Concept
Weighing Risk and Utility
Duty and negligent conduct framework
Weighing Risk and Utility
Negligence, in the Restatement (Second) of Torts, is conduct that falls below the standard the law sets for protecting others against unreasonable risk of harm (§ 282), and a negligent act is one the actor, as a reasonable person, should recognize as involving an unreasonable risk of invading another's interest (§ 284(a)).1 Both definitions turn on the word unreasonable. Duty settles whether the actor owed care to this plaintiff, and the reasonable person standard settles that conduct is measured objectively, yet neither says when a particular risk crosses the line. Courts draw the line by comparison: they set the danger the conduct created against the cost of avoiding it and against the value of the conduct.
Judge Learned Hand's formulation in United States v. Carroll Towing Co. remains the classic version of that comparison. A risk is unreasonable when the foreseeable probability and gravity of the harm outweigh the burden to the actor of alternative conduct that would have prevented it.2 Probability and gravity together fix the magnitude of the risk, and the burden of precautions sits on the other side of the scale. Courts reach for the same considerations when they restate a duty. In Sargent v. Ross, the court held that landlords, like other persons, must exercise reasonable care not to subject others to an unreasonable risk of harm, and it told the landlord to act as a reasonable person would under all the circumstances, including the likelihood of injury, the probable seriousness of any injury, and the burden of reducing or avoiding the risk.3
Foreseeability limits negligence liability without supplying it. A defendant is liable only for a risk that was both reasonably foreseeable and unreasonable.4 Nearly everything people do carries some foreseeable chance of harm to others, so a rule that condemned every foreseeable risk would condemn ordinary, useful activity.
The four subsections below take the comparison apart: the danger the conduct creates, the cost of reducing that danger, the value of the conduct, and the danger to persons other than the plaintiff.
Probability and Gravity of Harm
The magnitude of a risk, which some opinions call the hazard or the danger, is the foreseeable probability that harm will result combined with the foreseeable gravity of the harm if it does (Restatement (Second) § 293).5 Neither quantity measures danger alone. A chance of harm says nothing about whether the harm would be a bruise or a death, and a terrible harm says nothing about whether it is a realistic prospect or a remote one. Together they fix how much danger the conduct creates, and the amount of care that reason requires generally rises with that danger.6
A probability below even odds does not end the inquiry. One classic statement puts the point this way: “The test is not of balance of probabilities, but of the existence of some probability of sufficient moment to induce action to avoid it on the part of a reasonable mind.”7 A person who crosses a railroad track does not reasonably expect to meet a train, but because one may come, he must do what is reasonably required to find out. A driver on a sharp turn with an obstructed view may be careless toward oncoming traffic even though the odds are against meeting any one car.8 Tullgren v. Amoskeag Mfg. Co. applies the idea. A company escorted an employee suffering a “sick headache” part of the way home and left her to walk about 700 feet over a difficult road, and she died on reaching her house. Reversing a directed verdict for the company, the court said on rehearing that danger consists in the risk of harm as well as the likelihood of it, that a danger calling for anticipation need not be more probable than not, and that if there is some probability of harm serious enough that ordinary people would take precautions against it, failing to do so is negligence.9
The same point corrects a common misreading of opinions. Courts often say that injury was the “natural and probable consequence” of the defendant's act, but “probable” there means foreseeable, and a risk may be foreseeable and worth avoiding even though the chance that it will come to pass is far below fifty percent.10 On an exam, a probability under fifty percent therefore does not end the analysis. Odds also compound across repetitions, so a chance that is small on any one occasion can be high over the life of something used constantly. A car maker, for example, cannot say that it does not intend its cars to be involved in accidents when it can easily foresee, and knows that the probability over the life of its product is high, that the product will be involved in some injury-producing accident.11
The probability that negligence requires also sits far below the certainty that intent requires. Intent has traditionally included knowledge, so a person intends consequences when he desires them or knows they are substantially certain to result.12 Knowledge and knowing intention require consciousness of almost-certainty, while recklessness requires consciousness of something far less than certainty or even probability.13 Negligence is a separate category that does not include conduct recklessly disregardful of another's interest,14 and it asks what the actor should have recognized, not merely what he consciously perceived.
Gravity can outweigh a low probability. It may be negligence to dash across a little-traveled highway, though the chance of meeting a car is slight, because a collision would be so grave.15
Gravity does not decide every case, though. In Hauser v. Chicago, Rock Island & Pacific Railway, a passenger who fainted fell with her face against an uninsulated steam pipe that heated a coach's toilet and ran under the wash basin and water cooler, and she was seriously burned. The court held that leaving the pipe uninsulated was not unreasonable.16 The result is consistent with a balance that gives weight to how unlikely it was that anyone would fall against a pipe in that spot.
What the actor knew is a third variable. The actor is charged with what he actually knew and perceived as well as with what he ought to have known and perceived, and the ultimate issue is the reasonableness of his conduct, not the reasonableness of his belief that he was acting with care.17 Hand a loaded gun to a person you know to be mentally unstable and you create an unreasonable risk. Hand the same gun to the same person with no reason to suspect anything, and a reasonable person in your position would have seen nothing to weigh. Obvious dangers cut the other way. A reasonable person may expect others to avoid a danger that is quite obvious, and may expect a third person, such as a parent, to protect a child, and these are questions of fact for the jury rather than rules of law.18 The protection has a limit, because § 284(a) measures an act by what the actor as a reasonable person should recognize, so an actor who overlooks a warning that any reasonable person would heed is charged with what it would have revealed.19
Because probability, gravity, and knowledge interact, no single percentage marks the line between reasonable and unreasonable risk. The acceptable chance shifts with the seriousness of the harm, with what the actor knew, and with the burden and value on the other side of the scale.
Burden of Precautions
A person acts negligently, under § 3 of the Restatement (Third) of Torts: Liability for Physical and Emotional Harm, if the person does not exercise reasonable care under all the circumstances, and the primary factors are the foreseeable likelihood of harm, the foreseeable severity of any harm, and the burden of precautions to eliminate or reduce the risk.21 The first two factors measure how dangerous the conduct was. The third measures what it would cost to make the conduct less dangerous, whether in money, in effort, or in lost benefits.
Hand turned the comparison into arithmetic. The actor is negligent when the burden of adequate precautions is outweighed by the probable gravity of the harm multiplied by the probability that it will occur.22 A $200 precaution that eliminates a one-in-a-thousand chance of a $1,000,000 loss buys $1,000 of expected safety, so a reasonable actor takes it. The same precaution priced at $5,000 buys less than it costs, and a reasonable actor may decline it. Real cases seldom supply figures this clean, but the figures show the shape of the judgment a jury makes.
The type and amount of care that reason requires are generally a function of the type, likelihood, and degree of harm: a great risk normally calls for great precautions, and a small risk calls for small ones.23 Requiring only precautions that are worth their cost serves two ends. A precaution that saves less than it costs wastes resources that could have prevented greater harm elsewhere. And if every precaution that might have prevented an accident were required whatever its cost, every accident would prove negligence, and negligence would become insurance.
The burden is measured at the level of the rule being proposed, a point easy to miss because the cheap precaution is the first thing a plaintiff notices. In Lucero v. Holbrook, the plaintiffs argued that the burden on the owner of a car left running in her driveway was low, since she could easily have removed her keys and locked the car before returning to her house. The court affirmed summary judgment for the owner. It explained that the burden at issue was not the effortlessness of removing the keys but the burden of imposing a duty not to leave a motor running temporarily in a driveway, a duty that would apply across the board and create potential liability for every person in Wyoming who warms up a car on a cold winter day.24
That measure follows from what the court was deciding. Lucero treated the burden as part of whether a duty exists, and a duty rule governs every later case of the same type. Looking only at one defendant's ease would let hindsight make almost any precaution look trivial. The same factor appears among the primary considerations on breach under § 3, so the burden analysis can run at either stage, and an answer should follow the label the question and the jurisdiction use.25
The tenant and the Lucero owner could each have avoided the harm with a trivial step, yet the cases differ in the rule each step implies. Forbidding heavy objects to be heaved from upper windows burdens almost no one's ordinary activity and guards against a grave harm, while forbidding cars to idle for a few minutes in a driveway burdens a common practice across the whole community.
The tenant's shout was itself a precaution, and it is the cheapest kind. Warnings sit at the light end of the burden scale because adding words to a label costs little, and failure-to-warn cases such as Moran v. Faberge, Inc., in which a perfume maker did not warn that its product was flammable, are the typical setting. But a warning lowers the chance of harm only to the extent that the person warned notices it and acts on it, and it does nothing to reduce how serious the harm would be. When the risk is great, reasonable care normally requires great precautions,26 so a cheap warning cannot by itself make an unreasonably dangerous act reasonable when a safer way of doing the same thing was also cheap. Design cases carry the same weighing to manufacturers, who must compare the burdens of safety precautions they did not adopt with the safety benefits those precautions would have achieved,27 as in the Texas Supreme Court's products decision in Genie Industries v. Matak.
Because the care required scales with the magnitude of the risk, the same factual theory can require more or less proof depending on what the precaution costs. A plaintiff who says a store should have posted armed guards in its parking lot is demanding an expensive precaution, which a substantial foreseeable risk must justify. A plaintiff who says the store should have installed better lighting is demanding an inexpensive one, which a smaller risk can justify.28
Utility of the Activity
Negligence condemns only an unreasonable risk,30 so some risk to other people must be reasonable, and part of the sorting depends on what the risky conduct is worth. Utility reaches the balance through the burden side. A precaution that would force the actor to give up a valuable activity carries that loss as part of its cost, which is why the Lucero court counted the effect of a duty on everyone who warms up a car before driving.31 The primary factors in § 3 of the Third Restatement do not name utility separately, so an answer should treat it as part of the single balance, not as a fourth factor.
A risk can be tiny and still unreasonable when nothing is gained by running it. If there is no social utility in what the actor is doing, even a chance of harm below one percent can be enough for recklessness, a more culpable category than negligence.32 Suppose a person who knows that exactly one of a thousand pistols on a table is loaded picks one at random and fires it at another person. The chance of killing is one in a thousand, but the conduct buys nothing, so the risk is unreasonable.
The reverse also holds, because value can justify added danger that triviality could not.
The value that counts is social value, weighed impartially. A design that always protects passengers at the expense of bystanders has no social utility if the only reason for it is to privilege consumer interests over others, because tort law requires impartial consideration of consumers and bystanders and so makes the private utility consumers gain from unequal treatment irrelevant.33
Utility also has a ceiling, because the value of an activity is finite while the harm side of the scale grows with the seriousness of the injury and the number of people exposed.
Sports show another limit on what value can buy. A limited-duty rule among players, if it stays true to its consensual heritage, must follow the participants' reasonable expectations: professional football players must expect some rather serious misbehavior, but ping-pong players should not expect to be blinded by a thrown paddle.35 Hunters know that they are at risk from other hunters, yet given the seriousness of the potential injury they should not be viewed as consenting to being shot by negligent gunners.36 A valued activity does not license every risk it generates.
Land disputes reach the same questions in slower, costlier form. In Noone v. Price, the plaintiffs bought a mountainside house in Glen Ferris, West Virginia, that Union Carbide had built in 1928 or 1929, and four years later they found that the wall under their front porch was giving way and the living-room plaster had cracked.37 Whoever is blamed for such a loss will face the questions this chapter has developed: how likely and how grave the harm was, what avoiding it would have cost, and what the conduct was worth.
Steps two and three are the ones students most often skip. An answer that stops at “the activity was socially useful” has found the value without testing the manner, and an answer that ignores value in an ambulance run or an emergency repair overstates the risk by leaving out the one fact that makes it tolerable.
Risks to Third Persons
In deciding whether conduct is negligent toward one person, the fact that it involves an unreasonable risk of harm to a third person may in certain situations be considered (§ 294).39 When the actor's only alternative is a course of conduct involving an equal or greater risk of harm to a third person, that fact is a factor in determining the reasonable character of the actor's conduct (§ 295).40 The earlier subsections weighed probability, gravity, burden, and value, usually with a single person at risk, the plaintiff. These two sections widen the picture to everyone the conduct endangers. Both speak of what may be considered or what counts as a factor, so each tells the jury what belongs on the scale and leaves the jury to decide how the scale tips.
A third person's risk can belong in the balance because the balance measures the total danger the conduct creates, and that danger does not sort itself into separate packages for each person it might reach. A precaution also seldom protects only one person. A repair, a guard, or a decision to stop an activity usually removes the danger for everyone exposed at once, so counting only the plaintiff's share would understate what the precaution was worth. Design cases press the same point about bystanders: the duty of reasonable care calls for impartial consideration of consumers and bystanders alike.41
Section 294 speaks of third-person risk in “certain situations,” so an answer should not claim that any risk to anyone makes conduct negligent toward the plaintiff. The rule leaves untouched the separate requirement that the conduct be negligent with respect to the plaintiff or a class of persons that includes the plaintiff.42 In the plant illustration the worker stands in the path of the vapor and so belongs to the class the leak endangers. A person who could never have been in that path would not become a proper plaintiff merely because others were.
Section 295 addresses a different shape of problem, in which the actor faces a dilemma and every available course endangers someone. A precaution is worth requiring only when what it costs is less than the harm it prevents. When the only alternative endangers a third person equally or more, taking it would not reduce the total danger but would move it to someone else, so it buys no safety to set against its cost. Design law states the same idea: a design that eliminates one hazard only to create another of equal or greater magnitude is not a reasonable alternative, because the overall safety of the product must be considered.44 The rule is also fair to the actor, since a standard that condemned every available course would give the actor no way to comply.
A driver who had a clear, straight path on which to brake and stop safely, and who swerved onto the shoulder out of inattention, gets no help from § 295. The section reaches an actor whose only alternative carried an equal or greater risk to third persons, and this driver had a safe one. A defendant who invokes the factor must therefore show that the alternatives were truly forced, and a choice among dangerous courses the actor did not have to face is judged by the ordinary balance of risk, burden, and utility. Even when the alternative was forced, the section makes the fact a factor, so the jury still asks whether the choice was reasonable for someone deciding under those conditions. The next section takes up sudden emergencies, and one point belongs here: an actor who is careful once the emergency has arisen may still be liable for the tortious conduct that produced it.45