Concept
Policy Limits Within Foreseeability
Proximate Cause: Foreseeability and Scope of the Risk
Policy Limits Within Foreseeability
Courts limit negligence liability in two ways, and only one of them turns on foresight. The first denies recovery for harm a reasonable person could not have anticipated, the scope-of-the-risk limit of the earlier sections. The second denies recovery for harm that was foreseeable, because policy counsels against stretching the defendant's responsibility that far.1 Foreseeability cannot overcome the second limit, since it only limits liability and never creates it: a risk must be unreasonable as well as foreseeable before negligence liability attaches.2 The definition of a tort shows how foreseeable harm can go uncompensated. A tort is a civil wrong in which one person's conduct causes a compensable injury to another “in violation of a duty imposed by law,” and a court may decide that the law imposes no duty.3
Courts apply these labels inconsistently. Palsgraf was a dispute about foreseeability, yet Cardozo resolved it as a question of duty and Andrews as one of proximate cause. Courts that withhold recovery for foreseeable harm on policy grounds often write in proximate-cause language, as the early liquor-service cases did when they called the drinker's driving, not the service of liquor, the “proximate cause” of the injury.4 The reader's task is to look past the label to the ground the court actually relied on.
The fire-spread rule is the oldest example. In Ryan v. New York Central R.R. Co., the court refused recovery for property destroyed as a fire spread beyond the first building, speaking of “remote” and “immediate” causes.5 Its reasoning was candid about the burden. The court warned that following such a claim to its legitimate consequences would expose the defendant to a liability “against which no prudence could guard, and to meet which no private fortune would be adequate.” It added that a man may insure his own house or furniture but cannot insure his neighbor's building or furniture, since he has no interest in them.6 Remoteness was the label, and the burden was the reason. Anyone could foresee that a fire would burn beyond the next lot. The true rationale was the crushing burden such liability would impose, together with the general availability of fire insurance as an alternative source of protection, and one commentator, borrowing Dean Prosser's phrase, calls the proximate-cause analysis “moonshine and vapor.”7
The rule did not hold. New York soon allowed recovery to the first adjoining landowner, then to the first property the fire jumped to even if it did not adjoin, and other courts rejected the first-building test and held defendants liable for fires that spread a substantial distance.8 The retreat shows how contestable the policy judgment was: the same foreseeable harm falls inside or outside liability according to where a court decides the burden becomes unmanageable.
Secondary Economic Loss
A negligently started factory fire injures one worker and closes the plant for three months. The injured worker recovers for the injury. Unhurt coworkers who lose three months of wages do not, although a fire that damages a plant may well close it.9 Foresight cannot separate the two losses; the burden of liability does. Courts reason that liability for such secondary economic losses is too great, and the cases frequently say the defendant's conduct was “not the proximate cause” of them even though the denial has nothing to do with proximate cause. Imagine the exposure had the defendant burned down a bridge and interrupted the business affairs of an entire town.10
Insurance supplies the second half of the reasoning. Open-ended liability would make third-party liability insurance virtually impossible to administer, and reasonable limits are necessary for the system to work, whereas each affected person can insure against his own finite and readily discernible loss through first-party coverage.11 The general principle accordingly denies recovery for economic loss unaccompanied by physical damage. Suppose a negligent driver's chemical spill closes a highway for a week, and a travel agency whose clients cannot reach its office loses a week of business. The agency suffered no injury to its people and no damage to its property, and it recovers nothing.12
Commercial fishermen are the recognized exception. A longstanding line of cases awards them lost profits when pollution or another casualty disrupts their activities, because of their “particular and special nature” and because their losses from an event such as an oil spill are especially foreseeable and a direct consequence of it.13 The exception fits the rationale, since the worry about an indeterminate burden is weakest where the plaintiffs are a defined and predictable group.
Courts do not always name the burden as their ground. In Consolidated Aluminum Corp. v. C.F. Bean Corp., a dredger broke a gas main, and a manufacturer that depended on the main lost equipment and production when its gas stopped. The court denied recovery, treating the loss as beyond the harm a negligent dredger could reasonably have anticipated. That reasoning is hard to defend as foresight, because cutting a main foreseeably interrupts those who depend on it. Defining the risk is the court's tool here, and it can fail in either direction. Pitched too generally, it would let the plaintiff who tripped over a box lid, missed her train, and was hurt when a later train crashed recover, since the lid created a risk of personal injury and she suffered personal injury.14 The court must focus on the particular type of harm the defendant's act threatened, without bogging down in the idiosyncratic details of the accident.15 A candid opinion would rest on the burden of liability for secondary losses and call the limit a matter of policy, because courts that restrict liability within the circle of foreseeability for policy reasons often mask the decision in proximate-cause language.16
People Express Airlines, Inc. v. Consolidated Rail Corp. shows one court trying to relax the bar. It asks whether the plaintiffs form an identifiable class and stresses that an identifiable class is not simply a foreseeable one. Members of the public delayed near a negligently caused accident are certainly foreseeable, but their presence is fortuitous and the varied economic losses they suffer are hopelessly unpredictable, so the class is not sufficiently ascertainable.17 The line drawn even there is a policy line, because it asks how predictable the plaintiffs and their losses were, not whether the loss could have been foreseen.
Alcohol Service and Other No-Duty Rules
The early liquor-service cases show a policy limit dressed as proximate cause. The common-law rule held that one who sold or furnished intoxicating beverages was not liable to third persons later injured by the intoxicated person, because the damage was too remote: consumption, not the sale, was the immediate cause of the harm.18 In reality the rule was a no-duty rule based on public policy.19 Few acts are more foreseeable than a patron driving home from a roadside bar. The real thrust of the cases was a policy decision to place liability for obviously foreseeable harm on the more blameworthy of two negligent parties, and many jurisdictions have since repudiated them.20 Stated as a general rule for every sale to an intoxicated patron, and applied even though the harm was the very risk the sale created, the rule could not have been a finding about foresight.
The modern law keeps the policy question in view. Most courts hold that a bartender in a commercial bar owes third parties a duty not to serve a patron too much alcohol, while they hold, basically as a matter of policy, that a social host owes no such duty. Virtually all refuse to hold a host liable for serving an obviously drunk guest even when the host knows the guest means to drive soon.21 The danger is no less foreseeable at the host's table than at the bar, so foresight cannot explain the difference. Only when the duty question is answered yes, and the plaintiff proves breach, damages, and cause in fact, does the proximate-cause issue arise.22 A few courts, with Kelly v. Gwinnell the familiar example, have recognized a duty in social hosts, and because the danger is equally foreseeable on both sides, the dispute is over policy.
Policy reasoning can also lower a barrier. A court responding to DES plaintiffs, who generally cannot identify the manufacturer of the drug that injured them, observed that the legislature and the court had removed legal barriers to recovery, barriers which “may have had their place in other contexts, but which in DES litigation worked a peculiar injustice” because of how DES was developed, marketed, and sold and because of the insidious nature of its harm.23
Other categories follow the same pattern. Courts have limited the scope of duty despite obvious foreseeability of harm, for reasons of policy that sometimes resemble those operating in proximate cause. A firearms manufacturer, for example, has no duty to the victim of a shooting even when it knowingly permits distribution under circumstances that promote illegal sales to third parties likely to misuse the guns.24 A third party's crime does not by itself explain such a rule. In Kentucky, criminal conduct no longer automatically interrupts the causal link, because a spark that ignited negligently accumulated trash may have been foreseeable whether it was set accidentally, negligently, intentionally, or even criminally.25 In certain categories of cases courts hold that the duty owed is less than reasonable care or none at all, and the traditional duty of a possessor of land to a visitor was in many cases something less than reasonable care.26 Statutes can draw comparable lines, as when a statutory immunity, deemed inapplicable to highly developed recreational areas, was held to cover a lake but not the dam that created it.27
An opinion that calls a plainly foreseeable loss “not proximately caused” is naming the court's reason by the wrong label, and an answer that treats the phrase as a finding of unforeseeability will misstate the holding.