in effect have required plaintiffs to establish that defendants engaged in unreasonable conduct. Under this analysis, if the benefits of a product outweigh its known risks, and if the manufacturer has provided suitable warnings and directions for use, the defendant’s product will be deemed reasonably safe, and the plaintiff will not recover. Similarly, if the manufacturer has placed the product into the stream of commerce without knowledge of the dangers associated with its use or consumption, courts typically have refused to impose liability unless the exercise of reasonable care would have uncovered the hazards. 11 One notable exception to this trend is a recent decision by the New Jersey Supreme Court, holding that an asbestos producer might be strictly liable in tort for injuries caused by risks that were unknown despite reasonable investigation at the time of sale.
The reluctance of courts to impose strict liability in toxic-product cases corresponds to a trend, reflected in scholarly musings and adopted in recent congressional reform efforts, to limit strict liability to product defects attributable to the construction or manufacturing process. With respect to claims alleging inadequate product design, warnings, or instructions for use, the proponents of this limitation would apply a negligence test, either expressly or in a disguised form.
Although the desirability of imposing strict liability upon the pharmaceutical industry for adverse drug reactions has been debated, the larger issue of whether all manufacturers should be held liable without fault for other types of toxic adverse effects of their products largely has escaped scrutiny. Since courts in a number of jurisdictions may soon be addressing the merits of asbestos and DES cases, a fresh look at the subject seems in order.
The central focus of this Article is whether all “generic product risks” should be treated alike. The Article first will discuss the various types of generic risks-avoidable and unavoidable, known and unknown-including those risks associated with toxic products like asbestos and DES. It will then argue that section 402A of the Restatement and its comments provide little guidance in deciding cases that involve generic risks, and should not be accorded dispositive weight in product liability suits. The Article will then examine and evaluate the policy justifications for adopting a rule of strict tort liability in cases involving generic risks. Ultimately, the Article will conclude that a persuasive case can be made for imposing strict liability on manufacturers whose products contain unknown generic risks.
I THE NATURE OF PRODUCT RISKS
Risks attributable to flaws or impurities caused by the manufacturing process usually are present only in a small percentage of the units of a particular product and do not endanger every consumer of the product. Such product risks are nongeneric in nature. The presence of a foreign substance in a jar of mayonnaise and a malfunction in a television set due to poor workmanship exemplify this category of hazards. In contrast, asbestos and DES share a common characteristic: the capacity to create risks that endanger, but do not necessarily harm, every user or consumer of the product. Such product risks are generic in nature.
This Article will focus on generic product risks, of which there are two main types. One includes design risks, or risks that can be eliminated or at least reduced by changing the design of the product. For instance, the interior of an automobile can be made more crashworthy so that the occupant is more likely to survive a collision. Some design risks, however, may be impossible to eliminate or to reduce without frustrating the purpose for which the product is marketed. The sharpness of a knife, the heat of a stove, and the physical force generated by an automobile are examples of this type of risk. These hazards enable the products to do what they were meant to do; they are essential to the function of the product and cannot be designed away.
The hazards associated with toxic products like asbestos and DES represent the second main type of generic risk. The manufacturers of asbestos products and DES have no desire to create the hazards associated with their products because these hazards serve no useful purpose. Unlike the capacity of a knife to cut, which is essential to its intended use, the capacity of DES to cause cancer in the daughters of mothers who used the drug is irrelevant to the effectiveness of the drug; while the cutler consciously designs the cutting edge of a knife, the pharmaceutical company does not intentionally create the risk of cervical cancer. Toxic product risks are inherent in the nature of the product, regardless of its design, and cannot be eliminated, at least given the current state of scientific knowledge, by any means short of withdrawing the product from the market.
Other examples of generic, non design risks abound: adverse reactions to drugs and exposure to harmful chemicals; the risk of cancer from smoking cigarettes; the risk of “toxic shock” from using tampons; and the possibly deleterious effects of consuming food and beverages containing saccharin and caffeine, if these substances were someday linked conclusively to diseases in humans.
As the saccharin and caffeine examples suggest, different types of generic risks, whether designed into a product or inherent in its nature, may also be distinguished by the degree of existing knowledge about them. Some generic risks, such as the risk of cancer from smoking cigarettes, are well known to manufacturers and consumers alike. Other generic risks, such as the carcinogenic effects of DES, were unknown when the consumer was exposed to them. Still others, such as the possible side effects of caffeine, remain unknown today.
This Article discusses whether or not these various generic product risks-designed-in and inherent, known and unknown-should be treated alike for purposes of applying strict liability. Should the rights of a plaintiff whose hand is burned by a hot stove or whose eye is injured because a machine tool lacks a safety device be determined by the same theory of liability that determines the rights of a plaintiff disabled by exposure to toxic asbestos fibres or DES? Should the claim of a patient harmed by an adverse side effect known to be associated with a drug be governed by the same theory of liability as is the claim of a patient injured by an adverse side effect that was unknown at the time the drug was administered? The light shed on these questions by the Restatement (Second) of Torts, which has greatly influenced the development of product liability doctrine, is an appropriate starting point.
II GENERIC PRODUCT RISKS AND THE RESTATEMENT
Section 402A of the Restatement (Second) of Torts gave impetus to a profound and far-reaching change in the law of product liability. It subjected sellers, including manufacturers, of all products to strict liability and grounded the cause of action in tort rather than warranty. This change was important because a warranty cause of action was contractual in nature and was being preempted by the Uniform Commercial Code. More importantly, this change relieved plaintiffs of the need to establish a privity-of-contract relationship with defendants. This so-called “citadel of privity,” preventing plaintiffs from asserting breach of warranty against defendants with whom they were not in privity, already had almost totally collapsed in warranty cases involving products for internal human consumption, and was crumbling under the onslaught of plaintiffs injured by manufactured goods. The widespread judicial adoption of section 402A completed the demolition and seemed at the time to be the most dramatic aspect of the new rule.
This doctrinal revolution was remarkably swift. What began in 1958 as a modest proposal for strict tort liability for the sale of food “in a condition dangerous to the consumer,” was extended three years later to cover “other products for intimate bodily use” in a “defective condition unreasonably dangerous to the consumer.” By 1964, the final form of section 402A applied to “any product.” This expansion of the strict liability rule, however, was not accompanied by a thorough analysis of the implications of bringing new classes of products within the sweep of section 402A. As a result, the Restatement does not adequately address the issues raised by generic risks.
A. The Restatement Generally
When the drafters of the Restatement broadened the scope of section 402A to cover all manufactured goods, they apparently assumed that the doctrine and explanatory comments, which had been developed for food and other products “for intimate bodily use,” would apply equally well to all manufactured goods. The final version of the section and its comments, therefore, remained virtually intact.
In retrospect, the most significant impact of this rush to strict liability was the confusion and uncertainty that subsequently plagued product-design litigation. Although the concept of design defectiveness was not unknown in 1964, the proponents of section 402A saw no need to adjust the rules to determine explicitly when the new doctrine would impose strict liability for design defects. They retained the terms “defective” and “unreasonably dangerous” and added the requirement that the product “must be dangerous to an extent beyond that which would be contemplated by the ordinary consumer.” In subsequent years, courts and commentators alike have found this formulation inadequate and have struggled in vain to fashion an acceptable test for strict liability in product-design cases.
Although the issue of design defectiveness was not recognized as a problem during the evolutionary stages of section 402A, certain other generic risks did occupy the attention of Dean William E. Prosser (the Reporter of the Restatement (Second) of Torts), his advisers (the American Law Institute Council), and the American Law Institute (“ALI”) membership. In working out the new rule of strict liability, they were cognizant of the controversy over the causal relationship between cigarette smoking and cancer, as well as of the incidence of serious harm attributed to certain drugs and vaccines, and considered whether the tobacco and pharmaceutical industries should be subject to strict liability. In their floor debates, Dean Prosser and members of the ALI also considered how whiskey would fit into their scheme of liability.
With respect to cigarette-cancer litigation, the Restatement came out unequivocally on the side of the tobacco companies. During a 1961 floor debate on section 402A, a motion was made to delete the word “defective” on the ground that the “unreasonably dangerous” requirement was an adequate test for determining when strict liability should apply and that therefore the term “defective condition” constituted excess baggage. In response to this motion, Dean Prosser pointed out that the ALI Council wanted to retain the element of defectiveness in order to insulate from liability the sellers of dangerous products, such as whiskey, cigarettes, and certain drugs, which are inherently dangerous even though there is nothing “wrong” with them. The specter of alcoholics bringing a barrage of suits against distillers apparently haunted the drafters of section 402A. After a very brief discussion, the motion was defeated by a voice vote, and the “defective condition” standard remained a part of section 402A.
The notion that section 402A would apply only to defective products-products that have something wrong with them other than their inherent danger-would seem to exclude most generic risks. It is not clear, however, that this interpretation is what the majority of the ALI had in mind. During the 1961 debate, Dean Prosser agreed with other members that the “unreasonably dangerous” standard was sufficient to protect sellers of products such as cigarettes and whiskey. In drafting comment i to section 402A, he pointed out that many products, including food and drugs, involve “some risk of harm, if only from over-consumption,” but this risk did not render such products “unreasonably dangerous.” Dean Prosser concluded that the proper test was whether the product was “dangerous to an extent beyond that which would be contemplated by the ordinary consumer who purchases it, with the ordinary knowledge common to the community as to its characteristics.” Thus defined, the requirement of unreasonable danger would not be met in cases involving whiskey, the hazards of which are known universally, but might be met in cigarette cases, depending upon the court’s determination of what the ordinary consumer knew about the risks of smoking at the time of marketing. Toxic risks are not necessarily excluded, therefore, from section 402A.
Another way to approach the scope of section 402A is to ask whether a product with any kind of generic risk, which was found to be unreasonably dangerous, would meet the separate requirement of defectiveness. The comments to section 402A do not answer this question. Comment i presents examples that shed little light upon the problem. The examples contrast generic risks that are not considered unreasonable (“good” whiskey that makes some people drunk, “good” tobacco that causes harm, “good” butter that deposits cholesterol in the blood and leads eventually to heart attacks) with those that do present unreasonable dangers attributable to defects in the same products (whiskey contaminated with a dangerous amount of fuse oil, tobacco with marijuana, butter with poisonous fish oil). The former pose dangers widely known to the ordinary consumer; the latter present clear instances of something “wrong” with the product. Neither group of examples presents a product, not otherwise defective, with such unreasonable risks that strict liability ought to apply.
Comment g, elaborating upon the concept of “defective condition,” is similarly unhelpful. It limits strict liability to situations where ••the product is, at the time it leaves the seller’s hands, in a condition not contemplated by the ultimate consumer, which will be unreasonably dangerous to him.” The word ··condition,” like the contaminated product examples, seems to suggest that there must be something “wrong” with the product beyond any inherent capacity to cause harm.
Yet Dean Prosser and the ALI did not intend to exclude from section 402A all products creating generic risks. Comment j states that warnings may be required for “poisonous drugs or those unduly dangerous for other reasons” (categories broad enough to embrace medicines triggering deleterious reactions), a proposition compelling the conclusion that the failure to include such warnings might subject the manufacturer to strict liability. While the comment specifies that the absence of directions or warnings may render the product unreasonably dangerous, it does not explain whether unreasonably dangerous also means that the drug is in a “defective condition.” Does comment k shed any light on the meaning of “defective”?
B. The Meaning of Comment k
Comment k, dealing with so-called “unavoidably unsafe products,” is more expansive than these other comments. It declares that a drug with proper directions and warnings would be neither defective nor unreasonably dangerous, thus suggesting that the same characteristic (mislabeling) that made the drug unreasonably dangerous might also make it defective. This wording blurs the distinction between the two elements, and the requirement of a defect thus becomes superfluous.
The genesis of comment k may help explain this blurring and comment k’s other mysteries. Dean Prosser drafted the comment in response to a proposal at the 1961 ALI meeting that prescription drugs be specifically excluded from section 402A. The arguments and the discussion that followed were notably unfocused. The motion under consideration failed to distinguish between harm from adverse reactions and other kinds of drug-induced harm, such as that caused by improper formulation or toxic ingredients. Since no one could argue seriously that the latter risks should escape strict liability, the failure to separate the two categories muddled the debate. Moreover, neither Dean Prosser nor the ALI member who made the proposal indicated how he thought section 402A would apply to prescription drugs in the absence of an explicit exemption. A solution was being offered for a problem that never had been clearly defined. Nor were adverse reactions about which warnings had been issued at the time of marketing distinguished from other harmful effects not discovered until later.
There was also disagreement over the scope of the proposed exemption. The motion proposed to insulate all prescription drugs from strict tort liability. Dean Prosser suggested that a better case could be made for excluding “relatively new, experimental, and uncertain drugs, of which there are a great many on the market, and justifiably so.” He defined the term “experimental drug” to include virtually all prescription drugs and even some over-the-counter medicines. Dean Prosser’s use of the adjective “experimental” went far beyond clinical testing, an initial stage of the Food and Drug Administration (“FDA”) approval process, and covered drugs that had completed the entire approval process and had been marketed to consumers. Thus, he was suggesting an exemption even broader than that proposed by the motion. The motion to include an exemption for prescription drugs in section 402A ultimately was defeated, as was a subsequent motion to insert such an exception in the comments. On its face, this defeat did not seem to reflect a desire by the membership to exclude more than prescription drugs from section 402A, but Dean Prosser apparently saw things differently.
Reflecting the murkiness of its origins, the version of comment k that emerged from the Reporter’s hand failed to delineate in any meaningful way either the breadth of its coverage or its purpose. The comment first addresses “unavoidably unsafe products, .. which it defines as “products which, in the present state of human knowledge, are quite incapable of being made safe for their intended and ordinary use.” The comment then appears to focus on “the field of drugs,” where such products are “especially common,,. and presents three overlapping categories of unavoidably unsafe products: high-benefit, high-risk drugs, such as the vaccine used for the treatment of rabies; “many other drugs, vaccines, and the like, many of which [because of high risks involved] cannot legally be sold except to physicians, or under the prescription of a physician;” and “many new or experimental drugs.”
The comment furnishes no criteria for determining how risky and how beneficial a drug must be in order to qualify under the first category as “unavoidably unsafe.” In any event, such a determination would appear to be unnecessary for drugs. The second category may reasonably be read to include all prescription drugs, since federal law mandates that any medicine with toxic effects that render it unsafe as self-medication be sold under prescription -and a high-risk, high benefit drug surely would be limited to sale by prescription. The sweeping requirement of prescription status also makes the third category superfluous, a fortunate occurrence since the term “new or experimental drugs” is highly ambiguous.
Thus, if its examples are taken seriously, comment k reasonably could be read as excluding from section 402A only unavoidably unsafe prescription drugs. The comment, however, fails to explain what might render an unavoidably unsafe product “defective” and thus subject to section 402A in the first instance. Instead, it states that if the known benefits of one of these products outweigh its known risks, it would not be considered “unreasonably dangerous, .. provided that it was prepared properly and bore adequate warnings and directions for use. The negative implication of this statement radically expands the scope of the exemption. Since injury caused by any product whose risks outweigh its benefits presumably would be actionable under traditional negligence principles, comment k may be read to remove from the reach of section 402A any product that is unavoidably unsafe as long as the manufacturer will not be subject to liability under a negligence rule for injury caused by the product. Such an exemption includes but is not limited to prescription drugs, an ironic turn in light of the ALI vote rejecting the proposed exemption for prescription drugs alone.
To appreciate the effect of this interpretation of comment k, it is necessary to consider how sellers of unavoidably unsafe products might be held strictly liable in the absence of comment k. The consumer-contemplation test of comment F seems to preclude liability in cases where the risks generally were known and therefore within the contemplation of the ordinary consumer. Under this test, if a patient suffers harm from a high-risk, high-benefit drug and the harm falls within the scope of the contemplated risk, the drug would not be unreasonably dangerous. Similarly, a warning about an adverse reaction listed on the label of a prescription drug would be considered part of the contemplated risk, as would be true of known risks posed by experimental drugs. Given the broad sweep of comment i, one can salvage independent meaning for comment k only by surmising that, without comment k, harm from unknown risks, or harm from known risks which turns out to be much graver than expected, generally would be actionable under theories of strict tort liability. With comment k, therefore, one must surmise that a manufacturer of a product posing such risks would escape liability under section 402A if the product were “unavoidably unsafe.”
This analysis suggests that the function served by comment k is to exempt unknown risks created by unavoidably unsafe products, since comment i already excludes known risks. Yet this interpretation presents difficulties. The text of comment k is not at all specific on the point, and a matter as important as the treatment of unknown hazards merits direct mention. Moreover, the comment focuses on known risks. Two of the three categories listed in the comment involve products unavoidably unsafe because of known risks, such as a rabies vaccine. According to the comment, the manufacturers of these drugs should not be strictly liable for harm from the known risk, a proposition seemingly rendered superfluous by comment i. The third category, “new or experimental drugs,” however, does cover products that are unavoidably unsafe because of unknown risks. Indeed, one important purpose of the clinical testing of experimental drugs is to learn more about adverse reactions they might cause. On the other hand, since a patient participating in clinical trials must give an informed consent, which includes an understanding that the harmful effects of the drug are not yet fully known, any adverse reaction the patient suffers may be said to fall within the range of consumer contemplation.
Comment j, unlike comment k, speaks specifically to product risks unknown at the time of marketing; but comment j raises more questions than it answers and sheds little light on the meaning of comment k. In discussing the duty to give warnings and directions for use, Dean Prosser indicated that the sellers of food need not provide warnings about common allergic reactions to their products, since they might reasonably assume that consumers who suffer from the allergy are aware of it. This conclusion is consistent with the consumer-contemplation of unreasonable danger test in comment i: to the ordinary consumer with a common allergy, an allergic reaction would be an expected hazard, and hence not unreasonable. The Reporter went on to state, however, that
where . . . the product contains an ingredient to which a substantial number of the population are allergic, and the ingredient is one whose danger is not generally known, or if known is one which the consumer would reasonably not expect to find in the product, the seller is required to give a warning against it, if he has knowledge, or by the application of reasonable, developed human skill and foresight should have knowledge, of the presence of the ingredient and the danger. Likewise in the case of poisonous drugs, or those unduly dangerous for other reasons, warnings as to use may be required.
This language is unclear on a number of points. Why should the duty to warn unwary allergy victims be limited to cases in which a “substantial” segment of the populace is affected? Under ordinary negligence principles, one might find the risk of serious harm or death to a miniscule percentage of individuals, or even a single individual, to be sufficient justification for requiring a warning. Also, if the risk is undiscoverable in the exercise of due care and hence need not be mentioned in the warnings or instructions for use, does it follow that the manufacturer will not be strictly liable for harm resulting from the risk? This seems to be a fair reading of the text. If so, strict liability will not attach even though the product was dangerous beyond the contemplation of the ordinary consumer.
But what are the reasons for this departure from the comment i test? Does the last sentence of the paragraph indicate merely that drugs fall within the scope of the general duty to give warnings or directions in every case? Or does it mean that allergic reactions to drugs should be governed by the same principles applicable to reactions to food, i.e., that users need not be warned about common risks that are known by both the manufacturer and the consumer? Should it be read even more expansively to preclude liability for harm from all unknowable adverse drug reactions, and, by extension, from all unknowable generic risks? If this gloss on the language of comment j is correct, comment k again would serve no purpose.
Another noteworthy aspect of comment k is its suggestion that strict liability not be imposed on the manufacturers of “new or experimental” drugs containing harmful or impure ingredients that could not be eliminated “because of lack of time and opportunity for sufficient medical experience.” The scope of the “unavoidable product danger” exception would be extended beyond generic risks and would apply to garden variety defects, where something is actually “wrong” with some units of the product. Such a view reads into comment k an “‘impure ingredient” exception.
If the risk of an impure or otherwise deleterious ingredient is known when a drug is marketed, but the manufacturer could not discover Which doses contained the substance (as is the case of blood contaminated with serum hepatitis), an adequate warning on the label of the drug would place the defect within the scope of consumer expectations. The product thus would not be unreasonably dangerous under the comment i test. Impure ingredients whose presence is not known when the drug is sold (such as the offending agents in the polio vaccine case) pose a more difficult problem because of their similarity to impurities in food and manufacturing defects in mass-produced goods. The seller may be unaware of these defects and may be unable to discover them by economically feasible methods. But these instances are plainly covered by the strict liability rule of section 402A.
The comment k “impure ingredient” exemption should not apply to either of these cases. The exception should be narrowly limited to emergencies in which the usual precautions for assuring the purity of ingredients have not been taken, yet there is medical justification for using the drug. The appropriate scope of the exception is thus so narrow that the exception would make more sense as an interpretation of the consumer contemplation test of comment i than as an exception to the strict liability rule of section 402A: in this particular context, assuming an adequate warning has been given, the risk of harmful ingredients is within the ambit of consumer contemplation.
In conclusion, the Restatement’s treatment of generic risks falls short on several counts. The requirement of a “defect” as a distinct element of strict liability was inserted to serve a function already adequately addressed by the “unreasonably dangerous” test. The Restatement fails to make a clear distinction between known and unknown hazards, and never takes a forthright position on which of these two types of hazards strict liability should cover: either, neither, or both. This omission is surprising given the evident concern, reflected both in the ALI floor debates and the comments, over the effect of section 402A upon the manufacturers of drugs, vaccines, and cigarettes. Comment k also is vague in that it fails to make clear what kind of special rule it puts in place, what purposes it meets, and to what classes of products it applies. Finally, the ALI’s position on generic product risks, uncertain though it may be, reflects policy judgments. While the ALI is a distinguished body, it is a private, nongovernmental entity. The courts have ultimate responsibility for translating policy into common-law rules, and the matter of liability for generic risks, and for toxic products in particular, requires more comprehensive scrutiny than has been afforded by the Restatement.
III GENERIC PRODUCT RISKS RECONSIDERED
When the Restatement’s commentary on adverse reactions to drugs, food, and tobacco was drafted, the proposed rule of strict liability did not cover all products placed in the stream of commerce. Thus, there was no need to consider how the full range of generic risks should be integrated into the framework of a strict liability system. Even had the drafters reflected on this issue, their efforts may not have produced an internally consistent doctrine to cover harm from the ill effects of products for human consumption and intimate bodily use, and harm from the designed-in dangers of massproduced goods, for the problem is not an easy one.
There are two basic approaches to the issue of liability for the deleterious effects of generic risks. One approach is to focus on strict liability as it has evolved in design-defect and warning cases, and to ask whether the manufacturer’s duty to eliminate or warn of product dangers extends to the particular generic hazard in question. The other approach is to ask whether the policy justifications for imposing strict tort liability in cases involving nongeneric risks, i.e., construction defects, where there is general agreement that it should be imposed, support the extension of strict liability to cases involving generic risks. Each of these approaches will be considered in the remainder of this section.
A. Justifying Strict Liability for Generic Risks: Is the Duty in Design-Deject and Warning Cases Adequate?
Under well settled principles of negligence law, a manufacturer has a duty to use reasonable care in the design of a product. This obligation requires the manufacturer to use precautionary measures which are economically and technologically feasible, and which will eliminate unreasonable risks of harm. The duty extends to risks of which the manufacturer is aware and, in the exercise of due care, should be aware. If a hazard may be reduced by providing information to the user of a product, the duty of reasonable care may be discharged by providing instructions and warnings.
To have meaning in design cases, the concept of strict liability must make the manufacturer answerable for product-related harm for which negligence theories would provide no remedy. Strict liability potentially might extend to all generic risks, to risks that are designed into a product as well as to those naturally and unavoidably present. The failure to design out or to warn against these risks would render the manufacturer liable, even though the design change or warning might be economically or technologically infeasible, and even though the risk may have been unknown or unknowable at the time of production.
A theory of “ultra-strict” liability for harm from all generic hazards has found neither judicial nor scholarly acceptance. As Professor Gary Schwartz has argued in a similar context, if loss spreading is our goal, we ought not to adopt a rule that discriminates against the victims of nonproduct-related accidents. Courts adopting “ultrastrict” product liability would find themselves on the fabled slippery slope and would be unable to offer any logical reason for not extending the doctrine to other contexts in which the public is routinely exposed to the risk of injury, such as the operation of premises held open for business or public purposes or leased to tenants. Such radical changes in the common law surely and properly would encounter judicial hesitation, grounded upon the conviction that it would be more appropriate to leave the difficult policy judgments involved in adopting such an expansive rule to the legislature.
The rejection of “ultra-strict” liability leaves open, however, the theoretical possibility of imposing strict liability for some harm caused by generic risks. For example, suppose an automobile manufacturer is deemed not liable for all harm to occupants who collide with the interior of a vehicle. Is there any way to assign responsibility for some but not all injuries attributable to the generic risks of the so-called “second collision” -to assign responsibility in fewer than all cases. as would be done under a rule of ultra-strict liability, yet in more cases than would be done under a rule of negligence? In other words, are there second collisions that the manufacturer could not have avoided by exercising reasonable care but for which the manufacturer should be held liable? This question has provoked considerable academic debate, much of it sharply critical of courts that have answered “yes” and imposed liability for injuries that were not reasonably avoidable without articulating a clear, workable standard for deciding when an alleged design flaw is defective or unreasonably dangerous. The emerging consensus seems to be that design defects are best dealt with under a balancing test, which is indistinguishable from the negligence standard. Thus, the failure to develop judicially administrable criteria for strict liability has led to the conclusion that product manufacturers, absent negligence, should not be liable for failing to design out functional dangers. Commentators have concluded, in short, that there is no middle ground between negligence and “ultra-strict” liability, at least in cases involving design defects.
The one exception to this proposition, originally articulated by Deans Page Keeton and John Wade, and since adopted in several jurisdictions, is that knowledge of risks should be imputed to the manufacturer as of the time of production or sale. Thus, in determining whether to impose liability for failure to design out or warn of a danger, a jury might take into account hazards that were unknown, or even unknowable, to the manufacturer when the product was marketed. That the manufacturer could not have discovered these risks in the exercise of reasonable care would be irrelevant; if a hypothetical reasonable manufacturer, aware of these risks, would not have marketed the product or would have warned of the dangers, an injured plaintiff may recover.
This exception uses hindsight to achieve a genuine strict liability in certain cases of generic risks, such as adverse reactions to drugs, dusts, and chemicals. This hindsight approach, however, has not received much policy-oriented justification either by courts or commentators. The mere fact that it created a well-delineated area of strict liability in design and warning cases seemed to suffice. It was inevitable that a need for a firmer rationale would arise.
The recent decision of the New Jersey Supreme Court in Beshada v. ]ohm-Manville Products Corp. attempted to provide such a rationale. The court held that asbestos manufacturers might be liable for lung diseases caused by exposure to asbestos dust at a time when the risks were unknown and undiscoverable, offering three reasons to support this extension of strict tort liability: the allocation of the costs of injuries to the parties best able to bear them; the reduction of risks by increasing incentives for safety research; and the elimination of the need for plaintiffs to prove scientific knowability, a factual determination that is too complex and speculative for jury resolution. The potential problems with each of these reasons will be considered in turn.
The first rationale offered, the notion that manufacturers of defective or unreasonably dangerous products are in a superior position to allocate the costs of product-related injuries, does not really help to answer the question of what makes a product defective or unreasonably dangerous. Nor does it answer the question of which costs should be shifted. Compared with plaintiffs who are injured by products, manufacturers are almost always better able to bear risks by spreading losses through price adjustments and insurance. This rationale would therefore justify imposing liability for harm from risks known as well as unknown, reasonable as well as unreasonable, and ultimately would lead to “ultra-strict” liability. Because it proves too much, this rationale provides only weak justification for a narrower rule of strict liability.
Professor James Henderson has also criticized the risk-spreading rationale on the ground that a hindsight approach would misallocate the costs of liability from products creating risks that were unknown and unknowable at the time of sale. Manufacturers would add this cost to the prices of different, reasonably safe products or to the same products put to different, safe uses. Since the offending products would already have been priced and sold, their liability costs could not be assigned to them. Moreover, once manufacturers discover the danger, the product is removed from the market or redesigned, or appropriate warnings are given, and thus there is no longer any need to assign costs of liability.
Such a result-product prices reflecting costs other than those caused by the product itself-would lead to market distortions and destroy the optimality properties that flow from cost-based pricing in a perfectly competitive market. In a perfectly competitive market, cost minimization and profit maximization for a particular product, and not costs from earlier versions of a particular product, or different products altogether, will determine the price of the product. A manufacturer who tries to pass on these costs will be driven from the market by manufacturers who do not. Professor Henderson’s argument thus squarely poses a paradox: the market distorting effects of misallocation can occur only in a noncompetitive market, where the effects of misallocation are ambiguous. Because of competitive market pressures, unanticipated liability costs are more likely to be paid out of profits, loans, or sources other than price increases.
It is important to distinguish between the allocation that would result from the retroactive application of a hindsight rule and that from the prospective application. The court in Beshada pointed out that application of the rule of strict liability for unknowable risks “will force the price of any particular product to reflect the cost of insuring against the possibility that the product will turn out to be defective.” Thus, the threat of prospective liability would force a proper allocation of product prices. When a court initially adopts a hindsight rule and imposes it retroactively, however, the prices of products marketed years, or, in the case of asbestos, decades, earlier will not bear their own liability costs. In the case of asbestos, this “first shot” problem is enormous. The New Jersey Supreme Court did not ask whether considerations of fairness deriving from justifiable reliance by asbestos manufacturers, or the enormous potential liability to which the industry might be e:…:posed, supported the recognition of a hindsight rule that would operate prospectively only.
For all of these reasons-because it proves too much, because it may or may not apply depending on market conditions, and because its effectiveness depends on whether the application is prospective or retroactive-the risk-spreading rationale raises more questions than it answers and provides only weak support for a rule of strict liability.
The second policy justification offered in Beshada was that a rule of strict liability would spur safety research that might reveal hidden dangers. Put another way, a contrary rule would benefit producers who were unaware of risks and thus would tend to perpetuate ignorance, especially if plaintiffs could not easily establish that a hazard might have been detected in the exercise of due care. Admittedly, if the existence of a hazard were completely unknown at the time of marketing, a manufacturer would be unable to determine how much to spend in order to make the discovery, and there may be no increase in safety research. On the other hand, if a hazard were suspected or were known to exist but its full extent were not known, the incentive for additional investigation could produce some incremental level of safety. In either instance, though, this incentive for safety research would justify a rule of strict liability because the manufacturer can always uncover the known risks better and more cheaply than the potential victim.
It is worth noting that Beshada involved asbestos rather than a drug. Federal regulation prescribes the nature and amount of safety testing that must be done before the marketing of a new medication. In using stimulation of safety research as a rationale for a rule of strict liability for unknown risks, a court would be explicitly or implicitly recognizing a general need for more extensive premarket investigation than presently required by the FDA. This recognition, however, goes far beyond judicial determinations in individual cases that FDA approval of a particular new drug does not preclude a finding of negligence or strict liability. While the safety-incentive rationale is not indefensible, some courts might give it less weight than they would otherwise because of its far-reaching implications.
Another problem with the accident-avoidance rationale is that it leaves open the following question: why should courts impose strict liability upon manufacturers for harm from hazards of unknown scope as an incentive to discover the true scope of the risks, but not apply strict liability as a spur to technological development where at the time of production it was technologically infeasible to eliminate or to reduce risks? There is widespread agreement that in the latter cases, involving the so-called “state of the art” issue, manufacturers will not be liable, absent negligence, for having failed to use today’s safety technology yesterday . It is difficult to distinguish between technology that can detect the gravity of risk and technology that can eliminate or reduce risk, or to conclude that strict liability would act as a spur to the advancement of the former but not of the latter.
The third justification for strict liability offered by the Beshada court is that the litigation process cannot adequately determine scientific knowability. But although the same might be said of the need to decide whether a manufacturer failed to exercise reasonable care in designing a product, courts have not stopped resolving these is sues. The scientific speculation inherent in deciding whether a particular hazard was knowable may produce more uncertainty than a dispute about whether designing out a known danger was feasible; this greater degree of uncertainty might tip the balance in favor of giving at least some weight to this particular rationale for strict liability. The elimination of the need to establish knowability would certainly reduce trial costs, but so would dispensing with the burden of proving lack of due care in design cases.
Since design and warning cases generally are decided by balancing factors that are virtually identical to those used to determine negligence, it is difficult to justify treating unknown or unknowable generic risks as falling within the duty to design or warn but outside the balancing approach. Ultimately, however, a de facto negligence test for all generic risks is unsatisfactory because this standard does not take into account the compelling policy reasons for adopting a strict liability theory. I now turn to those policy reasons, which have been recognized in the context of non generic risks.
B. Justifying Strict Liability for Generic Risks: Are the Policies Underlying Strict Liability in Construction Deject Cases Adequate?
The conceptual treatment of liability for harm from unknowable generic risks as deriving from the manufacturer’s duty to design or to warn creates a discomforting impression: that liability is being imposed for a failure to do the impossible. An alternative approach is to view generic risk through the same lens that, when focused upon the risk of harm from construction defects, has produced a rule of strict liability even when it might have been economically infeasible or technologically impossible to eliminate the hazard. Here the theory does not rest so much on any real or presumed inadequacy in the manufacturing process as on a policy decision to impose liability without fault. Thus, it may be appropriate to inquire whether the bases of strict liability for construction defects support a similar rule for generic risks.
Manufacturers are strictly liable for harm from construction defects even if they could not have eliminated, or discovered, such defects by exercising reasonable care. Held to the standard of their own plans and specifications, manufacturers must answer for imperfections that arise from their production processes. Of the various reasons that have been advanced to justify this rule of strict liability in construction defect cases, three seem worthy of discussion in the context of generic risks: accident avoidance, loss spreading, and the satisfaction of justifiable consumer expectations.
Whether strict liability will actually foster accident avoidance has been seriously questioned. It has been argued that producers will avoid only those accidents worth avoiding-if it is cheaper to let an accident happen and to pay the resulting liability costs, the profit maximizing manufacturer will follow that course. Thus, if testing and quality-control procedures would cost more than projected liability costs, a rule of strict liability would not encourage manufacturers to adopt procedures to prevent accidents.
This argument, however, is not entirely persuasive. A manufacturer bound by negligence principles might foresee escaping some liability costs that should attach when it does not exercise due care. The difficulties of proving fault might be too great for injured plaintiffs in certain kinds of cases, or economic constraints might force plaintiffs to accept unfavorable settlements. Anticipating these lower liability costs, manufacturers might spend less on accident prevention. By reducing plaintiffs’ burdens, a strict liability rule might well encourage manufacturers to increase safety expenditures to the level they might reach under a negligence system that functioned optimally.
The adoption of a rule of strict liability in cases where a manufacturer knew a risk existed but did not know its full extent also might increase safety by providing an incentive to perform additional investigations. Indeed, assuming that manufacturers foresee that, under negligence principles, not every injured plaintiff will recover full damages for harm from a particular design feature or warning, the application of strict liability to all generic hazards, known and unknown, will increase the prospect of full recovery, encouraging safety expenditures and accident avoidance. This increase in safety enhancement standing alone, however, is probably insufficient to justify liability without fault in these cases.
The “loss-spreading” rationale rests on the manufacturers’ ability to use insurance to spread the costs of harm caused by construction defects more efficiently and more easily than product victims can. Construction defects are easily insurable for two reasons: the number of claims likely to arise from such defects is fairly predictable, and this number is likely to be relatively small in comparison with the total number of products placed into the market. Insurance against these risks, therefore, is readily available because the costs are predictable and the harm to be insured against normally will remain within modest bounds. The number of known generic risks likely to occur-ranging from adverse drug reactions for which warnings have been given to automobile accidents -can also be predicted with some certainty. Rough estimates can even be made about risks whose presence is known but whose extent cannot be calculated. The only type of hazard that would not permit even a guess would be the unknown and undiscoverable danger.
In the case of generic risks, however, the other aspect of insurability-a comparatively small number of risks-is absent. Unlike construction defects that affect only a small percentage of users, every generic risk will endanger every user of the product. Thus, the amount of damage attributable to generic product risks could be enormous, even if recoveries are reduced to take into account the comparative responsibilities of plaintiffs, third persons, and other enterprises that might appropriately share the losses. One might argue, then, that loss spreading makes sense only in the context of construction defects, where the relatively modest costs can be more easily absorbed by the manufacturer.
An intermediate position might hold manufacturers strictly liable for unavoidable hazards, such as adverse reactions to toxic products. but not for designed-in, functional dangers, such as the speed of an automobile. This compromise position, however, has several problems. As a practical matter, it is difficult to base a rule of strict liability on degrees of potential damage: the notion that the more harm a defendant may cause the less likely it is that liability will attach strikes a somewhat perverse chord. Moreover, the focus on the quantity of loss may well be misguided. If the purpose of loss spreading is to deflect the economic impact of product-related harm away from those who may not be able to absorb it, perhaps the focus should be on the victims’ capacity to pay for their own injuries, and not on the aggregate cost of all such injuries.
Consumers’ ability to foresee product risks is relevant to a determination of their ability to insure themselves against those risks, and thus to a determination of their capacity to absorb the cost of their own injuries. The policy of satisfying justifiable consumer expectations may shed light on this issue of cost absorption in particular and on the appropriateness of strict liability for generic product risks in general.
The notion that manufacturers should be strictly liable for harm from product frustration is rooted in the doctrine of implied warranty of merchantability, which holds goods to the standard of reasonable fitness for their intended use. Products placed into the stream of commerce carry with them a representation of safety, the scope of which is determined by what the ordinary consumer would expect of those products. This representation of safety underlies the consumer contemplation test set out in comment i of the Restatement.
It is important to distinguish between two uses of consumer expectations: the goal of meeting justifiable consumer expectations as a policy behind strict tort liability, and the use of consumer expectations as a criterion for deciding whether strict liability should apply in a particular instance. The former derives from the conviction that, as a matter of fairness, consumers should be entitled to rely on the representation of safety made by the seller of a product and by any information accompanying the product. Consumers depend on the manufacturer to provide goods that will meet these implied representations so that they can make rational judgments affecting their own well-being. The imposition of strict liability will encourage producers to satisfy these consumer expectations, will permit consumers to act on the assumption that expectations will be met, and will enable consumers to survive the economic hardship of unexpected losses.
When using consumer expectations as a criterion for applying strict liability, the critical task is to determine which consumer expectations are justifiable. The rule in construction defect cases suggests that courts have found such defects to lie outside the ambit of consumer contemplation; consumers, therefore, may justifiably expect products to be free of construction flaws, and manufacturers will be held strictly liable for all such flaws: known, unknown, and unknowable. In design defect cases, however, courts apply what amounts to a negligence test and say in effect that consumers justifiably may expect only that due care, measured as of the time of manufacture, will be exercised with regard to design and warning decisions.
Is this distinction tenable? Given what the average person undoubtedly knows about product quality (especially in light of the publicity given to recalls of automobiles and other household products), all types of risk-creating flaws, both in construction and design, are arguably within the contemplation of ordinary consumers. In some cases, awareness of a vague possibility that some defect might lurk somewhere within a product ought not to establish the risk as within the consumers’ contemplation. The wide range of potential flaws, especially in complex items such as automobiles and workplace machinery, and the varying degrees of potential risk associated with such flaws, renders a general awareness practically useless to the consumer. Moreover, the marketing image of a product may dim an already faint awareness of the risk. A rule of strict liability for construction defects, then, reflects a justifiable judicial determination that consumers merit protection under a standard requiring goods to be completely free of such defects.
A practical reason for limiting justifiable consumer expectations to the exercise of reasonable care in the design of products is that there is no other workable standard by which courts may determine whether a product is unreasonably dangerous. Consumers usually are unable to form an expectation about the extent to which design defects will be eliminated: it is not a matter of expecting one unit of a particular product to be as good as the next. Therefore, the best that consumers can justifiably expect in the design defect context is that manufacturers will use technologically and economically feasible methods to reduce or eliminate foreseeable risks.
The policy of satisfying justifiable consumer expectations also dictates the refusal to impose strict liability for harm from known generic risks. The ordinary consumer appreciates the dangers posed by a speeding automobile or a sharp knife, and would therefore have no cause to believe that a manufacturer would do more than use due care to reduce these hazards. Contemporary smokers know of the risk of cancer from cigarettes. The presence of warnings on the label of prescription drugs makes physicians, acting on their patients’ behalves, aware of the relevant risks. In each of these cases, consumers can make a rational judgment about the scope of the hazard and act accordingly.
But if the danger or its full dimensions do not become evident until after the plaintiff has been exposed to the product, the consumer-contemplation policy supports the imposition of strict liability. The product has inflicted an unpleasant surprise. Although the manufacturer could not have discovered the danger or its extent, the marketing of the product misled the consumer with an implied representation of safety that was not met and thus deprived the consumer of the opportunity to evaluate the risk and to decide whether to accept it. Under this new view of consumer e1:pectations, a product posing an unknown or unknowable generic hazard would stand on the same footing as a product with a construction flaw: each product would be considered unreasonably dangerous for purposes of strict liability because it frustrated justifiable consumer e1:pectations recognized by the law.
The need to integrate liability for product-related harm to nonconsumers into a scheme structured around consumer expectations raises a conceptual problem. Professor Gary Schwartz has noted that a third-party beneficiary theory can preserve the viability of the consumer-expectations test in instances where the consumer could reasonably be deemed to have contemplated the conferral of accident-avoidance benefits upon others. The extension of the implied warranty of merchantability, which under the Uniform Commercial Code protects anyone “who may reasonably be expected to use, consume or be affected by the goods,” lends support to this argument by analogy. But it would be stretching things beyond the breaking point to assume that a consumer intends to protect bystanders, especially those who are total strangers. As a practical matter, this problem will be limited to construction defect cases: the de facto negligence test used to determine liability in design defect cases applies equally well to consumers and bystanders; and unknown generic risks will rarely endanger anyone other than a product user. These limitations, however, do not eliminate the theoretical hurdle.
One answer is simply to recognize that the policy of satisfying justifiable expectations supports the imposition of strict liability only on behalf of consumers and their intended beneficiaries. To hold manufacturers liable without fault for harm to bystanders would then require a separate, independent rationale. A second, and perhaps preferable, solution lies in a reassessment of the consumer-contemplation policy. Its roots, as has been noted, go back to the doctrine of implied warranty of merchantability, the primary concern of which was the adjustment of the rights of parties to commercial transactions. Although courts fashioning tort doctrine may legitimately borrow from sales law, they need not feel fettered by sales law constraints. Where the same policy goals would be applicable to nonconsumers, it might be logical to extend strict liability protection beyond the purchaser. Thus, the user of a product personally relies upon the implied representations of safety inherent in the product. Certain bystanders may also entertain similar expectations that a product will not injure them. This approach would require courts to differentiate between two classes of bystanders: the first is exemplified by a pedestrian injured when an automobile goes out of control because of a construction defect; the second by the person harmed while asleep at home by an airplane that crashed because of a flaw in its assembly. In the latter case, the victim had no expectation generated or frustrated by the product. The falling airplane was like a falling meteorite-completely unexpected-an event for which there is no tort remedy. Hence the consumer-contemplation rationale, expanded to take into account the actual expectations of users and bystanders, would not support recovery by such victims under strict liability.
This Article has proposed a conceptual framework for determining when to apply strict liability to generic product risks. On the twentieth anniversary of the first decision to hold product manufacturers strictly liable in tort, the parameters of the doctrine remain in flux. Federal legislation threatens to restrict the doctrine to harm from non generic risks. Conflicts and uncertainties in the common law of product liability as it has evolved in the states have been cited as a major justification for federal action.
The case for salvaging some remnant of strict liability within the area of generic product risks is not an easy one. The use of a policy based analysis, however, makes it possible to link the accepted view that the rule should apply to construction defects to the admittedly controversial proposition that harm from unknown or unknowable generic risks should be compensated in the same fashion. The advantage of this approach is that it provides a coherent, principled basis for excluding other kinds of generic product risks from a rule of strict tort liability. Both the satisfaction of justifiable expectations on the part of product victims and the achievement of modest advances in safety justify the application of strict liability to harm from unknowable generic hazards.
Neither section 402A and comment k, interpreted as denying strict liability for unknowable generic risks, nor Beshada, forthrightly permitting recovery in such cases, presents a satisfactory resolution to the problem. The proposed federal Product Liability Act uncritically accepts comment k, while Beshada has provoked an outpouring of criticism. The tide at the moment apparently is running against strict liability in generic-risk cases. But the last words have not yet been spoken.
Joseph A. Page, 1983.
Click to download the full paper.