Concept
Applying the Balance
Risk-utility factors in judging unreasonable conduct
Applying the Balance
Restatement (Second) of Torts § 291 poses a single comparison. An act that a reasonable person would recognize as involving a risk of harm is negligent if the magnitude of that risk outweighs what the law regards as the utility of the act, or of the particular manner in which it is done.1 Sections 292 and 293 list the inputs to the two sides of the comparison.2 Because the comparison is a single one, the factors are weighed together and none works as a threshold. Large utility can justify a heavy risk, a trivial benefit can fail to justify a modest one, and a remote chance of grave harm can outweigh a likely chance of slight harm. Students who can recite every factor still misapply the balance in three recurring ways: they demand that harm be more probable than not, they assume that dangerous activities carry a stricter standard of their own, and they judge the actor by what the accident later revealed. The subsections below correct each error in turn.
Low Probability, Grave Harm
Nothing in § 293 sets a minimum chance of harm. Subsection (b) makes the extent of the chance that the actor’s conduct will invade the plaintiff’s interests one input to the magnitude of the risk, and § 291 then compares that magnitude, together with the other inputs, against the utility of the conduct.3 A risk can therefore be unreasonable long before harm becomes likelier than not. The treatise puts the point flatly: no mathematical rule of percentage applies; a risk is not necessarily unreasonable because harm is more likely than not to follow, nor reasonable because the chances are against it; and a very large risk may be reasonable in some circumstances and a small risk unreasonable in others.4
Students drift toward a probability threshold because “more probable than not” belongs to a different rule. A civil plaintiff must prove each element of the claim by a preponderance of the evidence, and that burden governs how convincing the proof of past events must be. The standard of care looks at another moment. It asks what a reasonable person would have done in choosing how to act, before anyone knew whether harm would follow. Importing the first rule into the second strips out the quantities the balance exists to weigh.
Passing on a blind rise shows the difference. A driver who pulls out where the view ahead is hidden will probably meet no one, and the chance of meeting a car at that spot may be small. Even so, a driver going around a sharp, obstructed turn may be careless toward oncoming traffic “though the probabilities may be against meeting any one,” because if the chance is so great that ordinary people would drive differently, it is careless not to do so.5 The Hand formulation explains why. A risk is unreasonable when the foreseeable probability and gravity of the harm outweigh the burden of the alternative conduct that would have prevented it.6 On the rise, the probability is low, the gravity of a head-on collision is great, and the burden is a few seconds spent waiting for a clear stretch. A trial court that told the jury the driver was negligent only if a collision was more probable than not would misstate the law.
Gravity does as much work as probability. Hold the chance at 1 percent and vary the worst case. If the maneuver risks a scraped bumper, a few seconds of delay may not be worth buying, but if it risks a death, the same chance weighs far more, because probability and gravity of harm together determine the magnitude of the risk.7 A threshold of even odds would also erase the burden of precaution from the analysis, since a cheap protection against a rare catastrophe would never be required. And because the odds on any single occasion would be long, the gamble would never count as negligent, even on the occasion when the harm arrives.
The treatise’s discussion of recklessness shows how far below even odds the law is willing to go. Where conduct has no social utility, one might be reckless though the chances of harm are something less than 1 percent, because recklessness requires a consciousness of something far less than certainty or even probability.8 If that holds for the more culpable state of mind, an even-odds threshold cannot be the test for negligence.
Probability misleads in the opposite direction too. Consider a railroad’s decision not to build an overpass. Suppose two injuries will occur each decade, the average injury costs $500,000, and the overpass will last 30 years. The expected injury cost is $3 million, set against $12 million to avert the six likely accidents. In Hand formula terms the decision may well be reasonable, because the investment may exceed what society, through tort law, is willing to require.9 Harm is expected in that example, and the conduct may still be reasonable. Hauser v. Chicago, Rock Island & Pacific Railway shows that a serious injury does not settle the question either. A passenger fainted, fell with her face against an uninsulated steam pipe under a wash basin, and was seriously burned, yet the court held it not unreasonable to leave the pipe uninsulated.10 A low probability therefore opens the analysis without closing it.
When a fact pattern stresses how unlikely the accident was, treat that as the start of the analysis. Ask how grave the harm would have been, how many people it could reach, and what the omitted precaution would have cost. The plaintiff must still prove by a preponderance that the defendant’s conduct fell short of the balance and caused the harm, but the odds of the accident do not measure that shortfall.
Care Proportionate to Risk
The comparison in § 291 stays fixed while its answer moves, because the magnitude of a risk changes with the circumstances. A court usually answers it by balancing the burdens of the safety precautions the defendant failed to adopt against the safety benefits those precautions would have achieved.12 The care that reason requires is generally a function of the type, likelihood, and degree of harm, viewed together as the magnitude of the risk, that precautions of a particular cost may be expected to prevent. If the risk is great, reasonable care normally requires great precautions, and if it is small, ordinarily only small ones.13 The amount of care a reasonable person takes therefore rises with the danger, and that rise comes from one standard applied to different facts.
Lawyers often reach instead for a formulation under which dangerous activities carry a heightened or highest-degree standard, and that formulation misdescribes the law. Suppose a plaintiff burned when gasoline overflowed at a filling station asks the judge to tell the jury that the defendant owed the highest degree of care. The judge should decline the label and instruct that the standard is reasonable care under the circumstances, and that when the risk is high and a reasonable person would have recognized it, the circumstances call for care commensurate with that risk.14 The two instructions sound alike, and a jury would probably hear both as an order to be very careful. They differ in what they say about the law: the requested instruction describes a different standard, while the correct one describes how the single standard applies.
Two features of the balance explain why one standard is enough. First, the balance already scales. Because the care reason requires depends on the type, likelihood, and degree of harm, a graver risk demands more precaution without any second tier.15 Second, a highest-degree label overshoots, since only reasonable precautions need be taken, not every conceivable or possible precaution.16 The burden of precaution stays on the scale opposite the risk. In the overpass example, $12 million to avert injuries expected to cost $3 million may be more than society, through tort law, is willing to require.17
The circumstances shape the standard from the inside and never replace it. A professional is expected to possess and employ the skill and knowledge of her profession, yet specialized knowledge does not subject her to a higher standard: her standard, like everyone else’s, is reasonable care under the circumstances. The facilities or resources available to the actor enter in the same way. Exploratory surgery may be reasonable where less invasive diagnosis is unavailable but not where it is available.18
Foresight and Knowledge
The balance in § 291 begins with an act that a reasonable person would recognize as involving a risk of harm, so only a risk recognizable when the actor chose how to behave enters the comparison.20 That starting point fixes the moment of judgment. The court asks what a reasonable person in the actor’s position would have recognized then, and it sets aside what only the accident revealed. Hindsight would otherwise distort the inquiry, because a plaintiff always sues after the harm and nearly every risk looks obvious after it. Liability would follow luck: the careful actor who suffered a freak accident would pay, and the careless actor who escaped would not. The Hand formulation is accordingly stated in terms of the foreseeable probability and gravity of the harm.21
Blyth v. Birmingham Waterworks Co. shows the rule at work. A reasonable person acts with reference to the average circumstances of the temperature in ordinary years, and the defendants had provided against such frosts as experience would have led prudent people to provide against. They were not negligent because their precautions proved insufficient against the extreme frost of 1855, which penetrated to a greater depth than any which ordinarily occurs south of the polar regions. That frost was a contingency against which no reasonable person can provide.22 The failure of the precautions proved nothing about negligence. The question was whether the frost lay within what the defendants’ experience gave them reason to expect, and it did not. The same inquiry explains why place matters. Whether a contractor building a skyscraper must guard against an earthquake may be answered differently in Chicago than in San Francisco, because what experience gives a reasonable builder reason to expect differs between the two cities.23
What a reasonable person would recognize depends on the information and capacities the actor had. An actor with specialized knowledge is expected to use it, and the resources available to the actor count as well.24 The foreseeable misconduct of others counts too, which is why it may be negligence to leave a car unlocked with the keys in the ignition, or to fail to slow down near a schoolyard where children may negligently dart into the street.25 Sudden incapacity applies the rule to the actor’s own body. A person who suddenly loses consciousness from a physical illness and injures another is negligent only if the actor could reasonably foresee the sudden incapacitation or loss of consciousness.26
Foresight measures the risk objectively and does not turn on what the actor hoped or intended. An actor cannot defend by proving that he meant to act carefully, because the ultimate issue is the reasonableness of his conduct, not the reasonableness of his belief that he was acting with appropriate care. The rule cuts the other way as well: if the conduct did not subject the plaintiff to an unreasonable risk, the actor is not negligent even though he was totally oblivious to the safety of others.27
Foreseeability admits a risk to the scale without deciding the weighing. Once a reasonable person would recognize a risk, § 291 still requires comparing its magnitude with the utility of the conduct, and a foreseeable risk may come out reasonable.28 A driver who speeds to carry a badly injured person to a hospital foresees the added danger to others, yet the social value of the interest the conduct advances may justify it.29 A test that stopped at foreseeability would condemn nearly every accident, because almost any harm can be described as something that might have been anticipated. Foresight decides which risks enter the scale, and the factors of §§ 292 and 293 decide whether those risks outweigh the utility.