www.lawrence.edu/fast/boardmaw/After_MacPherson.html reposted Thursday, March 15, 2007

After MacPherson: Comments on Strict Liability in Torts

(c)  Boardman: draft; fair use permitted

 

  After MacPherson until the 1950s, the tort system in America was recognizably traditional, requiring at least the negligence of a manufacturer to permit his liability for injuries caused by his product. Gradually through the 1950s, the liability of a negligent manufacturer had been extended from the ultimate purchaser of a product to his family, guests, and bystanders.

  Although the move in the second half of the twentieth century toward strict manufacturers' liability to the consumer was often articulated in court cases as simply a further extension of the fault system then in place in order to repair some few remaining technical flaws, it eventually resulted in the imposition of a new system of compensation using an altogether novel rationale. At mid-century, a person injured by a defective product would have had the burden of proving that his "injury has been caused by a defect in the product," "that the defect existed when the product left the hands of the manufacturer," and "that the defect was there because of the manufacturer's negligence." Minimizing the practical consequences of shifting the basis of tort from this negligence standard to a strict liability regimen, Prosser adds, "This [third task] is by far the easiest of the three, and it is one in which the plaintiff almost never fails. . . . It is true that [the plaintiff] has the burden of proof on the issue of negligence. . . . But in every jurisdiction, he is aided by the doctrine of res ipsa loquitur, or by its practical equivalent. In all jurisdictions this at least gives rise to a permissible inference of the defendant's negligence, which gets the plaintiff to the jury. And in cases against manufacturers, once the cause of the harm is laid at their doorstep, a jury verdict for the defendant on the negligence issue is virtually unknown."1 The standard for establishing a manufacturer's negligence is that the defendant should have foreseen the injury and was not reasonable in failing to prevent it. According to the traditional rules of negligence as articulated by O.W. Holmes, "sound policy lets losses lie where they fall, except where a special reason can be shown for interference."2 "Apart from a few exceptional situations such special reason was the tortfeasor's fault . . . A person at fault was not only guilty of a legal wrong but an ethical wrong."3

  Particularly after the 1950s, the policy of basing liability on fault came to be looked upon by academics and judges in America as superstitious. Skepticism about the suitability of the concept of "fault" arose in the case of automobile accidents, where the probability is high that every driver will have a number of lapses of attention during any brief time span;4 if misfortune so arranges the circumstances, anyone could be "at fault" for an ensuing accident. (That together with the large costs of litigation which the tort system added to the costs of compensating the victims spurred the call for non-fault auto accident insurance schemes.) Moreover, every product or activity foreseeably carries some risk of injury. In the instance of a manufactured product, it may be statistically likely, and therefore foreseeable, that a certain number of serious injuries will result unless certain steps are taken; yet even so, failure to have taken those steps might not constitute negligence, for it may be "reasonable" for the manufacturer not to have, e.g., installed prohibitively expensive modifications or not to have ceased production altogether in order to prevent the anticipated injuries. In applying the standard of care appropriate for judging the reasonableness of a defendant's conduct, "the courts frequently balance the degree of foreseeability or risk of harm against the costs to the defendant of avoiding the harm and the wider benefits foregone if a certain activity cannot be carried on."5 According to Gary Schwartz, "[T]he process of balancing the magnitude of the risk against the cost of risk prevention has been embedded in negligence law since the nineteenth century, and was rendered official by the First Restatement of Torts and Learned Hand's opinion in United States v. Carroll Towing Co. Indeed, many scholars have interpreted the products liability risk-benefit standard precisely as a very modern manifestation of very traditional negligence reasoning."6

  But perhaps a product or activity cannot be made safer in a cost-effective way, thus removing the issue of negligence; yet it may be that other, safer products and activities could be substituted for it. If building bridges over wide rivers inevitably leads to the deaths of construction workers, and the activity cannot be made safer without making bridges so expensive that few would be able to afford to use them, society might substitute a safer alternative, e.g., ferry service. Supposing that such a substitution might be rational, the traditional law of negligence does little to encourage the substitution.

  And imagine that, in an occasional case, a manufacturer produces a dangerously defective product in spite of his due care (e.g., as in the UK case of Daniels7): should the cost of injury fall on the faultless consumer or on the equally faultless manufacturer? The traditional rules of negligence could not in such an instance place liability on the manufacturer; nevertheless, he is at least better placed than the victim to anticipate and prevent such injuries—he is "the least cost avoider." (Compare the case to that of a drunk driver who causes an accident: at the time of the accident, he might have been unable to prevent it; but at an earlier time, when he began to drink, he might then have anticipated and taken steps to minimize the risks of continuing to drink and subsequently driving.) It seemed fairer and more reasonable to place these costs on the manufacturer; historically the move to strict manufacturers' liability was reached by incremental steps attempting to cure the injustices resulting from the negligence rule as modified by the doctrine of implied warranty.8 This led academics, and eventually judges, to a newer way of conceiving torts—as an instrument to reduce the incidence and costs of accidents by contriving to place liability on that party who was in the best position to avoid costs of injury. In general, that meant assigning strict liability to manufacturers for injuries caused by the proper use of their products.9

  But before we can discuss those changes, we will need to understand the economists' notion of "externality," and also notice how the advent of the insurance industry made it feasible to place specific monetary values on the risks of injury. When some of the costs or benefits associated with a person's choice are ones which the person making the choice does not himself pay or will not himself receive, those "orphaned" costs or benefits, not being internal to his deliberations, are "externalities" from his point of view. Suppose that a hobbyist in a residential neighborhood is choosing between two brands of power saw: although they perform their tasks equally well, that saw which is "whisper quiet" costs $30 more than the noisy saw; but because the hobbyist is willing to wear earplugs when he works, and these cost only $5, he chooses the noisier machine as having the better cost to benefit ratio—a net savings to him of $25. But suppose, further, that the noisy saw bothers his neighbors so much that they are induced to sound-proof their houses at a total neighborhood cost of $3,000. This further cost is external to the hobbyist's calculations, since he does not pay it—his disgruntled neighbors do. If we were to consider the total costs and benefits—to everyone—associated with his choice of saw, we would see—from a God's eye view—that the hobbyist's choice is unfortunate and inefficient: rather than saving $25 dollars, his choice actually wasted $2,975 (of other people's money). If we could somehow manage to get a single entity (perhaps a neighborhood improvement commission) to pay all the costs and receive all the benefits of such decisions, we could then expect it to make the most efficient choice: for since all of the costs and benefits would then be internal to its deliberations, it would purchase the "whisper quiet" saw; the extra $30 in initial purchase price would save the neighborhood $3,000. (We might also internalize such external costs if we could create a mechanism which would compensate the hobbyist for his extra cost in choosing the quiet saw from the savings to his neighbors which that choice would bring about; we would need, incidentally, to block attempts at extortion (or, what Polinsky calls "strategic behavior")10 which would seem possible—a person who had no genuine desire to use a noisy saw might threaten to buy one unless he were bribed by his neighbors not to do so. This is also called, "moral hazard.")

  I would conjecture that before the invention of insurance policies, no one really could quantify the risks of injury associated with various enterprises: torts law, with its liability rules, served to determine which of several parties would bear the losses due to injury; but there was no precise way to compare the risks of injury associated with different activities. The use of power saws, to use an anachronistic example, might result in more injuries than the use of hand saws, but so long as those injuries were not caused by the manufacturer's negligence, the greater cost of those injuries would not be reflected in the saw's price. The purchaser might become aware that power saws seemed to be associated with a greater incidence of injury than a hand saw, but he had no rational means for weighing the resulting costs since there was no way to quantify them. But with the advent of the insurance industry, and perhaps modern medicine, it became useful to collect accident data and feasible to place specific monetary figures on the costs of the risks of injury: actuarial tables combined the rate of accidents with the costs of those accidents, and premiums based on those tables varied with the risks. As a result, a purchaser of a saw could now bring into his deliberations about whether to buy an power saw or a hand saw not only the differing initial costs of the two kinds of saw, but also the differing insurance premiums for using them: insurance mechanisms allowed the costs of the risks of injury to him to be made internal to his deliberations. As a result of the ubiquity of insurance in the twentieth century tort law came to be seen not merely as a means of determining which party will bear the losses due to injury, but as a means of assigning the costs of the risks of injury; the question of liability seemed to change from "who will pay for the losses?" to "who will pay for the premiums which insure the expected losses?"

  Suppose, then, we consider two activities: one company is engaged in running a ferry service across large rivers, while a second is engaged in building enormous bridges across rivers of similar size. And let's suppose that the level of activity of these businesses is comparable, and that actuarial tables show that far more people will die as a as a result of bridge construction than as a result of instituting ferry service. Here, then, is a difference between the risk of death associated with different means of river transport. We might worry about whether the costs of the higher fatality rates associated with the bridge are borne, not by the bridge enterprise, but by other members of the public. If that were so, the costs to municipalities of building bridges would not include the costs of the higher death toll; and so, when these enterprises came to compete on the basis of price, one of the costs of using the bridges would not be passed on to the bridge-using public. And as a result, more people would use the bridge than was "optimal," since some of its real costs would be borne by others than the consumers. Perhaps if those costs were made internal to the choices of the consumer, if the bridge-tolls reflected the costs of supporting destitute families whose bread-winners had perished in bridge construction accidents, consumers would find it more cost-effective to patronize ferries than bridges. Then bridge enterprises would be forced either to find a way to reduce the costs of accidents in order to remain competitive or go out of business. On the other hand, if the workers' families continued to subsidize the cost of bridge-traffic, the bridge tolls would understate the full costs of using the bridge, leading to an inefficient preference for using bridges. If this situation held as described, we could talk about the costs associated with fatalities in bridge construction as an externality: it is a genuine cost of the use of bridges—since people can use bridges only after they are constructed (and, so, only after some workers have perished)—but it would not have been included in the price which people pay to use the bridges. As a result, the consumers' toll for using bridges would make their use appear to be cheaper than it really is: the cost borne by the workers' families would be one of the costs of using bridges external to the costs of the consumer. One might suppose that if, instead, such costs were placed on the activities which generate them, it might turn out that the ferry rates would be substantially less than the bridge-tolls, since the latter would now include insurance premiums for the expected higher fatality rate associated with the construction of bridges.

  There certainly exist examples of economic externalities; but even if my example takes place before 1960, it might not really turn out to be one. For even under the laws as they were then, at least a part of the costs of bridge construction fatalities were borne by people who paid the tolls. Because the fatality rate of bridge construction was greater than for running a ferry, its workman compensation costs were higher; and it had to pay its workers higher wages than prevailed in safer industries.11 And these costs would be passed on to the bridge-users through the tolls.

  But, then, consider automobiles as one form of transportation. Were the costs of auto accidents external to automobile manufacturing? Well, at first it might seem so: each individual purchases his own liability policy rather than having it provided by the automobile manufacturer. But upon a second look it seems that the costs of auto accidents were indeed internalized. If the automobile industry had insured each customer's driving, it would have passed on those additional costs as part of the price for the automobile; in purchasing the automobile, the consumer would automatically have paid for the machine and also for the insurance premium.12 Yet as it worked in 1960, the consumer still had to pay for each of these things; the fact that he paid for them separately, in two transactions, is irrelevant. To use an automobile for one's transportation required that someone pay the costs of constructing the automobile and also that someone pay the costs of insuring its use. On the other hand, if one used other means of transportation, he would not have to purchase an automobile insurance policy but an alternative policy for his alternative form of transportation; if the costs of automobile accidents were higher than the costs of accidents for the alternative transportation, then the auto insurance would be higher than the other insurance, raising the cost of using automobiles relative to the cost of using the alternative means of transportation. In either scheme, the automobile purchaser does in fact pay the costs of automobile accidents; so no other party is subsidizing these costs.13 And as the cost of automobile insurance for a given model of car increased, fewer consumers would purchase the model made by that manufacturer. So, if a manufacturer could make his car safer and thereby lower the user's cost of insurance, he would be able to sell more cars. Moreover, tort law after MacPherson imposes the costs of foreseeably catastrophic accidents resulting from the manufacturer's negligence upon the manufacturer; and this will also provide an incentive to manufacturers to design and manufacturer their products to reduce these risks of injury. And so, it is difficult to understand what was supposed to be the problem with liability law prior to the 1960s; looking back, it is not obvious that they encouraged inefficiencies.14

  Why should one think it necessary to make the automobile manufacturer insure his customers rather than letting them insure themselves? In order to understand the argument, we must talk about the "causes" of accidents in a very technical way. We will speak of something as a cause of an accident when the accident would not have happened in its absence: a "but-for cause."15 Suppose I get drunk at a party and, while driving home, injure someone. It is obvious that my having drunk seven beers just prior to driving, and consequently having impaired my skill in driving, was a cause of the accident. But suppose that Detroit is able to manufacture an automobile which cannot be operated by an intoxicated driver: a device monitors the driver's breath for alcohol and will not allow the car to be operated if he is intoxicated.16 In that case, one cause of my accident was the fact that the breath sampling device was not installed in my car. And, suppose, too, that Detroit is able to manufacture an automobile which continuously monitors on-coming traffic with a computer which will guide the car into evasive action when an impending impact is sensed: in that case, one of the causes of my accident was that this monitoring system was not installed in my automobile.17 You will say that it is cheaper for me to refrain from driving while drunk than to install such equipment; you are correct, of course, but we need to notice that this regime of voluntary self-monitoring is not a particularly reliable system for the entire population of automobile users. While the impact-sensing automobile will be more expensive than self-monitoring, it might be tremendously more effective in preventing expensive accidents. Suppose we plot the cost and effectiveness of each of these means of avoiding automobile accidents against the costs of accidents that would be avoided if every automobile relied on it. We may find that the very cheap method of abstinence will reduce accident costs by a comparatively paltry amount, while the more expensive impact monitoring device would pay for itself in the accident costs it avoids.

  But who is going to weigh the costs and benefits of these alternative methods and choose the most cost-effective method? When Ralph considers whether to purchase an automobile having such an expensive computer system, he doesn't really have the incentive to weigh all the benefits of the system against its costs, since many of these benefits will not benefit Ralph. Consider the case where you run through a red light and as a result Ralph's car slams into yours; here is an instance where an accident-avoidance system in Ralph's automobile would have saved the costs of an accident. Yet Ralph wouldn't be willing to trade his costs (for the device) for your benefit (your savings in accident costs); for even without the device Ralph would not have been liable for the costs of this accident since he was not at fault. In this case, the imagined benefit to you is an externality to Ralph. On the other hand, because both you and Ralph must purchase your automobiles from a manufacturer, it would seem that the manufacturer would be in the best position to judge the possibilities and effectiveness of accident prevention technology and to judge whether such a device is cost effective and, by including the cost of the device in the cost of each automobile, to make such costs internal to the cost of buying an automobile:18 if the law were to impose strict liability on the manufacturer, there would be a single entity with the incentive to find and implement the most cost-effective means of reducing accident costs. For since all the costs and benefits would go through the manufacturer's calculator, the more effective the methods he devises to prevent costly accidents, the less he needs to charge for his automobile to recoup his investment in those methods; if he can save the costs of your injury, he can reduce the price of the computerized automobile to Ralph. Making manufacturers strictly liable for accidents caused by their products is seen by some theorists as a device to make them the insurers of consumers, which would provide a mechanism for internalizing a number of the costs and benefits of product accident prevention.19 If the courts were to impose such strict liability, it is theorized, greater efficiency in accident prevention would be achieved.20

  Consider this recent attempt, by a Judge of the U.S. Court of Appeals for the D.C. Circuit, to articulate the rationale behind strict liability in tort:21

  Deterrence theory justifies tort law as a cure for market failures related to risky products and services. Negligence liability gives the providers of goods and services an incentive to take all cost-justified precautions in the design and manufacture of what they sell. Strict liability not only creates those incentives, but also brings about the economically correct levels of production: suppliers' prices will reflect the relative riskiness of their goods, making consumers adjust their purchases to the level they would have demanded if they were fully informed about the products' residual hazards. Thus, under either a negligence or a strict liability system, soft drink and beer manufacturers will take cost-justified steps to prevent bottles and cans from exploding, cutting drinkers' lips, and the like. Under strict liability, if bottles are more dangerous than cans even after these safety measures have been taken, the price differential between the two will reflect the risks and thus spur consumers to adjust their consumption decisions accordingly. The allocation of resources as between bottles and cans will be optimal.

  The validity of deterrence theory, as applied to claims by people who contract to buy the goods and services that cause their injuries, depends on an assumption ... that consumers will rationally invest little time or effort in informing themselves [about low-probability risks]. ... Thus, but for tort liability, producers would have inadequate incentives to compete either in reducing risk or in offering warranties. If consumer error runs systematically toward optimism, as the architects of strict liability assumed, the distortion will allow risky products to enjoy a more-than-optimal share of the market. To correct this market failure, courts impose liability regardless of fault, even in the face of contractual waiver.

  The rationale behind replacing negligence liability with strict liability can best be explored from Ronald H. Coase's "The Problem of Social Costs."22 Coase's discussion implies that where market transactions are costless, it should not matter whether tort liability for dangerous products is imposed on the manufacturer or whether the costs of such injuries are left on the individual, since the efficient allocation of resources (e.g., into making a dangerous product "optimally" safer) would come about in either case from the bargaining of these rational actors. In a simple world—one in which there are no "transaction costs"—when the liability is imposed on the manufacturer, he will anticipate that he will be liable in similar suits unless the product is made safer; and he will, accordingly, make an investment in safety so long as it will be less than the likely tort payouts which it would avoid. The manufacturer will then pass along to the consumer the added cost of those safety investments (together with those costs of liability which cannot be avoided at a cheaper investment in safety). Notice that the manufacturer, in this simple world, has no incentive to refuse to invest in the prevention of accidents since his net costs are less when he has invested in safety. Thus, in the simple world, placing liability for accident costs on the manufacturer will reduce those costs to the "optimum" level.23

  But now suppose that there is no tort liability in this simple world, so that the costs of accidents remain on the consumer. In that case, the rational consumer will bargain with the manufacturer to induce him to invest in product safety an amount of money less than the accident costs it would avoid. Through a "bribe" given to the manufacturer, the consumer reduces his accident costs by paying an increased price for a safer product; since the increase in price is smaller than the accident costs he would have incurred otherwise, the consumer is better off. Thus, in the simple world, placing liability for accident costs on the consumer rather than on the manufacturer will also reduce those costs to the optimum level. So what Coase implies is that, in this simple world, the same amount of money would be invested in safety no matter whether the initial costs of accidents were imposed on the manufacturer through tort liability or on the consumer upon whom they originally fell. In a simple world, if the costs of accidents can be reduced through cost-effective investment in safety, rational actors will bring about those investments, and the consumer will eventually pay the higher price for an optimally safer product.24 But we must remember that these implications hold only where there are no "transaction costs."

  The economist's "transaction costs" will include such things as the cost of acquiring information, the cost of bargaining, the cost of adjudicating law cases, and the cost of government regulation of manufacturers. As Coase recognized, while these may be zero in a simple world, they are not generally zero in our world of manufactured products. Suppose that accident costs remain on the consumer on whom they originally fall: although he has an incentive to bargain with the manufacturer for a safer but more expensive product (or to patronize the manufacturer's competitor who would have an incentive to make a safer product), an individual consumer would find that acquiring such information about the relative safety of products and cost-effective ways to improve their safety would be more expensive than the likely savings he would gain from a safer product.25 If the various consumers could get together and share the cost of gathering information about the design and costs of an "optimally" safe product, then any one consumer's cost of information would not be prohibitive; but now the cost of such bargaining with other consumers in order to form their alliance would itself be prohibitive. Indeed, because of the large transaction costs of safety information, manufacturers could raise their prices and advertise their unimproved products as being safer than their competitors' without the consumer's being able, in a cost-effective way, to verify such claims.26 And there are further problems in the more complicated world. Different potential consumers of a product will place differing values on a safer product: because of their different incomes, preventing an accident which will deprive them of a year's salary will be worth much less to a poorer customer than to a richer customer. So even if they could form an alliance to gain information about the methods of accident prevention, when their alliance began bargaining with the manufacturer about how much the group's members were willing to pay for accident-saving improvements, they would not agree with each other about the increase in product cost which would be warranted by its increased safety. (For rich and poor, "optimal safety" will be different, depending on what each would rationally afford to pay for such safety.) Of course, in a simple world, the manufacturer could bargain individually with each consumer, selling the product to a richer person for a higher price than to a poorer person; but, alas, in the complicated world, it would be quite expensive for a manufacturer to acquire reliable information about each of his customers and then to bargain individually with each.27

  On the other hand it might seem, in the more complicated world, that placing strict liability on the manufacturer would succeed in bringing about the efficient levels of investment in safety and in insurance for those accidents which could not efficiently be prevented. The manufacturer will have an incentive to avoid the cost of liability by investments in safety which will be less than his costs of liability judgments; and he will recoup those investments by raising the price of the product. Some accident costs will, however, have to be absorbed by the manufacturer since some accidents cannot be efficiently avoided by investments in changing the design and method of manufacturing the product or in giving warnings against those risks which cannot be prevented cheaply by the manufacturer; these residual costs will be spread among the consumers by increasing the per-unit price of the product. So, at first glance, it would appear that the inefficiencies in our complicated world which are brought about by high transaction costs would be solved by imposing strict liability on the manufacturers: manufacturers will be induced to invest the right amounts in accident avoidance, and consumers—who, if left to their own devices, might not possess the information which would lead them to insure themselves against the costs of inevitable accidents—will be forced to pay just the right amount for this insurance (i.e., the total cost of unavoidable accidents divided by the number of purchasers). At first glance, then, it would appear that the problems have been solved.

  Nevertheless, if the accident-costs are placed on the manufacturer through tort liability, that also will not bring about an efficient state of affairs. For after the manufacturer recoups his investments in safety and in insurance by raising the price of the product, the product may now be too expensive for many of his former customers. This result would be efficient if the value of the product to the consumer were accurately reflected in its new price; the economic advantage to internalizing externalities is that consumers will then be presented with the true costs of a product and will no longer make inefficient choices based upon the fact that some actual costs are hidden.28 But we must remember something we mentioned earlier: because of their differing incomes (and other differences), the various consumers will not place the same value on improvements in safety and on insurance for residual accidents; in an efficient arrangement, they would not pay the same amount for accident avoidance. In spreading these total costs by dividing them by the number of purchasers, the consumer will pay the average cost of accident avoidance, not the cost to him of avoiding those accidents he would choose to avoid.29 So the poorer customers would subsidize the accident avoidance costs of the richer. And because those who use the product frequently will tend to have more accidents and therefore value accident avoidance more than consumers who use the product less frequently (since the probability of an accident from the product increases with the frequency of its use), the frequent users will be subsidized by the infrequent users. Some risks are unavoidable, and imposing strict liability on manufacturers may slow product innovation or force useful products off the market altogether.30 Finally, imposing strict liability on manufacturers will tend to reduce the incentives on individual consumers to prevent accidents voluntarily and cheaply by simply being careful.31 Recall that we have been talking about the "causes" of accidents in the sense of "but-for" causes; many accidents which would not have happened but for a particular product design would also not have happened but for a particular choice made by the consumer in using the product; and it is often the case that the accident could have been prevented more cheaply by the consumer's taking care in using the product than by the manufacturer's altering the product's design. Yet in placing strict liability on the manufacturer, we seem to focus on the more expensive ways to prevent accidents to the exclusion of the cheapest ways.32

  Prior to 1970, the effect of the MacPherson rule was to place liability on manufacturers for those accidents which they could have foreseen to be very serious, which they could have avoided efficiently, and for which the design or manufacture of the product was the direct cause (in contrast to being a mere "but for" cause); while other accident costs would remain on the consumer. That created an incentive for manufacturers to avoid the most serious accidents which they could directly prevent, leaving consumers to insure themselves against the remaining accidents, and creating an incentive for consumers to try to avoid them.33 Yet in the 1970s, the idea behind replacing negligence liability with strict liability seemed wonderfully ingenious: rather than setting forth standards to which citizens are expected to conform their behavior, the proposed strict liability device would manipulate citizens and product manufacturers to make efficient choices. It seemed wise to repudiate the old fashioned, common sense standards of behavior in order to look forward to the brave new judicial policy of economic efficiency.

  A number of critics of strict liability frame the issue as though the main question is whether insurance for the consumer should be purchased by the consumer or by the manufacturer34 If the consumer is rational and well-informed, then he can more efficiently purchase the insurance coverage he desires, deciding for himself what the optimum cost of accidents is, and avoiding subsidizing other, often wealthier, consumers by purchasing the insurance for himself. In a world in which the consumer knows that he must purchase insurance against injuries caused by defects in the products which he purchases new, and where he has the information (in a form usable by lay persons) to assess the risks of such defects, purchasing insurance coverage for oneself from the market would be efficient. (And, evidently, every citizen would have to be informed of the need to purchase insurance against injury caused by someone else's new products.) But in our world, would one suspect that he must insure himself against injury from new and, often, expensive products? For what little it is worth, as a consumer since the 1950s, it would never have occurred to me that I might have to bear the costs of injuries caused by defective products. When manufacturers of automobiles and appliances routinely present information about the possibility of buying the product on credit, even in the fifties none of which I was aware presented information about the need to insure oneself against defective products; and when manufacturers and other companies used, standardly, to specify a waiver of all liability in tiny type on the back of sales agreements and tickets, they did not add a warning that the consumer should seek out insurance on his own. So at the present time, the consumer would not know that he was expected to purchase insurance for himself for the new products and services he purchased. In addition, the consumer is ordinarily in a poor position to know about the potential risks of products' being defective and to determine whether it is sufficiently serious for him to reopen negotiations on his insurance policies. When soft-drinks were typically sold in glass bottles, the bottles from time to time exploded and caused injury to consumers; how many consumers would think to insist on a clause in their insurance to cover such a possibility?

  As we have seen, there are difficulties in our complicated world in attempting to use civil liability law as an engine to internalize accident costs. Imagine that there is a certain sort of woodworking saw which requires substantial experience and skill in order to be used safely. When introduced to the market, it has been accompanied by warnings that it is intended for professional use, and its hazards for unskilled consumers have been pointed out. Nevertheless, a certain number of laymen (who, foreseeably, imprudently discount the warnings) purchase the saw each year, and a certain number of serious injuries result—primarily from the laymen group of users. Under a scheme by which strict liability on manufacturers imposes the requirement that they insure their customers, the saw will require very high premiums; for it would be prohibitively expensive for the manufacturer to negotiate individual prices of the product with individual customers, lowering the cost of the insurance premium for the skilled consumers, and raising the cost for unskilled consumers; but, as a result of Henningsen v. Bloomfield Motors,35 a standard form contract (which would avoid the transaction costs of negotiating individually with customers) in which duffers waived the manufacturer's liability would probably not hold up in court. One might propose that such practical difficulties in face-to-face negotiations over individual contracts might be overcome by requiring that standard-form contracts must be "reasonable" and then specifying their details through subsequent litigation. But it is worth remembering that the economist's "transaction costs" include legal and court costs, and that these costs are considerable. "Estimates of the share of the insurance premium dollar that is actually received by victims vary widely, ranging from a low of 18 cents on the dollar to a high of 60 cents. ... This implies transaction costs ranging from a low of 40% to a high of 82%. Huber estimates the administrative and legal costs at 60% of the amounts spent on medical malpractice and products liability insurance, and 50% of traffic-accident liability."36 As a result, this very useful saw might be forced from the market and be replaced by another which is much less efficient for skilled workers, but which is safer for duffers. And so the cost of wood cabinetwork rises, making it less competitive than plastic cabinetwork. But the saw company will not have been driven out of business because it has manufactured a defective saw: it will be driven out of business because the most efficient way to insure consequent losses is foreclosed by the strict liability requirements of the law of torts. It would have been more efficient for individuals to take out their own insurance, since there is a significant contribution of individual factors—e.g., one's lack of skill—to the cause of saw-related accidents.

  Consider automobiles again. When each individual must insure himself (and if the system works rationally—in, among other things, requiring the owner of a car to carry liability insurance for his car), then if someone has a poor driving record, his insurance premium for a "performance car" will be astronomical, and he will be unable to make the purchase of that sort of car, since he will be unable to afford its insurance; on the other hand, a very skilled driver with an excellent driving record will not be charged an astronomical premium for the same car, and may be able to afford it. What we need to remember is that there is always a number of different "but for" causal factors for most accidental injuries; while some of these might be inherent in the product, others rest on the circumstances in which the product is used, and still others rest on the individual user. If what is wanted is efficiency, we would sometimes choose the manufacturer to be the insurance carrier, and other times choose the user. (Compare the case of automobiles to that of airline flight: it makes more sense for the airline to carry the insurance because, unlike the person using an automobile, the behavior of a passenger on an airline plays virtually no role in the safety of the flight—though, of course, his presence is a "but-for" cause of his injuries.) But the strict liability torts rule which Priest 37 describes requires that the insurance carrier always be the manufacturer; as a result, since the manufacturer has no legal way to discriminate between the circumstances in which his product is used once it is delivered to the retailer, nor a feasible and legal way to discriminate between those potential customers who are likely to be low risk users and those likely to be high risk users, he must pool the risks of all—which drives up the cost of the product and requires the low-risk users to subsidize the high-risk users. And if there is a sufficient number of potential high risk customers, the product might be driven from the market altogether.

  One of the main consequences of the brave new liability rules is to heighten a business' uncertainty about the possible range of liability to which its actions may expose it; and since there is no stable experience under the new rules upon which to base reliable actuarial tables, projections for insurance sky-rocket (if, indeed, any insurance company is willing to underwrite a policy at all). And what seems to me an exquisite irony is this: surely the ghost of Winterbottom haunts the current judicial landscape; for current events seem to confirm the wisdom of that court's worry about the dangers of extending liability indefinitely.38 Although Britain had declined to go down this garden path prior to joining the European Union, its tort law resembling that of the U.S. at mid-century, a recent European Community Directive "closely resembles the form of strict liability which has evolved in the United States. Somewhat ironically, the Directive's adoption and implementation have coincided with signs of a change of mood toward the strict-liability principle in the United States. This is evident first of all in the growing use [in the U.S.] of the 'risk-utility test' in preference to the test of 'consumer expectation.'"39

  Return for a moment to the auto accident prevention devices we imagined earlier. What would be more efficient is focusing involuntary accident prevention measures on those people who are least successful in avoiding accidents cheaply through voluntary means: the person who proves himself able to restrain himself from driving while drunk should be able to drive an automobile without a breath-monitor; it is wasteful to put this device in every car when it is only needed in some. But if it is installed in only some cars, the unit price will be prohibitive.40 If we are unwilling to "ground" drivers who are inclined to drive while drunk—but why should we be unwilling to do that?—then we might consider requiring others to subsidize the costs of their monitors: that would still seem to be at least cheaper than installing a monitor in every car. Under the fault regime, careful drivers pay less for their insurance since their premiums are individualized to their driving record. But if insurance were made part of the cost of a car, it would seem that the good drivers would end up subsidizing the premiums for the poor drivers. It is hard for me to see how that makes sense: there are genuine problems in internalizing the various costs of accidents; but why should we suppose that there must be a simple policy, e.g., strict liability, which will solve all of them? At any rate, this strikes me as precisely the sort of half-baked theory that could not have been spawned by a traditional court acting with traditional caution in moving case-by-case in refashioning rules according to the what had been learned from the details of previous cases.

  I gather that this experiment in devising legal policy from the bench is currently confined to fairly large manufacturing concerns—with the huge costs of litigation and insurance being passed on to their many consumers. It is worth asking whether the changes will not inevitably and perhaps rather quickly be applied to other sorts of actions performed by other sorts of concern. Suppose a small service company helps a small business "computerize" its various operations; will it then be strictly liable for its service? (If so, would any small business be able to afford the insurance premiums? Perhaps only ones which are "judgment proof"!) Again, suppose a small firm mows my grass on a regular basis: unknown to it, its decision to cut the grass to a conventional length encourages the production of cut-worms, which would damage my lawn. Is the firm strictly liable for its service notwithstanding the fact that the small extent of its business does not make feasible the systematic gathering of current horticultural scholarship? (It might be argued that only huge lawn-mowing firms will be truly efficient, considering the costs of information; nevertheless, such risks of encouraging cut-worm infestation may be worth their costs of prevention for a golf course enterprise but not be worth their costs to the average homeowner. The homeowner may be willing to bear such risks which would be catastrophic only to one whose income depended upon the certainty of a beautiful lawn.) Since these new tort rules are not rooted in common sense—in the customs of every-day behavior, it is unrealistic to expect them to be used as standards of behavior by typical individuals.41 Indeed, is it clear how one should act given these rules—beyond, perhaps, the acquiring of insurance whose reach of possible liability runs wildly further than any typical policy encountered up to now?

  The main difficulty I see with these new rules of tort liability is the way in which they evidently have been instituted. For all I know, the difficulties in the system can be rectified. But the special genius of Anglo-American common law, as I see it, lies in making changes in the law incrementally. Traditionally, changes have been made gradually, step-by-step, with many opportunities for courts to pause to consider whether the apparent changes are ill-starred, or whether a recent flood of undesirable consequences of the ruling can be handled by finding a distinction which the rule permits and which helps the rule's rationality, but which has not yet been explicitly stated by any court. The earlier series of cases which we have studied (Winterbottom to MacPherson) nicely illustrates this traditional judicial process of change. Yet, as I understand the discussions by Priest and by others, a radical new change has been implemented by various state courts virtually at a single stroke, by judicial fiat;42 and I really don't think that the courts are the places to make such comprehensive changes. If comprehensive changes are made by the legislative branch, there would be the opportunity for hearings so that experts from a huge variety of fields could testify on the likely impact of the proposed rules and on the difficulties which might be caused, and could suggest ways of revising the system to meet those problems which are likely to arise in the decades ahead.43 But a common law court simply does not have the ability to look at a policy proposal in this way: it cannot systematically consider the variety of parties—in addition to the defendant and the plaintiff—who might potentially be affected by its decision; it has no means to anticipate in detail the full consequences of the rule on the variety of activities which would be affected, as Paul Weiler44 argues. And more importantly, it cannot establish quickly a systematic policy which will be relatively certain in its scope and application. One difficulty of the current judicial tort reform is that a large number of businesses, municipalities, and insurance underwriters do not know where things stand. When such a comprehensive change in policy is implemented through court decisions in individual cases, it must be done in a tiny fraction of the possible instances where the policy would apply (in contrast with the implementation of such changes through a general statute). This results in greater uncertainty over whether the policy is intended for any given enterprise, and such uncertainty causes potential defendants to acquire wastefully redundant insurance coverage, in order to protect themselves against all possible avenues of liability, and to defer those decisions—many of them major ones—which might be effected by the outcome.

 

Endnotes:

res ipsa loquitur: Although the plaintiff in a civil action is required to prove his case on a balance of the probabilities (and not, as in the criminal law, "beyond a reasonable doubt") and is ordinarily required to produce evidence of the defendant's lack of care to which only then must the defendant respond, the doctrine of res ipsa loquitur ("the thing speaks for itself") places upon the defendant the burden of going forward—the burden of initial explanation for the accident—by creating a rebuttable presumption that the accident was caused by the defendant's negligence; it relieves the plaintiff from the need to "prove precisely what was the relevant act of omission which set in train the events leading to the accident." Lloyde v. West Midlands Gas Board, 1 WLR 749 (1971), 755, as quoted in B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 160. According to Prosser's textbook on Torts, the doctrine has three conditions:

Prosser, [The Law of] Torts [1941], p. 295, as quoted in Ybarra v. Spangard 25 Cal.2d. 486 (1944). Wigmore explains that the "particular force and justice of the rule, regarded as a presumption throwing upon the party charged the duty of presenting evidence, consists in the circumstance that the chief evidence of the true cause, whether culpable or innocent, is practically accessible to him but inaccessible to the injured person." 9 Wigmore, Evidence, 3d Ed, ί 2509, p. 382, as quoted in Ybarra v. Spangard, Ibid. [Return]

1 William L. Prosser, "The Assault Upon the Citadel (Strict Liability to the Consumer)," 69 Yale Law Journal, 1099 (1960), 1114-5. Markesinis and Deakin say that the use of res ipsa loquitur, "may effectively settle a case where neither side can offer a convincing explanation of the event in question." B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 161, emphasis added. Nevertheless, the defendant might rebut the presumption by producing a plausible explanation of the events which does not involve his carelessness. [Return]

2 Holmes, The Common Law (1923), p. 50, as quoted in B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page. 41, citation omitted. [Return]

3 B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page. 41 [Return]

4 "[T]he vast majority of [automobile] accidents are not due to 'fault', but to errors, which fallible human beings cannot avoid however carefully they may be temperamentally. For example, a study conducted by the World Health Organization in 1962 revealed that on average even a good driver makes a mistake every two miles of driving, while a comparable American study commissioned by the Department of Transportation and published in 1970 showed that in the United States the average good driver commits about nine violations (of the criminal law concerning traffic circulation) for every five minutes of driving." B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), pages 272-3. [Return]

5 B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 71, citing Latimer v. AEC Ltd. [1953] AC 643. This is similar to the test of liability first used by Judge Learned Hand: "The degree of care demanded of a person by an occasion is the resultant of three factors: the likelihood that his conduct will injure others, taken with the seriousness of the injury if it happens, and balanced against the interest which he must sacrifice to avoid the risk." Conway v. O'Brien 111 F.2d. 611 (1940), 612, as quoted in B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 145; as Markesinis and Deakin note, it is not only the interests of the defendant that must be weighed, but also the benefits to society which will be foregone if the activity is abandoned. (Ibid., page 155.) [Return]

6 Gary Schwartz, "The Myth of the Ford Pinto Case," 43 Rutgers Law Review 1013 (1991), 1037-8, citing: Henderson and Twerski "Doctrinal Collapse in Products Liability: the Empty Shell of Failure to Warn," 65 N.Y.U.L. Rev. 265 (1990), at 271-2 & n.23; and Schwartz, "Foreword: Understanding Products Liability," 67 Calif. L. Rev. 435 (1979), 449-51, 462-3. Markesinis and Deakin write that in Britain such cost/benefit notions "often seem to underly the judges' apparently intuitive notions of what is reasonable in a particular case, and there is some sign that their use is becoming more explicit." B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 155.
  Although he supposed the elements not susceptible of quantitative measurements in practice, Judge Hand cast the test in "algebraic terms" in US v. Carroll Towing Co. 159 F.2d. 169 (1947), 173: "if the probability be called P; the injury, L; and the burden, B; liability depends upon whether B is less than L multiplied by P: i.e., whether B less than PL." Markesinis and Deakin add, "An essential qualification to the test adopted by Hand is made by Professors Landes and Posner, namely that what matters in the comparison are not the total or average values but the marginal costs and benefits of eliminating a particular risk . . . If an incremental gain in safety can only be made at enormous expense, failure to take this precaution is unlikely to amount to negligence in the sense applied by the courts." B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 155. But see below the discussion of an "optimum level of accidents" and the public's evident disapproval of the use of a cost-benefit procedure to decide which accidents to prevent.
  After noting that the Hand formula, "loosely conceived" is followed by the English courts "in appropriate cases," Markesinis and Deakin observe, "If the carelessness of defendant's behaviour is to be judged, at least in part, by undertaking a calculus of the wider social costs and benefits of imposing liability, the result is to dilute the idea of fault based on individual responsibility." They go on to note, in Britain, a "tendency for negligence to verge towards strict liability in areas such as road traffic and employers' liability where the courts see defendants (or their insurerers) as better equipped than plaintiffs to absorb or shift the losses in question [via their insurance]." B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 146. [Return]

7 Daniels and Daniels v. R. White & Sons and Tabard ([1938] 4 All E.R. 258) [despite carbolic acid's having been found in the bottle of lemonade, defendant company was not proved guilty of a breach of its duty to take reasonable care in bottling the lemonade]; Neil MacCormick provides an extensive discussion of the case against the manufacturer and the retailer in his Legal Reasoning and Legal Theory (Oxford: 1978). [Return]

8 "Thus far [as of Henningsen v. Bloomfield Motors in 1960 and Greenman v. Yuba Power Products in 1963] the innovation was seen more as a cleaning-up operation, freeing manufacturers' liability merely from the law of sales and dispensing with the need to stretch res ipsa loquitur rather than as opening up a novel and potentially explosive source of liability. Indeed, one of its mooted attractions was that it would reduce litigation by stopping up undeserving loopholes through which a guilty defendant might hope to escape. Unhappily, these sanguine expectations were soon turned to ashes as it became clearer that the courts were opening a Pandora's box. Products liability was to transform the whole agenda of tort litigation: inaugurating a new era of mass actions, mass costs, and mass awards." John G. Fleming, The American Tort Process (Oxford: 1988), page 61, footnotes omitted. [Return]

9 Priest writes about the history of changes in tort law, "The conceptual foundations of tort law began to shift, however, roughly at the turn of the century as legal scholars and policymakers began to consider tort law more seriously as an affirmative public policy tool." George L. Priest, "The modern expansion of tort liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991) page? . Also see George L. Priest, "The New Legal Structure of Risk Control," 119 Daedalus: Journal of the American Academy of Arts and Sciences 207 (Fall, 1990). According to Atiyah and Summers, "almost the whole of the modern law of strict products liability in tort has originated with academic writings." Atiyah and Summers, Form and Substance in Anglo-American Law (Oxford: 1987), page 402. In a footnote, they mention Justice Traynor in California, a former law professor, as having been a leader in this enterprise, together with Professor William Prosser.

  Fleming writes,

John G. Fleming, The American Tort Process (Oxford: 1988), page 60, footnotes omitted. [Return]

10 A. Mitchell Polinsky, An Introduction to Law and Economics second edition (Aspen Publ: 1989 [formerly published by Little, Brown, & Co.]), page 18. [Return]

11 Whether the higher wages accurately reflect the relatively greater risks to life and limb from working in the more dangerous industry will depend upon whether the workers have reliable information about the nature and extent of its risks and about their alternative employment opportunities. It is, of course, difficult and expensive to gain such information (consider the artificiality of judicial attribution of such knowledge to workers which was common in the late nineteenth century and early twentieth century); and it may be more difficult for a worker to obtain such information than for an employer. In addition, while it may be evident to a sophisticated observer that the worker has viable alternatives to working at a dangerous occupation bearing only a marginally greater wage, the individual himself may lack useful information about alternative occupations open to him. In any case, workman compensation costs understate the total costs to a workman and his family from work-related injuries. [Return]

12 Of course, we are speaking about only one clause of an insurance contract, insuring oneself against injuries due to defects in the automobile; the more familiar clauses would, e.g., insure others for injuries due to one's negligent driving and would be needed with or without strict product liability. [Return]

13 Actually, this requires a qualification. If a poor person declines to purchase insurance against the risk his driving imposes on others, because insuring against this risk to others is not worth its cost to him, then other drivers will subsidize the uninsured motorist. And if he declines to insure himself against defects in his automobile and is injured as a result of those defects, the costs of those injuries may be thrown upon the public via unreimbursed emergency room expenses. [Return]

14 Whitford argues that, as of the late 1960s, even when theories of strict liability were applicable, plaintiffs in Wisconsin often continued to pursue on a negligence basis. Perhaps as a result of cases such as the 1960 New Jersey Henningsen v. Bloomfield, in legal cases and even in informal attempts at resolution, auto manufacturers rarely advanced claims that implied or express warranties were explicitly disclaimed in the purchase agreement. Whitford thinks that most cases which proceeded on a negligence-basis were virtually indistinguishable from the way they would have proceeded under a strict liability basis. William C. Whitford, "Strict Products Liability and the Automobile Industry: Much Ado about Nothing," 1968 Wisconsin Law Review 83 (1968).

  While many recent essays conceive the move to strict liability as a break from the traditional fault doctrines, it is possible to look upon the move as an attempt to remove flaws in the traditional procedures—as was noted in the second paragraph, above. Markesinis and Deakin write:

B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), pages 531-2.
 
In case law, a "defective" product is variously judged in comparison to "the average quality of like products" or the average quality of normal samples of the manufacturer's product. R. Traynor, "The Ways and Meanings of Defective Products and Strict Liability," 32 Tennessee Law Review 363 (1965), 367. The (Second) Restatement of Torts uses "the surprise element of danger," comparing the dangerousness of an article to what "would be contemplated by the ordinary consumer who purchases it, with the ordinary knowledge common to the community as to its characteristics." Restatement § 402A, Comment i (Tent. Draft No. 10, 1964), as quoted by Traynor, ibid., p. 370; Traynor adds that a "warning or notice might cure a defect attributable to the product." Ibid., p. 372. [Return]

15 Priest writes, "A law concerned with internalizing costs and providing insurance is vastly different from a law seeking only to penalize the abnormal. Whether the controlling doctrine is negligence or strict liability, central concepts of causation are changed substantially. The earlier regime that imposed a sharp distinction between particularly extreme sources of harm and all others was necessarily committed to a very strict concept of causation. Actions were subject to liability for causing harm chiefly if they constituted the sole or exclusive source of the harm. In contrast, our modern law devoted to cost internalization focusses less upon strict causation than upon contribution to the occurrence of a harm. Any increase in risk must be regarded as a cost, and may constitute a cost worth internalizing, though it was only one of many simultaneously contributing sources of the loss." George L. Priest, "The modern expansion of tort liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991) page? . [Return]

16 A story in June 15, 2000 New York Times, Technology/Circuits section, "Device Helps Keep D.W.I. Offenders and Others Safe," by KAREN J. BANNAN, describes an ignition interlock system which uses an alcohol breath analyzer to prevent the car's being started when alcohol is detected. Currently, 40,000 of such devices are in use in the United States. "The cost of the installation and monthly rental, which are paid for by the offender, vary from state to state. In New York, the installation costs $60 and the rental fee is $2.14 per day or about $65 per month," according to the report.
    In another story in November 20, 2006 New York Times, "A New Strategy to Discourage Driving Drunk," MATTHEW L. WALD writes

“ The threat of arrest and punishment, for decades the primary tactic against drunken drivers, is no longer working in the struggle to reduce the death toll, officials say, and they are proposing turning to technology — alcohol detection devices in every vehicle — to address the problem. . . .
“ 'We've seen no progress in 10 years; we're completely stalled,' said Susan A. Ferguson, a researcher at the Insurance Institute for Highway Safety.
“ Ms. Ferguson said the most promising technologies would work automatically, like air bags. 'We don't want the soccer mom dropping kids off, going to the grocery store and the preschool, and having to blow into something every time,' she said. . . .
“ Future devices may read alcohol content when a driver's palm touches the steering wheel or the gear shift lever, said Jim McNally, the chief executive of TruTouch [a company developing interlock devices]. . . .
[Return]

17 Peter Godwin, "The Car That Can't Crash," 6-10-00 New York Times, reports riding in such an experimental car: "Once the Cadillac [which is just ahead] moves off, our vehicle resumes its previous speed, following at a safe distance, and my companion explains what has just happened. Our car is equipped with Adaptive Cruise Control, a technology developed by Delphi Automotive Systems to prevent cars from rear-ending one another. Radar in our front grille has spotted the offending car, calculated its speed relative to ours and sent that information back to our automatic throttle and braking controls. When the system senses that we are on a collision course, it overrides the driver to take its precautionary action. " [Return]

18 While any entity can acquire information about the possibilities and effectiveness of accident prevention technology, it will be expensive to acquire that information; the manufacturer can, however, cheaply apportion the costs of acquiring this information among the purchasers of his automobiles. [Return]

19 In his influential concurrence (in result) in Escola, Justice Traynor proposed the use of strict liability in product injury cases on grounds of public policy; he suggested, "Even if there is no negligence [of the manufacturer], however, public policy demands that responsibility be fixed wherever it will most effectively reduce the hazards to life and health inherent in defective products that reach the market. It is evident that the manufacturer can anticipate some hazards and guard against the recurrence of others, as the public can not. . . . The cost of an injury and the loss of time or health may be an overwhelming misfortune to the person injured, and a needless one, for the risks of injury can be insured by the manufacturer and distributed among the public as a cost of doing business. It is to the public interest to discourage the marketing of products having defects that are a menace to the public." Escola v. Coca Cola Bottling Co. of Fresno, (Supreme Court of California: 1944), 24 Cal. 2d. 453 (1944), 462. Later, Chief Justice Traynor persuaded his court on the merits of strict liability in Greenman v. Yuba Power Products, Inc., (Supreme Ct. of California: 1962), 377 P. 2d. 897 (1962) [manufacturer's strict liability for defective products]. Also see an account of the expansion of the rejection of privity as a bar to tort liability from MacPherson, in William L. Prosser, "The Assault Upon the Citadel (Strict Liability to the Consumer)," 69 Yale Law Journal 1099 (1960), and William L. Prosser, "The Fall of the Citadel (Strict Liability to the Consumer)," 50 Minnesota Law Review 791 (1966).
  In my illustration using automobile accidents, it might be thought that the manufacturer of an automobile would not be liable in an accident in which one motorist might have brought about the accident by driving carelessly or recklessly. Yet there are many such lawsuits in which a manufacturer is charged with a proportion of the total liability because a different design of the automobile might have prevented or mitigated the injuries despite the motorist's carelessness; municipalities are also frequently sued on the ground that a different roadway design might have taken away the opportunity for such an accident. I suppose that it is obvious that an important motive for making a manufacturer a party to the suit is that it has "deep pockets," the financial ability to pay a huge judgment; such judgments will eventually be recovered by the manufacturer through raising the prices of his cars. [Return]

20 Fleming suggests that Justice Traynor advanced "a new policy argument that shifting of loss suffered by victims of accident could be justified on the ground that the defendant was in a strategic position, if not to prevent accidents, at least to provide compensation. The cost might act as an incentive to greater investment in safety and could in any event be spread by insurance or higher prices among the public.
  "This theorem not only broke with the nineteenth-century commitment to fault liability; even more radical was its departure from the age-old view of tort liability as an instrument of corrective justice. For it lent respectability to the view that compensation could be justified by appeal to the broader notions of social and economic policy, transcending interpersonal equities, like economic efficiency or redistribution of wealth on the pattern of social insurance. . . . If the legislators were unwilling to impose a new direct tax burden on the public for [social insurance program] expenditures, the courts were prepared to convert tort law into a welfare system." John G. Fleming, The American Tort Process 11-12 (Oxford: 1988. [Return]

21 Stephen F. Williams, "COMMENTARY. Second Best: The Soft Underbelly of Deterrence Theory in Tort," 106 Harvard Law Review 932 (1993), 932-3, citations omitted—[arguing from theory of the second best that strict liability "will not necessarily enhance safety or health, or bring the economy closer to optimal performance" since it will not correct error about the hazards of alternatives not subject to liability (e.g., foregoing DTP vaccination rather than paying the liability-increased price), nor about the positive health and safety features of liability-covered goods and services. (933).] [Return]

22 Ronald H. Coase, "The Problem of Social Costs," 3 Journal of Law & Economics 1 (1960). [Return]

23 There is an "optimum" level of accident costs since it will not be efficient to prevent every accident; some accidents will cause injuries costing less than the expenditure required to have prevented them. Imagine that in handling paper for our computer printers, we sustain a paper cut for every thousand sheets we use. Suppose that the cheapest way to prevent paper cuts would be to sand the edges of the paper reams at the time of manufacture, adding 7 cents to each thousand sheets; but for each paper cut, we could instead supply a band-aid for 2 cents. In this case, preventing these accidents would not be efficient; it is cheaper not to prevent them. So, the "optimum" level of accident costs is that level at which it would be more expensive to avoid further accidents than to bear their costs.
  (But suppose it is foreseen by the manufacturer that, statistically, in every billion cases of paper cuts there will be one consequent instance of blood-poisoning resulting in amputation. And suppose, further, that the costs of this risk (the monetary costs of medical expense and loss of income multiplied by the probability of the injury) will be less than the costs of prevention; then it would seem that the "optimum" level of accident costs would exclude prevention of this possible injury. Efficiency (and rationality, on the Learned Hand test) would seem to dictate that the manufacturer should not prevent such accidents. Yet, in many famous suits, juries impose "punitive" or "exemplary" damages to punish the defendant for his "outrageous" willingness to refrain from preventing the injury because he has conducted just such a cost/benefit analysis. Compare Moseley v. General Motors (1993), concerning the design of a GM truck in which the gas tank was located outside the frame, and the earlier Ford Pinto cases.
  After noting how the popular press, including Mother Jones and Mike Wallace on Sixty Minutes, excoriated Ford for having deliberately used cost-benefit considerations in deciding questions of potential safety, Schwartz writes: "What my previous article suggested, in a Calabresian way [à la Calabresi and Bobbit, Tragic Choices], is that the public subscribes to the idea of the pricelessnesss of life, and therefore is firmly opposed to processes of risk-benefit balancing. Now I should make clear that this is an opposition which I myself do not share. For over twenty years I have been immersed in the academic literature on the negligence standard, and more generally in the public policy literature on the management of risk. Given this experience, I am convinced that risk-benefit analysis is not only proper, but just about essential." Gary T. Schwartz, "The Myth of the Ford Pinto Case," 43 Rutgers Law Review 1013 (1991), 1041, citations omitted. Elsewhere he notes that the risk/benefit calculations would not balance $10 against one death prevented, but instead, for example, $10 multiplied by ten million cars sold against the probable number of deaths prevented—which might be as few as five. Ibid., 1060.) [Return]

24 Although both parties will be better off after the exchange, the initial assignment of liability may have distributional consequences affecting the relative wealth of the parties; e.g., the party bearing the liability costs may have a poorer bargaining position than the other party. See B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 24. [Return]

25 For automobiles, this would include the costs of destroying existing and experimental automobiles in crash tests and the costs of compiling accident statistics from accidents categorized by vehicle brands and models. [Return]

26 Of course, a government might intervene in the market in order to require that advertisements meet certain criteria of accuracy. But note that such regulation would carry a substantial transaction cost, not only for the bureaucracy required to acquire the relevant information and then to create and enforce appropriate regulations, but also for the manufacturers to understand and conform to them. (You should not think of these costs as irrelevant to the consumer: he will eventually pay for them in taxes and in the increased price of the products; while government regulation may frequently be worth its cost, it should not be thought of as a free good.) [Return]

27 Still, a manufacturer might try, through product differentiation and advertising, to induce consumers to choose that product priced so as to recoup the average insurance costs for the group of consumers to which the product is likely to appeal. According to Priest, "Some firms have made greater investments in product or model differentiation in order to attempt to segregate low-risk from high-risk consumers, such as the differentiation of domestic from backcountry four-wheel-drive vehicles . . . . It is difficult for others than industry experts to distinguish insurance from consumer preference trend, at least with respect to more extreme risks." George L. Priest, "The Current Insurance Crisis and Modern Tort Law" 96 Yale Law Journal 1521, as abridged and reprinted in Saul Levmore, Foundations of Tort Law (Oxford: 1994), page 297. [Return]

28 Imagine two rival products which serve the same purpose, one of which has been designed and manufactured cost-effectively to be safe, and the other which has not; although the one product's purchase price is more expensive (as a result of the manufacturer's investment in safety) than the purchase price of the other, when the costs of accidents which use of the products will eventually entail are considered, the product costing most initially is much cheaper to purchase and use than its rival product. If the consumer lacks relevant information about the ultimate differences in relative savings in accident costs, he may falsely suppose that the product which is initially cheaper is the better buy. The hope of a policy of strict manufacturers' liability is to make the costs of accidents part of the initial purchase price of the product so that consumers will make their choices on the basis of the actual total costs of the product. As I go on to explain in the text, the success of this policy will depend upon those costs of accidents being approximately the same for the various consumers. [Return]

29 We can see how different people will value accident prevention differently by returning to our earlier explanation of an "optimum" level of accident prevention. We imagined that while the cheapest way to prevent paper cuts would add 7 cents to the cost of each thousand sheets, it would be cheaper still to suffer the accidents, since a band-aid would be cheaper at 2 cents per thousand sheets of paper. But a eye surgeon, a pianist, or a jeweler, might well prefer to pay 7 cents to avoid such accidents than to suffer them: the actual costs of a paper cut accident to these people will be greater than the costs to the average person. If the manufacturer prevents these accidents by sanding the edges of all reams of paper, the average paper user will pay more to avoid paper cut accidents than their avoidance is worth to him; he will be subsidizing the higher costs of accident prevention to others.
  Note that these are differences in individuals' rational preferences for accident avoidance costs. It is, of course, notorious that young people tend to place an irrationally high discount on future accident costs (believing that catastrophic accidents won't happen to them) in indulging in bungie-jumping and similar risky behavior; but it is only differences in rational preferences which an efficient market would accommodate. (If the existence of wide spread irrational preferences are foreseeable, that creates a problem for a market solution.)

  We need to recognize a serious practical problem in a market system's allowing consumers to choose the level of risk they are willing to pay for: its efficient operation requires that these consumers have access to accurate information about those risks. But consider a New York Times story on the Grace Manufacturing Company: while the substantial health hazards of asbestos were becoming publicly known in the 1970s, the Grace Company promoted an alternative to asbestos, tremolite (in its product, Monokote), without disclosing the information available to its researchers that its product, though safer than asbestos, might well pose a similar kind of risk. In considering its costs/benefits in marketing the product, it (cynically) discounted the risks of liability suits by the unlikelihood that potential plaintiffs would be successful in documenting harm from their product:

  Many recent news stories about tobacco and automobile and drug lawsuits suggest that many companies frequently pursue a policy of not disclosing information about the risks of their products to consumers (and a policy of not disclosing the greater risks to their employees in manufacturing the products); when pertinent information is standardly withheld by manufacturers, the consumer cannot make efficient and rational choices about his costs of accident prevention unless he has some independent source of information. It should be noted that such undisclosed information about risks is often unearthed in tort suits using discovery procedures (allowing internal company memoranda to be subpoenaed) which are not otherwise available to individual consumers. So what would be inefficiencies of government regulation and strict manufacturer liability in a truly open and fair market may be efficient responses to widespread manufacturer predation.
  Indeed, it may be that the government's regulatory machinery is inadequate to block the intentional thwarting of occupational safety regulations; and unlike the case of manufacturing defects, Workmen's Compensation rules may preclude resort to tort suits by the injured party. Consider the study of the McWane family of companies conducted by the New York Times, PBS's Frontline, and the Canadian Broadcasting Corporation program, "The Fifth Estate":
[Return]

30 "Some risks [may be] unavoidable, in particular in relation to hidden design defects; a strict-liability regime which holds the manufacturer automatically liable for the resulting harm may lead to manufacturers taking excessive amounts of care, pushing prices up beyond a level which reflects the potential costs to society of product defects, driving producers out of the market, or inhibiting innovation. Recent empirical studies have backed up [such] claims with evidence that product innovation has been stymied by the threat of high damages awards based on strict liability. [fn. 24: See Financial Times, Feb. 1993, referring to a Brookings Institute study which found that product-liability litigation 'was not an overwhelming force for safety' and that uncertainty over the scale of damages awards had 'probably reduced innovation.']" B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 542, articulating challenges, made to the assumption that manufacturers must always be the best risk avoiders, by critics such as R.A. Epstein, Modern Product Liability Law (1970); demurring to this position in the case of personal injury, they add, "The less desirable features of the American system may well be the result not of strict liability as such, but rather of the prominent role of the jury, and the incentive to litigation provided by the contingency-fee system." Ibid., p. 543. [Return]

31 Priest writes, "Most courts have accepted three central presumptions in warning cases . . . The three presumptions that the manufacturers possess complete knowledge about product dangers, that warnings of these dangers are easily and cheaply given, and that all consumers will read, comprehend and act on them, has led warning law close to absolute liability. Except where intentionally self-inflicted, all product-related injuries can be said to derive from insufficient information about product risks. Where a consumer is injured despite the existence of a warning, the warning must have been inadequate. Though consumers do not always recover in warning cases, and though courts continue to claim that the controlling standard is strict, not absolute liability (that is, that some defenses to manufacturers remain), these presumptions leave no principled grounds for denying recovery." George L. Priest, "The modern expansion of tort liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991) page? . [Return]

32 There is some ambiguity in speaking here of "more expensive ways of preventing accidents." Suppose that advertisements which warn people to slow down in construction zones (or, not to drink alcohol before driving) and which remind them of the expensive penalties in being caught speeding in such zones (or, in being arrested for DWI) will prevent the majority of such accidents; the average cost of preventing these accidents (those accidents preventable by this method) might be small even though very many accidents remain unprevented by this method. But suppose that through more expensive alterations in the design of automobiles, we could prevent a greater number of accidents albeit at a greater average cost; the total savings in the costs of accidents (the costs of accidents prevented minus the costs of the design alterations) from the second method might be greater than the total savings from the first method—since although the costs of prevention for the first method are less, its savings in accident costs are also less. [Return]

33 Priest characterizes the traditional negligence standard in this way: "Throughout most of the 19th century, negligence was defined chiefly in terms of the extent to which the injury-causing behavior deviated from the normal. Obviously, intentional harm-causing actions justified liability and, beyond the intentional, unusually egregious actions would justify liability as well. At this time, the standard 'strict liability' meant absolute liability largely insensitive to any defense and was applied in specific factual circumstances where the abnormality of the injurer's activity was so obvious as to be beyond debate. In the 19th century, for example, individuals were strictly liable for injuries caused by wild animals that they kept as pets or injuries caused by explosives.
  "It is obvious that a legal regime built upon a standard of abnormality is a regime of very limited scope. By definition, egregiously abnormal behavior occurs infrequently. . . .
  "Before the 1960s, a ladder was a ladder, and a fall was a deviation from normal consumer use. Today, ladder manufacturers are viewed as in a superior position either to design the ladder to make it more secure, to warn consumers effectively how to use it (say, to warn them in a way that keeps them from standing on the top rungs) or, if those actions fail, to provide insurance to consumers for ladder-related injuries. Similarly, before the revolution, doctors or hospitals could be presumed to have done the best they could, but it was acknowledged that all medical procedures involve hazards that will afflict some patients. After the revolution, in contrast, medical malpractice litigation under the negligence standard has expanded dramatically, because it is almost always possible to allege--and thus to justify reference to a jury--that the doctor or hospital could have obtained superior equipment, could have used more sophisticated procedures, could have employed larger or more skilled support staffs, or would be likely to invest more fully in technological advances if faced with the prospect of liability judgments." George L. Priest, "The Modern Expansion of Tort Liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991), in the sections entitled, "The Triumph of Strict Liability in 20th Century Tort Law," and "The Modern Tort Regime Described." [Return]

34 For example, in a rather abstractly framed argument, Alan Schwartz advocates free contract with regulated disclosure. Alan Schwartz, "The Case Against Strict Liability," 60 Fordham Law Review 819 (1992), 840-1. Priest also seems to frame the discussion about strict liability as a question of who might purchase insurance most efficiently. [Return]

35 Priest writes, "In 1960, the New Jersey Supreme Court ruled in the case Henningsen v. Bloomfield Motors, Inc., that cases involving personal injury from product use would no longer be governed by warranty law that had controlled such actions for the preceding 100 years (32 N.J. 358, 161 A.2d 69 [1960]), The Court described product warranties as tools for exploiting consumers, denying such warranties any future effect." George L. Priest, "The modern expansion of tort liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991) page? .
  See also a further historical note on exclusions of warranties in standard form-contracts. [Return]

36 Stephen F. Williams, "COMMENTARY. Second Best: The Soft Underbelly of Deterrence Theory in Tort," 106 Harvard Law Review 932 (1993), fn. 22 on page 938, citing Peter Huber, Liability: The Legal Revolution and Its Consequences (1988). [Return]

37 "Our courts have defined two basic principles of decision making to internalize costs to create incentives to reduce the risk level as much as is practicable. First, if the injury could have been practicably prevented, liability will be placed on that party in the relatively better position to prevent it. Second, if the injury could not have been practicably prevented, liability will be placed on that party in the relatively better position to spread the risks of the injury [i.e., through insurance]." . . .
  "But the increase in insurance costs alone still does not explain why so many products and services have been made unprofitable. The very large majority of American consumers possess first-party and health insurance. . . . It is not clear that the expansion of civil law to spread risks creates subsantial new insurance obligations. The expansion of law to provide insurance could only make products and services unprofitable, thus, if the costs of providing insurance through tort law were substantially greater than the costs of similar first-party insurance.
  "There are strong reasons to believe that tort law insurance is substantially more costly than equivalent first-party insurance and that this cost difference is the underlying cause of the widespread product and service withdrawals. First, because of the way damages are measured in tort law, for similar losses the benefits provided are roughly twice to three times the benefits provided under any first-person insurance system. Second, the operational costs of providing insurance through an adversarial tort system are roughly five to ten times the operational costs of first-party insurance. Third, the ability of insurers to differentiate among the insured population in order to reduce costs is vastly constrained where insurance is delivered through tort law rather than through first-party coverage. For these reasons, tort law insurance is conservatively estimated to cost five to ten times more than first-party insurance for coverage of the identical injury [citing Priest, "The Current Insurance Crisis and Modern Tort Law," 96 Yale Law Journal (1987): 1521-90].
  ". . . [Additionally,] spreading risks by expanding civil liability compels the low-income workers and the poor in our society to subsidize the insurance costs of the high-income earners and the wealthy." George L. Priest, "The New Legal Structure of Risk Control," 119 Daedalus: Journal of American Academy of Arts and Sciences 207 (1990), pp. 216 & 224-5. [Return]

38 Fleming writes, "Perhaps, in days gone by, the spectre of 'opening floodgates' was overdrawn; still, it revealed a keen sense of the long-term cost implications of a given decision. By contrast, the new ideology betrays little concern for the cost of its judicial welfare programme, in the insouciant or naïve belief that insurance and deep pockets will take care of the problem. The tort system has thus become about the only segment of the economy not subject to the discipline of prudent resource allocation: it is a programme without a budget." John G. Fleming, The American Tort Process (Oxford: 1988), page 13. [Return]

39 B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 539; see pages 533-9 for a discussion of Directive 85/374/EEC.
  The basis for the consumer expectation test which has been used in various American jurisdictions is Comment i to section 402A of the Second Restatement of Torts, which provides a gloss on "any product in a defective condition unreasonably dangerous to the user or consumer": "the article sold must be 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." Restatement (Second) of Torts, s. 402A, as quoted in B.S. Markesinis & S.F. Deakin, Tort Law—third edition, (Oxford: 1994), page 533. Critics charge that this test fails to work in cases of defective design, where "the consumer simply does not have adequate information to know what to expect." O'Brien v. Muskin Corp., 463 A.2d 298 (1982), 308, as quoted at page 540, Ibid.
  Although resembling the Learned Hand test in saying "that on balance the benefits of the challenged design outweigh the risk of danger inherent in such design . . . [nevertheless, the newer U.S. version of the risk-utility test] explicitly focuses the trier of fact's attention on the adequacy of the product itself, rather than on the manufacturer's conduct, and places the burden on the manufacturer, rather than on the plaintiff, to establish that because of the complexity of, or trade-offs implicit in, the design process, an injury-producing product should nevertheless not be found defective." Ibid., 539-40, as quoted from Barker v. Lull Engineering Co. 573 P.2d 443 (1978), 452-6. After pointing out that this new risk-utility test is claimed not to make a manufacturer "an automatic insurer of consumer injuries, and [that] in practice it may not be too difficult for a manufacturer to present enough evidence to meet the presumption against him," (but see Priest's warning about juries' uses of this standard), Markesinis and Deakin add: "A recent review of the American case-law suggested that in most jurisdictions consumer-expectation is retained only for cases of non-design defect (meaning products which are badly manufactured according to their specification, such as a car with faulty brakes). . . . Empirical Research suggests that the [American] courts are moving away from high damages claims of their own accord, and are at the same time taking a more restrictive view of the scope of manufacturers' liability." Ibid., pages 540-1, citing a study of judgments of appellate courts by Henderson and Eisenberg, "The Quiet Revolution in Products Liability," 36 UCLA Law Review 479 (1990). [Return]

40 But perhaps ways might be found to make such things affordable: see endnote 16. [Return]

41 Also, there is a problem with how juries may be expected to apply these brave new standards. Priest writes, "Many states retain a 'consumer expectations' test for design defects, allowing the claim to be referred to the jury based on the injured consumer's testimony concerning what level of safety he or she expected. . . . The more common standard for whether a product has been defectively designed, however, is what is called the 'risk-utility' test, according to which the jury is instructed that the manufacturer can be held liable for the injury if the risk of product injuries given current product design exceeds the product's utility. . . . But there is no agreed upon way to measure product utility: the question is left for differing expert testimony. . . . The risk-utility test survives as it does because only juries, and not judges, are forced to interpret and apply it. Its operational significance is that it allows courts to justify presentation to a jury of a wide range of expert testimony, for both plaintiffs and defendants, concerning the product design process. As might be expected, outcomes have been widely variant, both for individual products and across products. Some decisions have approached the absurd." George L. Priest, "The modern expansion of tort liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991) page? . Moreover, in adding "punitive damages" to the awards of compensation, juries seem to be combining a negligence regime with a strict liability regime.

  (Yet, despite such problems, if we want to evaluate current tort law in America, we must compare the entire package—with its inefficiencies—to other likely packages of tort law together with the systems of consumer protection which would accompany them. For as things stand now, much of the job of consumer protection is accomplished through the tort system by individuals' suing manufacturers for damages caused by defective products. Government regulation seems unequal to the task, and even government mandated disclosures of the various risks of products seems largely driven by tort suits. In one's evaluation, one must factor-in what strikes this observer as a more predatory relationship of manufacturers towards their consumers than was typical in the early 1900s; while "buyer beware" was then the rule, the seller did not actively disguise relevant information. It is possible, of course, that the changes in the tort system have provided an important incentive to manufacturers to alter their historical relationships to their consumers: to disclose information and warnings in a timely way may be thought to invite and nourish tort suits, including frivolous ones.) [Return]

42 Priest writes, "in 1964, the prestigious American Law Institute, in its Second Restatement of Torts, announced its acceptance and recommendation of the standard of strict liability for defective products without regard to the manufacturers' negligence (Restatement Section 402A). . . . By 1971, 28 states had adopted the Restatement provision or a similar version of strict products liability; by 1976, 41 states; today [1991], all but two." George L. Priest, "The modern expansion of tort liability: Its sources, its effects, and its reform," 5 Journal of Economic Perspectives 31 (1991) page? (citations omitted). [Return]

43 Yet legislative action in the various states appears to have been stalemated by opposing lobbies. Fleming writes, "Reform of tort law in the American context assumes a highly adversarial posture . . . by opposing organizations of lawyers: on the one hand, the trial lawyers [especially the American Trial Lawyers Association]; on the other side, the defence bar . . . representing potential tort defendants, like the American Manufacturers Association on issues of products liability, and of course liability insurers. . . . The effect of this highly structured confrontation is to substantially paralyze legislative reform in the torts area. This has not always been so. Historically, state legislatures were in general dominated by conservative majorities and interests." John G. Fleming, The American Tort Process (Oxford: 1988), pages 40-1, footnotes omitted.
  In contrast to the U.S., Britain has seemed able to revise its legal system systematically through effective legislative means. [Return]

44 Paul Weiler, "Legal Values and Judicial Decision-Making," 48 Canadian Bar Review 1 (1970); and his "Two Models of Judicial Reasoning," 46 Canadian Bar Review 406 (1968). But Fleming warns, "Not that reform statutes invariably exploit this [theoretical] advantage [of issuing a comprehensive policy] . . . [since some leave] all the details, including important policy questions . . . to be worked out by the courts through the protracted process of random litigation." John G. Fleming, The American Tort Process (Oxford: 1988), pages 52-3 footnotes omitted. [Return]