Posner on Strict Liability

March 14, 2006

The roots of section 520 [of the Restatement 2d of Torts] are in nineteenth-century cases. The most famous one is Rylands v. Fletcher, but a more illuminating one in the present context is Guille v. Swan. A man took off in a hot-air balloon and landed, without intending to, in a vegetable garden in New York City. A crowd that had been anxiously watching his involuntary descent trampled the vegetables in their endeavor to rescue him when he landed. The owner of the garden sued the balloonist for the resulting damage, and won. Yet the balloonist had not been careless. In the then state of ballooning it was impossible to make a pinpoint landing.

Guille is a paradigmatic case for strict liability. (a) The risk (probability) of harm was great, and (b) the harm that would ensue if the risk materialized could be, although luckily was not, great (the balloonist could have crashed into the crowd rather than into the vegetables). The confluence of these two factors established the urgency of seeking to prevent such accidents. (c) Yet such accidents could not be prevented by the exercise of due care; the technology of care in ballooning was insufficiently developed. (d) The activity was not a matter of common usage, so there was no presumption that it was a highly valuable activity despite its unavoidable riskiness. (e) The activity was inappropriate to the place in which it took place — densely populated New York City. The risk of serious harm to others (other than the balloonist himself, that is) could have been reduced by shifting the activity to the sparsely inhabited areas that surrounded the city in those days. (f) Reinforcing (d), the value to the community of the activity of recreational ballooning did not appear to be great enough to offset its unavoidable risks.

These are, of course, the six factors in section 520. They are related to each other in that each is a different facet of a common quest for a proper legal regime to govern accidents that negligence liability cannot adequately control. The interrelations might be more perspicuous if the six factors were reordered. One might for example start with (c), inability to eliminate the risk of accident by the exercise of due care. The baseline common law regime of tort liability is negligence. When it is a workable regime, because the hazards of an activity can be avoided by being careful (which is to say, nonnegligent), there is no need to switch to strict liability. Sometimes, however, a particular type of accident cannot be prevented by taking care but can be avoided, or its consequences minimized, by shifting the activity in which the accident occurs to another locale, where the risk or harm of an accident will be less ((e)), or by reducing the scale of the activity in order to minimize the number of accidents caused by it ((f)). By making the actor strictly liable — by denying him in other words an excuse based on his inability to avoid accidents by being more careful — we give him an incentive, missing in a negligence regime, to experiment with methods of preventing accidents that involve not greater exertions of care, assumed to be futile, but instead relocating, changing, or reducing (perhaps to the vanishing point) the activity giving rise to the accident. The greater the risk of an accident ((a)) and the costs of an accident if one occurs ((b)), the more we want the actor to consider the possibility of making accident-reducing activity changes; the stronger, therefore, is the case for strict liability. Finally, if an activity is extremely common ((d)), like driving an automobile, it is unlikely either that its hazards are perceived as great or that there is no technology of care available to minimize them; so the case for strict liability is weakened.

The largest class of cases in which strict liability has been imposed under the standard codified in the Second Restatement of Torts involves the use of dynamite and other explosives for demolition in residential or urban areas. Explosives are dangerous even when handled carefully, and we therefore want blasters to choose the location of the activity with care and also to explore the feasibility of using safer substitutes (such as a wrecking ball), as well as to be careful in the blasting itself. Blasting is not a commonplace activity like driving a car, or so superior to substitute methods of demolition that the imposition of liability is unlikely to have any effect except to raise the activity’s costs.

Against this background we turn to the particulars of acrylonitrile. Acrylonitrile is one of a large number of chemicals that are hazardous in the sense of being flammable, toxic, or both; acrylonitrile is both, as are many others. A table in the record, drawn from Glickman & Harvey, Statistical Trends in Railroad Hazardous Material Safety, 1978 to 1984, at pp. 63-65 (Draft Final Report to the Environmental & Hazardous Material Studies Division of the Association of American Railroads, April 1986) (tab. 4.1), contains a list of the 125 hazardous materials that are shipped in highest volume on the nation’s railroads. Acrylonitrile is the fifty-third most hazardous on the list. Number 1 is phosphorus (white or yellow), and among the other materials that rank higher than acrylonitrile on the hazard scale are anhydrous ammonia, liquified petroleum gas, vinyl chloride, gasoline, crude petroleum, motor fuel antiknock compound, methyl and ethyl chloride, sulphuric acid, sodium metal, and chloroform. The plaintiff’s lawyer acknowledged at argument that the logic of the district court’s opinion dictated strict liability for all 52 materials that rank higher than acrylonitrile on the list, and quite possibly for the 72 that rank lower as well, since all are hazardous if spilled in quantity while being shipped by rail. Every shipper of any of these materials would therefore be strictly liable for the consequences of a spill or other accident that occurred while the material was being shipped through a metropolitan area. The plaintiff’s lawyer further acknowledged the irrelevance, on her view of the case, of the fact that Cyanamid had leased and filled the car that spilled the acrylonitrile; all she thought important is that Cyanamid introduced the product into the stream of commerce that happened to pass through the Chicago metropolitan area. Her concession may have been incautious. One might want to distinguish between the shipper who merely places his goods on his loading dock to be picked up by the carrier and the shipper who, as in this case, participates actively in the transportation. But the concession is illustrative of the potential scope of the district court’s decision.

No cases recognize so sweeping a liability. Several reject it, though none has facts much like those of the present case. With National Steel Service Center v. Gibbons, which held a railroad strictly liable for transporting propane gas — but under Iowa law, which uses a different standard from that of the Restatement — we may pair Seaboard Coast Line R.R. v. Mobil Chemical Co. which refused to impose strict liability on facts similar to those in this case, but again on the basis of a standard different from that of the Restatement. Zero Wholesale Co. v. Stroud refused to hold that the delivery of propane gas was not an ultrahazardous activity as a matter of law. But the delivery in question was to a gas-storage facility, and the explosion occurred while gas was being pumped from the tank truck into a storage tank. This was a highly, perhaps unavoidably, dangerous activity.

Siegler v. Kuhlman also imposed strict liability on a transporter of hazardous materials, but the circumstances were again rather special. A gasoline truck blew up, obliterating the plaintiff’s decedent and her car. The court emphasized that the explosion had destroyed the evidence necessary to establish whether the accident had been due to negligence; so, unless liability was strict, there would be no liability — and this as the very consequence of the defendant’s hazardous activity. But when the Supreme Court of Washington came to decide the New Meadows case, supra, it did not distinguish Siegler on this ground, perhaps realizing that the plaintiff in Siegler could have overcome the destruction of the evidence by basing a negligence claim on the doctrine of res ipsa loquitur. Instead it stressed that the transmission of natural gas through underground pipes, the activity in New Meadows, is less dangerous than the transportation of gasoline by highway, where the risk of an accident is omnipresent. We shall see that a further distinction of great importance between the present case and Siegler is that the defendant there was the transporter, and here it is the shipper.

Cases that impose strict liability for the storage of a dangerous chemical provide a potentially helpful analogy to our case. But they can be distinguished on the ground that the storer (like the transporter, as in Siegler) has more control than the shipper.

So we can get little help from precedent, and might as well apply section 520 to the acrylonitrile problem from the ground up. To begin with, we have been given no reason, whether the reason in Siegler or any other, for believing that a negligence regime is not perfectly adequate to remedy and deter, at reasonable cost, the accidental spillage of acrylonitrile from rail cars. Acrylonitrile could explode and destroy evidence, but of course did not here, making imposition of strict liability on the theory of the Siegler decision premature. More important, although acrylonitrile is flammable even at relatively low temperatures, and toxic, it is not so corrosive or otherwise destructive that it will eat through or otherwise damage or weaken a tank car’s valves although they are maintained with due (which essentially means, with average) care. No one suggests, therefore, that the leak in this case was caused by the inherent properties of acrylonitrile. It was caused by carelessness — whether that of the North American Car Corporation in failing to maintain or inspect the car properly, or that of Cyanamid in failing to maintain or inspect it, or that of the Missouri Pacific when it had custody of the car, or that of the switching line itself in failing to notice the ruptured lid, or some combination of these possible failures of care. Accidents that are due to a lack of care can be prevented by taking care; and when a lack of care can (unlike Siegler) be shown in court, such accidents are adequately deterred by the threat of liability for negligence.

It is true that the district court purported to find as a fact that there is an inevitable risk of derailment or other calamity in transporting “large quantities of anything.” This is not a finding of fact, but a truism: anything can happen. The question is, how likely is this type of accident if the actor uses due care? For all that appears from the record of the case or any other sources of information that we have found, if a tank car is carefully maintained the danger of a spill of acrylonitrile is negligible. If this is right, there is no compelling reason to move to a regime of strict liability, especially one that might embrace all other hazardous materials shipped by rail as well. This also means, however, that the amici curiae who have filed briefs in support of Cyanamid cry wolf in predicting “devastating” effects on the chemical industry if the district court’s decision is affirmed. If the vast majority of chemical spills by railroads are preventable by due care, the imposition of strict liability should cause only a slight, not as they argue a substantial, rise in liability insurance rates, because the incremental liability should be slight. The amici have momentarily lost sight of the fact that the feasibility of avoiding accidents simply by being careful is an argument against strict liability.

This discussion helps to show why Siegler is indeed distinguishable even as interpreted in New Meadows. There are so many highway hazards that the transportation of gasoline by truck is, or at least might plausibly be thought, inherently dangerous in the sense that a serious danger of accident would remain even if the truckdriver used all due care (though Hawkins and other cases are contra). Which in turn means, contrary to our earlier suggestion, that the plaintiff really might have difficulty invoking res ipsa loquitur, because a gasoline truck might well blow up without negligence on the part of the driver. The plaintiff in this case has not shown that the danger of a comparable disaster to a tank car filled with acrylonitrile is as great and might have similar consequences for proof of negligence. And to repeat a previous point, if the reason for strict liability is fear that an accident might destroy the critical evidence of negligence we should wait to impose such liability until such a case appears.

The district judge and the plaintiff’s lawyer make much of the fact that the spill occurred in a densely inhabited metropolitan area. Only 4,000 gallons spilled; what if all 20,000 had done so? Isn’t the risk that this might happen even if everybody were careful sufficient to warrant giving the shipper an incentive to explore alternative routes? Strict liability would supply that incentive. But this argument overlooks the fact that, like other transportation networks, the railroad network is a hub-and-spoke system. And the hubs are in metropolitan areas. Chicago is one of the nation’s largest railroad hubs. In 1983, the latest date for which we have figures, Chicago’s railroad yards handled the third highest volume of hazardous-material shipments in the nation. East St. Louis, which is also in Illinois, handled the second highest volume. With most hazardous chemicals (by volume of shipments) being at least as hazardous as acrylonitrile, it is unlikely — and certainly not demonstrated by the plaintiff — that they can be rerouted around all the metropolitan areas in the country, except at prohibitive cost. Even if it were feasible to reroute them one would hardly expect shippers, as distinct from carriers, to be the firms best situated to do the rerouting. Granted, the usual view is that common carriers are not subject to strict liability for the carriage of materials that make the transportation of them abnormally dangerous, because a common carrier cannot refuse service to a shipper of a lawful commodity. Restatement, supra, § 521. Two courts, however, have rejected the common carrier exception. If it were rejected in Illinois, this would weaken still further the case for imposing strict liability on shippers whose goods pass through the densely inhabited portions of the state.

The difference between shipper and carrier points to a deep flaw in the plaintiff’s case. Unlike Guille, and unlike Siegler, and unlike the storage cases, beginning with Rylands itself, here it is not the actors — that is, the transporters of acrylonitrile and other chemicals — but the manufacturers, who are sought to be held strictly liable. A shipper can in the bill of lading designate the route of his shipment if he likes, (a)(1), but is it realistic to suppose that shippers will become students of railroading in order to lay out the safest route by which to ship their goods? Anyway, rerouting is no panacea. Often it will increase the length of the journey, or compel the use of poorer track, or both. When this happens, the probability of an accident is increased, even if the consequences of an accident if one occurs are reduced; so the expected accident cost, being the product of the probability of an accident and the harm if the accident occurs, may rise. Glickman, Analysis of a National Policy for Routing Hazardous Materials on Railroads (Department of Transportation, Research and Special Programs Administration, Transportation Systems Center, May 1980). It is easy to see how the accident in this case might have been prevented at reasonable cost by greater care on the part of those who handled the tank car of acrylonitrile. It is difficult to see how it might have been prevented at reasonable cost by a change in the activity of transporting the chemical. This is therefore not an apt case for strict liability.

We said earlier that Cyanamid, because of the role it played in the transportation of the acrylonitrile — leasing, and especially loading, and also it appears undertaking by contract with North American Car Corporation to maintain, the tank car in which the railroad carried Cyanamid’s acrylonitrile to Riverdale — might be viewed as a special type of shipper (call it a “transporter”), rather than as a passive shipper. But neither the district judge nor the plaintiff’s counsel has attempted to distinguish Cyanamid from an ordinary manufacturer of chemicals on this ground, and we consider it waived. Which is not to say that had it not been waived it would have changed the outcome of the case. The very fact that Cyanamid participated actively in the transportation of the acrylonitrile imposed upon it a duty of due care and by doing so brought into play a threat of negligence liability that, for all we know, may provide an adequate regime of accident control in the transportation of this particular chemical.

In emphasizing the flammability and toxicity of acrylonitrile rather than the hazards of transporting it, as in failing to distinguish between the active and the passive shipper, ultrahazardousness or abnormal dangerousness is, in the contemplation of the law at least, a property not of substances, but of activities: not of acrylonitrile, but of the transportation of acrylonitrile by rail through populated areas. Natural gas is both flammable and poisonous, but the operation of a natural gas well is not an ultrahazardous activity. Whatever the situation under products liability law (section 402A of the Restatement), the manufacturer of a product is not considered to be engaged in an abnormally dangerous activity merely because the product becomes dangerous when it is handled or used in some way after it leaves his premises, even if the danger is foreseeable. The plaintiff does not suggest that Cyanamid should switch to making some less hazardous chemical that would substitute for acrylonitrile in the textiles and other goods in which acrylonitrile is used. Were this a feasible method of accident avoidance, there would be an argument for making manufacturers strictly liable for accidents that occur during the shipment of their products (how strong an argument we need not decide). Apparently it is not a feasible method.

The relevant activity is transportation, not manufacturing and shipping. This essential distinction the plaintiff ignores. But even if the plaintiff is treated as a transporter and not merely a shipper, it has not shown that the transportation of acrylonitrile in bulk by rail through populated areas is so hazardous an activity, even when due care is exercised, that the law should seek to create — perhaps quixotically — incentives to relocate the activity to nonpopulated areas, or to reduce the scale of the activity, or to switch to transporting acrylonitrile by road rather than by rail, perhaps to set the stage for a replay of Siegler v. Kuhlman. It is no more realistic to propose to reroute the shipment of all hazardous materials around Chicago than it is to propose the relocation of homes adjacent to the Blue Island switching yard to more distant suburbs. It may be less realistic. Brutal though it may seem to say it, the inappropriate use to which land is being put in the Blue Island yard and neighborhood may be, not the transportation of hazardous chemicals, but residential living. The analogy is to building your home between the runways at O’Hare.

The briefs hew closely to the Restatement, whose approach to the issue of strict liability is mainly allocative rather than distributive. By this we mean that the emphasis is on picking a liability regime (negligence or strict liability) that will control the particular class of accidents in question most effectively, rather than on finding the deepest pocket and placing liability there. At argument, however, the plaintiff’s lawyer invoked distributive considerations by pointing out that Cyanamid is a huge firm and the Indiana Harbor Belt Railroad a fifty-mile-long switching line that almost went broke in the winter of 1979, when the accident occurred. Well, so what? A corporation is not a living person but a set of contracts the terms of which determine who will bear the brunt of liability. Tracing the incidence of a cost is a complex undertaking which the plaintiff sensibly has made no effort to assume, since its legal relevance would be dubious. We add only that however small the plaintiff may be, it has mighty parents: it is a jointly owned subsidiary of Conrail and the Soo line.

The case for strict liability has not been made. Not in this suit in any event.

Indiana Harbor Belt R. Co. v. American Cyanamid Co., 916 F.2d 1174, 1176-1182 (7th Cir. 1990).


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: