T
HE PRECEDING CHAPTER introduced the fundamental concepts of tort law and
developed an economic analysis of tort liability. In this chapter we wish to advance
the understanding of the economic analysis of the tort liability system in two ways.
First, we relax some simplifying assumptions in order to bring the model closer to reality.
Second, we shall examine some arguments that the tort liability system does not work
well and needs thorough reform. In the course of this examination, we shall look at some
recent evidence on how well the tort system minimizes the social costs of accidents.
I. Extending the Economic Model
The model that we introduced in the last chapter made some implicit simplifying
assumptions. The grand tradition in economics would have us assert our intention to
relax these simplifications but then forget to do so. But we aspire to do better. We turn
immediately to the task of exploring the conclusions of our simple model when we relax
our simplifying assumptions.
A. Relaxing the Core Assumptions
In the previous chapter we implicitly made five simplifying assumptions before we
developed our economic analysis of tort law:
1. Decision makers are rationally self-interested.
2. There are no regulations designed to reduce external costs.
3. There is no insurance.
4. All injurers are solvent and pay damages in full.
5. Litigation costs are zero.
The purpose of this section is to relax these assumptions and to see the effect, if
any, on the conclusions from the economic theory of tort liability.
1. Rationality One of the central assumptions in economic theory is that decision
makers are rationally self-interested. As a technical matter, this means (as we saw in
our review of microeconomics in Chapter 2) that decision makers have stable, wellordered
preferences,1
which implies something about the decision maker’s cognitive
230
7 Topics in the Economics
of Tort Liability
1
Recall that such preferences are stable in the sense that they do not change too rapidly or quixotically and
that they are well ordered in the sense that they are transitive, which means that if A is preferred to B and B
is preferred to C, then A must be preferred to C.
I. Extending the Economic Model 231
and reasoning abilities. Specifically, it suggests that decision makers can calculate the
costs and benefits of the alternatives available to them and that they choose to follow
the alternative that offers the greatest net benefit.
There is a vital connection between the assumption of rationality and the economic
model of tort liability in Chapter 6. We saw that the rules for assigning tort liability
are, economically speaking, designed to send signals to potential victims and
potential injurers about how they ought to behave. For the tort liability system to have
this effect, it must be the case that those whose behavior the law is seeking to affect
are rational: If they do not behave in the manner predicted by the assumption of rational
self-interest, then we may need to amend the tort liability system in light of how
people actually behave.
But do people really make decisions about potential liability in this way? Some
people do, and others do not. Recent academic literature suggests that many decision
makers commit predictable errors in making calculations of the sort that tort liability
encourages them to make. For instance, Kahneman and Tversky report two disturbing
conclusions.2
First, they find that most people simply cannot accurately estimate lowprobability
events; they seem to deal with them by assuming that “low probability”
means that the event will never happen—that the probability of the event’s happening
is zero. Second, they find that, for some well-publicized, potentially catastrophic
outcomes—such as accidents from nuclear power plants—most people systematically
exaggerate the probability of an accident’s occurring, regardless of objective information
to the contrary.
These opposite errors—underestimating most low-probability events and overestimating
some low-probability events—apparently have a common cause: the frequency
and vividness with which people are reminded of these risks. Infrequent and dull
reminders of risk cause people to underestimate them, whereas frequent and vivid
reminders cause people to overestimate them. Most low-probability events are seldom
discussed or portrayed in the media; so, people tend to act as if their probability is close
to zero, whereas potentially catastrophic events such as nuclear risks are much discussed
and portrayed in the media, so people tend to overestimate their probability.3
These findings have implications for the economic model of tort liability. If many
people do not accurately estimate risks, then they cannot make the appropriate calculations
of net benefits and costs that the economic theory assumes that they make.
Using the symbols of the previous chapters, some people may inaccurately set p(x)
equal to 0 for low-probability events and, therefore, take no precaution, when, in fact,
2
See generally DANIEL KAHNEMAN & AMOS TVERSKY, EDS., JUDGMENT UNDER UNCERTAINTY: BIASES AND
HEURISTICS (1981). For behavioral insights applied to law, see Russell Korobkin & Thomas S. Ulen, Law
and Behavioral Science: Removing the Rationality Assumption from Law and Economics, 88 CAL. L. REV.
1051 (2000).
3
A common example of this phenomenon arises from this question: “Which is more common—homicide or
suicide?” Many people answer, “Homicide,” largely because information about homicide is vivid and
widely reported while that on suicide is typically not reported. In fact, suicide is approximately twice as
common as homicide. The U.S. Centers for Disease Control and Prevention reports that in 2009 there were
approximately 18,000 homicides in the United States and about 34,500 suicides.
232 C H A P T E R 7 Topics in the Economics of Tort Liability
that probability is positive, and they should take precaution. The inability of these decision
makers to make accurate calculations may lead to too many or too severe accidents.
In other cases, decision makers may overestimate p(x)—that is, they may think
that an accident is far more likely than it, in fact, is—and may, therefore, take far too
much precaution. As a result of these inabilities to calculate correctly, the tort liability
system—to the extent that it presumes that people can and do calculate, may not induce
people to take actions that minimize the social costs of accidents.
The economic theory of tort liability not only draws our attention to the importance
of the rationality assumption in analyzing tort law, but it also suggests a corrective
measure when that assumption is violated. Consider accidents involving power
tools. One might suppose that precaution in such cases is bilateral: There is something
that both the consumer and the producer can do to reduce the probability and severity
of an accident. As a result, the economic theory would suggest that some form of the
negligence rule should be used to induce efficient precaution by both consumers and
producers. However, suppose that there is strong evidence that consumers do not accurately
assess the risks associated with the use of power tools. They might presume that
the tools are so safe that they need not take any particular care in how they are used. In
short, consumers might mistakenly assume that the probability of an accident is zero
and take very little precaution. That fact would make this a situation of unilateral,
rather than bilateral, precaution: Only manufacturers could realistically be expected to
take steps to reduce the probability and severity of an accident.4
In these circumstances,
manufacturers might be held liable for failing to design a product that would prevent
foreseeable misuse by less than fully rational consumers.
Besides misperceived probabilities and other cognitive errors, many accidents result
from tangled feet, quavering hands, distracted eyes, slips of the tongue, wandering
minds, weak wills, emotional outbursts, misjudged distances, or miscalculated consequences.
Described more abstractly, accidents result from clumsiness, inattention, misjudgment,
misperception, or weakness of will. Occasional acts of this kind are called
“lapses.” Chapter 12 explains how lapses can cause crimes. Here we focus on lapses
that cause unintended negligence, which in turn causes an accident. In these cases an
actor aims for a given level of precaution and fails to achieve it. Negligence rules determine
liability by comparing the legal standard to the injurer’s actual level of care, not
the injurer’s intended level of care. So, actors are liable under a negligence rule for the
harm caused by their lapses.
Here is an example.
Example 1: Unintended Negligence by a Motorist: A motorist sets out
on the long, straight drive from San Francisco to Los Angeles on Interstate 5.
The road is uncongested, it is night, and the speed limit is 70 miles per hour.
Under these conditions, a reasonable driver of a car with a mechanism to maintain
constant speed (“cruise control”) would set it at the speed limit of 70 miles
per hour. The car, however, lacks such a mechanism. Not being a machine, the
4
Note, further, that if the rationality assumption fails, then there is not a great deal to be said for a policy of
better informing the parties about the objective values of the risks. They either discount or ignore that
information.
I. Extending the Economic Model 233
driver cannot possibly go 70 all the time. The driver aims for 65. Occasional
lapses in attention cause the driver to exceed or fall short of the intended speed
of 65. Near the end of the trip, the driver has an accident while going 73 that
he would have avoided if he had been going 70. Under a negligence rule, the
motorist is liable for harm caused by the accident.
In this example, a safe driver is held liable for an accident that he caused by accidentally
going too fast. Now consider the symmetrically opposite example.
Example 2: Intended Negligence by a Motorist: The facts in the preceding
example remain the same except that the driver aims for 75, so he intends to
drive at an unreasonable speed. Occasional lapses in attention cause the driver to
fall short of the intended speed of 75. Near the end of the trip, the driver has an
accident while going 67. Under a negligence rule, the motorist is not liable for
harm caused by the accident.
In this example, a dangerous driver is held not liable for an accident that occurred while
he was accidentally going at a safe speed.
The first driver was intentionally nonnegligent most of the time and accidentally
negligent part of the time. He had “bad moral luck”: He accidentally went too fast at
just the wrong time and caused an accident. The second driver, in contrast, was intentionally
negligent most of the time and accidentally nonnegligent part of the time. He
had “good moral luck” with respect to liability: He accidentally went too slow at just
the right time and escaped liability.5
Allowing moral luck to determine liability may seem unfair to you. Fairness aside,
reducing the role of moral luck in liability improves incentives and reduces inefficiencies.
To see why, we construct a graph to represent the safe driver who had bad moral
luck. The vertical axis in Figure 7.1 represents the probability of driving at a particular
5
Because he could not have avoided this accident by driving 67, speeding was not the cause of this accident,
so he should not be held liable even if he had been speeding. In practice, however, the court may be unsure
of these facts, and it is likely to find him liable if he were going 75 when the accident occurred.
Probability
Precaution
margin
of error
probability
of lapse
causing
negligence
legal
~
intended
x x*
FIGURE 7.1
Probability of a lapse causing negligent
precaution.
234 C H A P T E R 7 Topics in the Economics of Tort Liability
speed. The horizontal axis represents precaution, which corresponds to driving slowly
in the preceding example. We model a lapse as a probabilistic connection between intended
and actual precaution. The actor in Figure 7.1 intends to achieve precaution
level x*. The actual level of precaution x the actor achieves depends on his intention x*
plus a random variable so To provide for a margin of error, the actor
intends to exceed the legal standard, so An actor lapses when actual precaution
falls below intended precaution, or In Figure 7.1, the probability of a
lapse resulting in negligence is the shaded area under the probability density function
that lies below the legal standard.
In Figure 7.1, the actor gains a private advantage by exceeding the legal standard,
but social efficiency requires the actor not to exceed the legal standard. We already explained
this fact in the preceding chapter, which we restate briefly. Recall that precaution
is socially optimal when the cost of taking a little more equals the social benefit of
fewer accidents. If the legal standard is set at the social optimum, then exceeding the
legal standard of precaution has more social costs than benefits. For the actor, however,
private benefits increase significantly when his precaution reaches the legal standard
because he escapes liability. So the actor takes excessive precaution. (Moral luck also
has another bad incentive effect that we cannot discuss here.6
)
We have explained that allowing moral luck to affect liability seems unfair and distorts
incentives for precaution. Replacing a rule of liability for accidents caused by negligence
with a rule of liability for accidents caused by intentional negligence would reduce
the role of moral luck in determining liability. Unfortunately, this remedy is usually worse
than the problem. Compared to a rule of liability for objective negligence, a rule of liability
for intentional negligence requires the plaintiff to prove much more before recovering
damages. If the plaintiff had to prove intentional negligence in order to recover damages,
the burden of proof would be crushing, and recoveries would seldom occur. Thus, the victims
of automobile accidents could seldom recover if they had to prove that the speeding
driver intended to speed. Practical problems of information cause courts to condition liability
on objective negligence rather than intentional negligence. (Consider, however, that
global positioning systems or some other technological advance may someday provide a
complete record of a driver’s speed on any journey, which would often allow a driver to
prove that his speeding was merely an unintended lapse.)
These thoughts raise concerns about whether tort liability induces the appropriate
precautionary action by potential injurers and victims. We shall tentatively maintain the
rationality assumption but shall be ready to amend our conclusions about efficient tort
rules when there is sound evidence that the appropriate decision makers are not behaving
rationally.
QUESTION 7.1: Wearing seat belts and shoulder harnesses is an efficient
means of minimizing the costs of automobile accidents. Assuming that the
x 6 x*.
x* 7 x
'
.
x = x* + e.e,
6
While people cannot choose whether to lapse, they can control the frequency and magnitude of lapses
through concentration, preparation, conditioning, and training. Moral luck causes excessive investment in
these activities by a rational actor. We relegate this fact to a footnote, however, because irrational people
who invest too little in cultivating self-control pose a greater danger to society.
I. Extending the Economic Model 235
benefits of these passive restraints exceed their costs, but that not all drivers and
passengers use seat belts, how might the rules of tort liability be changed so as
to induce a greater number of people to wear seat belts and shoulder harnesses?
Web Note 7.1
We have previously mentioned the burgeoning literature in behavioral law and
economics. Much of that literature relates to the examination of the economics
of tort liability. See our website for much more on the connections between
behavioral law and economics and tort law.
2. Regulations Fire regulations usually require a store to have a fire extinguisher.
Inspectors will check from time to time to confirm that the store complies with the regulation.
If it fails to comply, the regulators may impose a fine. Even if the store complies
and a fire injures a customer, the store may be liable to that customer in a private
cause of action. In this example, the store is subject to safety regulations and liability.
In Chapter 6, the basic model assumed that injurers face liability but not regulations.
The fact that injurers often face both liability and regulations poses the question, “Why
have both safety regulation and liability?” If, in our example, liability law is adequate
at inducing safety precaution, the store will presumably keep a fire extinguisher even
without the regulation. And if the store complies with the regulation, perhaps the injured
customer should seek compensation from his insurance company, not the store.
Comparing liability and regulation, sometimes one is more efficient than the other,
and sometimes the two together are more efficient than either one by itself. A general
theory of safety regulation and tort liability must distinguish these alternatives. Instead
of attempting a comprehensive theory, we will sketch some determinants.7
Administrators have the power to order potential injurers to correct a hazard before
an accident occurs, whereas courts have the power to order injurers to compensate
victims after an accident occurs. Regulation is ex ante enforcement by administrators,
and liability is ex post enforcement by victims. This difference determines many of the
advantages and disadvantages of each.
Administrators and courts differ with respect to information. Administrators can
often acquire technical knowledge needed to evaluate the safety of specialized industries,
whereas courts of general jurisdiction have difficulty accumulating technical
knowledge about specialized industries. In these circumstances, administrators may set
standards better than courts, so the court may use a safety regulation as the standard of
care for determining liability. By accepting safety regulations as defining the legal standard
of care for tort liability, courts defer to administrators. If safety regulation and liability
law impose the same standard of care, then potential injurers will conform to that
standard in order to avoid both ex ante fines and ex post liability.
7
See Steven Shavell, Liability for Harm versus Regulation of Safety, 13 J. LEGAL STUD. 357–374 (1984),
and Charles Kolstad, Thomas Ulen, & Gary Johnson, Ex Post Liability for Harm vs. Ex Ante Safety
Regulation: Substitutes or Complements?, 80 AM. ECON. REV. 888 (1990).
236 C H A P T E R 7 Topics in the Economics of Tort Liability
Sometimes, however, courts have better safety information than administrators do.
For example, a trial may provide judges and juries with better information about the
harm caused by an accident than administrators can predict. In addition, courts often
have fewer political motives than administrators. Problems of information or motivation
can cause a court to distrust the legal standard imposed by a safety regulation.
Courts may feel that the regulators set the standard too low in order to avoid liability
for politically powerful businesses. In these circumstances, the standard of care imposed
by the court for liability may exceed the safety regulation. If liability law
imposes a higher standard than safety regulations, then most potential injurers will conform
to the higher standard in order to avoid liability.
Alternatively, courts may feel that the regulators set the standard too high in order
to reduce competition. For example, U.S. automobile manufacturers may seek high
safety standards to increase compliance costs for foreign competitors. If safety regulations
impose a higher standard than liability law, then most potential injurers will conform
to the regulation in order to avoid fines.
Safety regulations provide a rich source of bribes for corrupt officials in many countries.
Sometimes officials want tough regulations to guarantee that bribing an official is
cheaper than conforming to the regulations. In countries where the administrators are
more political and corrupt than judges, liability has a distinct advantage over regulation.
When tort liability exceeds an injurer’s wealth, the injurer is bankrupt. Some risky
activities attract undercapitalized firms that can escape liability through bankruptcy.
Highly capitalized firms may avoid these same activities to avoid the risk of liability. In
those industries where undercapitalized firms risk bankruptcy, safety regulations have
an advantage over liability. By collecting fines before an accident occurs, officials can
force an undercapitalized firm to comply with safety standards that it would violate if
the only sanction were liability.
Finally, consider the administrative costs of regulations and liability. Sometimes
accidents impose small harm on a large group of people. When the cost of trial for each
victim exceeds his damages, making liability law work requires aggregating claims, as
in a class action suit. Sometimes claims are easier to aggregate in an administrative proceeding
than in a court trial. In general, safety regulations dominate liability as a remedy
for accidents that impose small harm on a large group of people.
Web Note 7.2
The remarkable story of the attempts to regulate the harms from tobacco use
and to hold the tobacco companies liable for those harms constitutes an instructive
case study of the relationships between liability and regulation. See
our website for a discussion of the tobacco cases and the settlement reached in
the U.S. litigation.
3. Insurance How does the availability of insurance affect our analysis of tort liability?
So far, our analysis of alternative tort rules and institutions has proceeded as if no
one were insured. In reality, insurance is pervasive for accidents and tort liability.
I. Extending the Economic Model 237
Now we need to discuss how insurance interacts with tort liability and consider whether
insurance advances or retards the goals of tort law, which we formulated as minimizing
the sum of the costs of precaution, accidental harm, and administrative costs.
A person who faces the risk of accidental harm may buy insurance. When an accident
occurs, a victim with insurance files an insurance claim for compensation with his
insurance company. In addition to recovering from his insurance company, the victim
may have a right in tort law to recover from the injurer. In principle, the accident victim
could recover twice—once from the insurance company and once from the injurer. The
insurance contract, however, usually transfers the victim’s recovery rights to the insurance
company by means of what is called “subrogation”: The insurance company stands
in place of the insured in the tort suit for the harm covered in the insurance claim. To illustrate,
the accident may cause the victim to lose $100 in medical costs and $200 in
pain. The victim may have health insurance to cover the medical costs and no insurance
to cover pain. So, the victim will recover $100 from the insurance company and $200
from the injurer, and the insurance company will recover $100 from the injurer.
Besides insurance against accidental harm, many people buy insurance against liability.
In the preceding example, the injurer is liable for $300. If the injurer has full
liability insurance, the injurer’s insurance company will pay the injurer’s liability of
$300 ($200 to the victim and $100 to the victim’s insurance company).
In the preceding example, the victim’s insurance is incomplete (it covers medical
costs but not pain), and the injurer’s insurance is complete (it covers medical costs and
pain). As insurance becomes more complete, we approach a situation where victims recover
all of their compensation from insurance companies, injurers recover all of their
liability from insurance companies, and the insurance companies resolve disputes
among themselves. In effect, insurance is a private system of liability law that reallocates
the costs of accidents according to contracts between insurer and insured. As this
private system becomes more complete, injurers and victims deal directly with their insurance
companies, not with each other. In these circumstances, people care more
about insurance rates and terms of coverage, and they care less about the underlying
law of accidents (except insofar as the latter affects the former).
Insurance companies set premiums, which provide revenues. They also process
claims and pay them, which are costs of doing business. In perfectly competitive markets,
companies earn zero profits. Applied to insurance markets, this proposition implies
that the premiums equal the claims plus administration costs. Earlier we
formulated the goal of tort law as minimizing the sum of the cost of the harm from accidents,
the costs of avoiding accidents, and the costs of administration. In a system of
universal insurance and competitive insurance markets, the goal of tort law can be
described as minimizing the total cost of insurance to policyholders.
To illustrate this proposition, we contrast no liability and the rule of strict liability.
With a rule of no liability, potential victims buy accident insurance, and potential injurers
have little need for liability insurance. In contrast, with a rule of strict liability potential
injurers need liability insurance, and potential victims have little need for
insurance against those accidents for which the injurers are liable. So, a rule of no liability
causes victims to buy relatively more insurance, and a rule of strict liability
causes injurers to buy relatively more insurance.
238 C H A P T E R 7 Topics in the Economics of Tort Liability
Which rule is more efficient? This policy debate is important historically. In the
nineteenth century, consumers injured by defective products seldom recovered in court,
so consumers who wanted insurance had to buy it themselves.8
In the twentieth century,
the emergence of strict products liability in tort law effectively caused manufacturers
to insure consumers, and manufacturers often bought liability insurance for
themselves. These facts provoked an argument about whether victims’ insurance is
more or less efficient than injurers’ insurance. We cannot answer this question fully,
but we can give the flavor of debate.
In general, insurance transfers risk from the insured party to the insurer. Transfer
is another name for externalize. Externalizing risk gives the insured an incentive to reduce
precaution. The insurance industry, which is old and has its own language, calls
the reduction in precaution caused by insurance moral hazard. To illustrate moral hazard,
a person who insures his car against theft may not be so careful about locking it.
Insurance companies employ various means to reduce moral hazard, notably coinsurance,
deductibles, and experience rating. Under a deductible, the insured pays
a fixed dollar amount of his accidental losses. Under coinsurance, the insured pays a
fixed percentage of his accidental losses. Under experience rating, the insurance company
sets the insured’s rates according to the experience of the insured’s claims. A
claim in year two, for example, usually means a rate increase in year three. While these
devices reduce moral hazard, they cannot eliminate it. Consequently, insurance inevitably
undermines the insured’s incentives for precaution.9
To combat this problem, liability insurers impose safety standards that policyholders
must meet to remain covered by insurance. To illustrate, a fire insurance company
may require a business to maintain fire extinguishers as a condition for writing an insurance
policy. In the preceding section, we contrasted ex ante regulation and ex post
liability. Insurance companies impose standards ex ante. Officials of the company may
inspect for compliance. Insurance safety standards are private regulations imposed
by contract and enforced by private parties, as opposed to public regulations imposed by
law and enforced by state officials. On balance, legal scholars generally think that insurance
promotes the goals of tort law and should be encouraged.10
Having discussed the incentive effects of insurance, we return to the question of
whether no liability or strict liability is a better rule for consumer product injuries. The
rule of strict liability has a distinct advantage over no liability in terms of the efficiency of
insurance markets. As explained, insurance companies usually set rates according to the
history of an individual’s claims through the process of “experience rating.” Many claims
trigger a surcharge, and few claims trigger a discount. Under a rule of strict liability,
8
Note, however, that many forms of insurance that we take for granted were unavailable in the nineteenth
century. Insurance markets took time to develop.
9
An additional problem for insurers is adverse selection, which we discuss in the section on insurance in
Chapter 2 and also discuss below as a possible source of periodic insurance crises. Will adverse selection,
if uncorrected, create efficiency issues in tort liability? Describe and contrast those problems in regimes of
no liability and strict liability.
10
Some states, however, prohibit liability insurance for punitive damages, presumably for the same reason
that states prohibit insurance against criminal fines.
I. Extending the Economic Model 239
a company that produces many defective products makes many claims to its insurer and
thus pays higher rates. This fact creates incentives for more precaution by the manufacturer
to reduce its claims by reducing consumer accidents. However, these incentives for
producers disappear under a rule of no liability, where consumers must buy their own insurance.
One of the main arguments in favor of strict liability for consumer product
injuries is that this rule causes liability insurers to monitor the safety of manufacturers.
Some scholars argue against the rule of strict liability on the ground that it provides
consumers with unwanted insurance. By “unwanted” we mean that consumers would not
voluntarily buy the insurance if they had to pay for it. To illustrate, a well-insured U.S.
motorist who injures his knee when his car slips on ice and crashes will receive compensation
from his insurance company for medical costs, lost wages, and damage to the car.
If, however, the motorist suffers the identical injury due to the fault of another driver, the
motorist will also receive additional compensation in tort for pain and suffering. In general,
the tort liability system effectively provides consumers with insurance against pain
and suffering that they would not buy for themselves. In personal injury cases, pain-andsuffering
compensation can be relatively large, which implies that consumers have a lot
of unwanted insurance. (Later we discuss how a market for tort claims can solve this
problem by allowing potential accident victims to sell unwanted liability rights.)
We have introduced the complex interaction of tort liability rules and insurance
markets. Before leaving the topic, we want to explain a problem with insurance markets.
In 1985 and 1986 and again in the mid-1990s, a “crisis” over liability insurance—
particularly with respect to medical malpractice insurance—occurred in the United
States and elsewhere. During the crisis, insurance companies abandoned some lines of
insurance, refused to renew policies for some persons and companies, lowered the limits
on some insurance coverage, and sharply increased some premiums. The crisis
poses the question, “Is the insurance industry inherently unstable?” Answering this
question explains some fundamental characteristics of the insurance industry.
The answer may be “Yes” for two reasons. The first reason concerns the “reserves”
held by an insurance company. Sound business policy and state regulations require insurance
companies to hold a fraction of their revenue from premiums in reserve to cover
future claims. For some risks, however, many claims can occur at once. To illustrate,
earthquake insurance results in no claims in most years and vast claims when a large
earthquake actually occurs. As a consequence, the insurance company must use the premiums
in years with no claims to build up its reserves (“reserve funding”). Sometimes
an insurance company has larger reserves than it needs to cover a risk. If an insurance
company has excess reserves, it can expand the supply of insurance at little cost. At
other times, insurance companies have no excess reserves, so they cannot write more insurance
policies without increasing their reserves or liquidating the investments they
have made with their other premium revenues. Changing the level of reserves can be
very costly because of tax consequences.11
One theory holds that insurance crises occur
11
Capital removed from reserves becomes taxable as profit; so, insurance companies do not like to remove
capital from reserves except in years when they have losses to offset their tax liability. Similarly, additions
to reserves can reduce tax liabilities (especially under the U.S. tax law before the 1986 reforms), so insurance
companies prefer to add to reserves in years when they have large profits from their other activities.
240 C H A P T E R 7 Topics in the Economics of Tort Liability
because insurance companies exhaust their reserves and must raise premiums relatively
quickly to replace those reserves.
The second explanation is quite different in character. Suppose that insurance premiums
rise, and some people stop purchasing insurance. The people who retain their
insurance represent the worst risk, so the insurance company may have to increase its
premiums again. To illustrate, suppose there is a 20 percent chance that I will suffer an
accident costing $10,000 and only a 15 percent chance that you will suffer such an accident.
If we are both insured, I expect to recover $2000 from the insurance company
in claims, whereas you expect to recover $1500 in claims. Suppose the insurance company
offers us the same insurance policy for the same premium, say $1500. It does not
charge me more than you because it does not have enough information to tell us apart
(or perhaps the regulators won’t permit “price discrimination”). Both of us purchase
the insurance because the premium is equal to or (in my case) less than the expected
value of the claim, and we are both probably risk-averse. The insurance company collects
$3000 in premiums from the two of us and expects to pay $3000 in claims, so it
earns zero profits, as required in perfect competition.
Observe what would happen if the insurance company increased the rates to
$1600. You are more likely than I am to decide to drop your policy. (Why?) But if you
drop your policy, the insurance company loses a good risk (you) and retains a bad risk
(me). Now the insurance company collects $1600 in premiums and it expects to pay out
$2000 in claims, so it expects to lose $400. To overcome this loss, it must raise its rates
again. (To what level does the company now need to raise its premium?) The rates have
to be raised a second time because the first rate increase caused good risks to stop buying
the policy while bad risks continued buying it.
This phenomenon, whereby an increase in insurance premiums drives out good
risks while retaining bad risks, is called “adverse selection.” The second explanation of
the insurance crisis holds that increased claims set off a cascade of increased premiums
due to adverse selection. Exhaustible reserves and adverse selection can create instability
in the supply of insurance.12
QUESTION 7.2: Are subrogation clauses efficient? Be sure to review your
answer in light of the section below about litigation costs and their effect on
the efficiency of the tort liability system.
4. Bankruptcy Under assumptions explained in Chapter 6, strict liability causes
the firm to internalize the social costs of accidents, so it chooses the socially optimal
activity and care levels. The possibility of escaping liability through bankruptcy
changes this conclusion. When potential damages to tort victims exceed the firm’s net
worth, the firm externalizes part of the risk, thus eroding its incentives to take precaution
and restrain its activity level. Limited liability can cause too little precaution and
too much dangerous activity.
12
See Ralph Winter, The Liability Crisis and the Dynamics of Competitive Insurance Markets, 45 YALE J.
REG. 455 (1988); and Michael J. Trebilcock, The Role of Insurance Considerations in the Choice of
Efficient Civil Liability Rules, 4 J. LAW, ECON. & ORG. 243 (1988).
I. Extending the Economic Model 241
Consider the example of a disposal company for hazardous waste. If such a company
planned to remain in business indefinitely, it might use extreme care in dumping
hazardous waste in order to avoid future liability. Alternatively, it might follow the
strategy of dumping recklessly and accumulating potential tort liabilities that exceed
its assets. Anticipating future liability and bankruptcy, the firm continually distributes
profits and remains undercapitalized. When harm materializes and suits begin, the
firm declares bankruptcy, and its tort victims take their place with other unsatisfied
creditors.
This scenario suggests that firms in risky industries may have too many accidents
and too little capital, thus lowering production and distorting the capital labor ratio. In
addition, if tort liability causes bankruptcy and liquidation, the firm’s nontransferable
assets are destroyed, such as its reputation (“goodwill”), organization, and its employees’
knowledge of how the company conducts business (“firm-specific human capital”).
Thus, avoiding liability through insolvency causes significant inefficiencies.
A recent article, however, proposes that new judgment-proofing techniques enable
corporations to avoid tort liability without being undercapitalized.13
First, a corporation
can place risky activities in a subsidiary, which is a separate company owned by
the parent corporation. Courts seldom reach past a bankrupt subsidiary to the parent’s
assets. Some evidence exists that liability causes U.S. firms to divest and locate hazardous
activities in smaller firms.14
Some scholars urge U.S. courts to “pierce the corporate
veil” and extend tort liability to the parent of a subsidiary, or even to people who
own shares in bankrupt corporations. Research by Richard Brooks on the Exxon Valdez,
which was an Exxon oil tanker whose wreck contaminated the Alaskan coast in spring
1989, demonstrates that the oil companies apparently believe that courts will pierce the
corporate veil.15
Specifically, large oil companies responded to massive liability by
shipping more oil in their own tankers, which they control, rather than attempting to
escape liability by contracting with tanker companies to ship their oil.
Second, in bankruptcy the secured creditors get priority over other, unsecured
creditors, including tort victims. By lending a greater proportion of the corporation’s
debt to secure creditors, a firm can shield a larger portion of its assets from the claims
of tort victims.
Third, firms often have expected income, such as future payments from buyers of
the firm’s products (“accounts receivable”). In a process called “securitization,” firms
convert expected income into securities and sell them to investors.16
After securitizing,
the future income belongs to the owners of the securities, so tort victims cannot tap this
income as a source of compensation.
13
Lynn M. LoPucki, The Death of Liability, 106 YALE L. J. 1 (1996).
14
Al H. Ringleb & Steven N. Wiggins, Liability and Large-Scale, Long-Term Hazards, 98 J. POL. ECON. 574
(1990).
15
Richard R. W. Brooks, Liability and Organizational Choice, 45 J. LAW & ECON. 91 (2002).
16
The bonds allow the issuer to convert the stream of future income into a lump sum. One of the first to use
the method of securitization was the rock singer David Bowie, who issued bonds that gave the bondholders
a claim on Bowie’s future income. In honor of this use, securitization bonds are sometimes called
“David Bowie bonds.”
242 C H A P T E R 7 Topics in the Economics of Tort Liability
The distortion of incentives for precaution caused by insolvency has no perfect solution.
Imperfect solutions include compulsory insurance, posting bond, or replacing
ex post liability with ex ante regulations. In addition, replacing rules of strict liability
with negligence rules ameliorates the problem. A negligence rule allows a firm to escape
liability by conforming to the legal standard of care. Having escaped liability, the
firm has no need to shield assets from tort suits. In contrast, a rule of strict liability only
allows a firm to escape liability by insolvency.17
5. Litigation Costs The final core assumption of the economic theory of tort liability
was that litigation is costless. Of course, nothing could be further from the truth:
Litigation is expensive and sometimes ruinously so. A more complete analysis of the
efficiency of the various liability rules we have discussed should introduce these costs
explicitly.18
Costly litigation will have different effects on potential victims and potential injurers.
Moreover, these different effects will have very different implications for the efficiency
analysis of Chapter 6.
Consider, first, the impact of costly litigation on potential victims. If victims must
incur a cost to assert their claims for compensation, then they may assert fewer claims.
Consider an extreme case in which litigation costs exceed the expected compensatory
damages. Victims will not bring suit, and so the potential injurers will not receive the
signal from the tort liability system that what they are doing is unacceptable. They may,
as a result, take less precaution than they should, with the consequence that there may
be more accidents (and more severe accidents) than there should be.
However, costly litigation may have a contrary effect on the decisions of potential
injurers. If it is expensive for an injurer to litigate, then it may make sense to take more
precaution than would be the case if litigation were costless. By taking more precaution,
the potential injurer makes an accident less likely or less severe; if the cost of this
additional precaution is less than the cost of litigation, then we should expect potential
injurers to take additional precaution when litigation is costly. Similarly, high litigation
costs may cause investors to withdraw funds and reduce activities that risk lawsuits. As
a result, there should be fewer and less severe accidents.
Because the effects of costly litigation on potential victims and on potential injurers
pull in different directions (one suggests less precaution; the other suggests more precaution),
we cannot be sure of the net effect of relaxing the assumption of costless litigation.
High litigation costs have an unsettling, counterintuitive implication, which we illustrate
by contrasting two possible rules of legal procedure. Under the first procedure,
assume that when a plaintiff complains that a wrongdoer caused harm of $100, the
judge hears the case and awards damages of $100 whenever the plaintiff proves the
17
For a detailed analysis of how bankruptcy impacts alternative tort rules, see Alexander Stremnitzer &
Avraham Tabbach, Insolvency and Biased Standards—The Case for Proportional Liability, Yale Law
School Faculty Papers, No. 24 (2009).
18
See Janusz Ordover, Costly Litigation in the Model of Single Activity Accidents, 7 J. LEGAL STUD. 243
(1978); and A. Mitchell Polinsky & Daniel Rubinfeld, The Welfare Implications of Costly Litigation in the
Theory of Liability, 17 J. LEGAL STUD. 151 (1988).
I. Extending the Economic Model 243
necessary facts. Under the second procedure, assume that when a plaintiff complains
that a wrongdoer caused harm of $100, the judge flips a coin and dismisses the complaint
without a trial whenever the coin shows “heads.” If the coin lands “tails,” however,
the court decides the case and awards damages equal to 200 percent of the actual
harm. The injurer’s expected liability is the same for both procedures— under
the old procedure, and under the new procedure. If potential injurers decide
how much precaution to take based on expected liability, then the change in procedure
will not affect their behavior, so the sum of the cost of harm from accidents and the cost
of avoiding accidents apparently remains the same. However, the new procedure has
reduced the number of trials by 50 percent, so administrative costs are much lower under
the second procedure.19
Changing our legal system from the first to the second procedure would save costs,
but this will not happen. A judge who actually decided a case by flipping a coin would
provoke outrage and censure. Even so, our example makes this important point: A legal
system can save administrative costs by reducing the probability of liability and offsetting
this fall with an increase in damages. Chapter 13 returns to the equivalent point
in criminal law when we discuss criminal fines rather than civil liability.
QUESTION 7.3: Use the economic theory of bargaining to characterize the
torts in which the transaction costs of settling disputes are likely to be large.
(Hint: Recall the distinction between public bads and private bads.)
QUESTION 7.4: For which liability standard would you expect the litigation
costs to be greater—negligence or strict liability? Why? Is that an additional
efficiency argument for preferring one standard to the other?
6. Conclusion Taken altogether, what is the ultimate result of relaxing the core
assumptions for the conclusions of the previous chapter? Perhaps somewhat surprisingly,
the conclusions of the economic model survive almost intact. We have seen that
relaxing the rationality assumption may be warranted, and that where it is relaxed, the
economic theory helps us to see how tort law ought to take into account the cognitive
imperfections of those whose behavior it seeks to affect. We also saw that relaxing the
assumption that there is no first- or third-party insurance does not change the results of
the economic theory of tort liability. Coinsurance, deductibles, subrogation clauses, and
the implied threat of higher premiums or of policy cancellations preserve the incentives
of potential injurers to take optimal care and of potential victims (through their insurers)
to bring actions in order to induce potential injurers to internalize the social costs
of their carelessness. Nor does the presence of other social policies, such as safety regulation,
necessitate our changing any of the economic conclusions. These alternative
social policies require some account of how best to coordinate tort liability and safety
regulation, and that coordination is likely to require an understanding of the economic
tradeoffs involved. Finally, the fact that litigation is costly does not necessitate a change
in our economic model. Rather, we have seen that costly litigation points in different
.5 * 200
1.0 * 100
19
See David Rosenberg & Steven Shavell, A Simple Proposal to Halve Litigation Costs, 61 VA. L. REV. 1721
(2005).
244 C H A P T E R 7 Topics in the Economics of Tort Liability
directions: On the one hand, it may induce potential victims not to file actions (thus
allowing potential injurers not to bear the full costs of their carelessness and inviting
them to take less care in the future), but, on the other hand, it may induce potential injurers
to take more care (if taking additional care makes accidents less likely or less
severe and is cheaper than the costs of litigating).
B. Extending the Basic Model
The economic model that we have been exploring in this and the previous chapter
explains not just the broad questions of tort liability’s purposes and the differences between
negligence and strict liability, but it also helps us to understand some of the more
special doctrines of tort liability. In this section we shall show how the economic theory
applies to certain special cases—for example, the liability of employers for the torts
of their employees—and to some issues at the frontiers of tort liability.
1. Vicarious Liability There are circumstances in which one person may be held
responsible for the torts committed by another. Where this happens, the third party is
said to be vicariously liable for the tortfeasor’s acts. Vicarious liability may extend
from an agent to his or her principal or from a dependent child to a parent, but by far
the most common instance of vicarious liability is that of employers’ responsibility
for the tortious wrongs of their employees under the doctrine of respondeat superior
(“let the master answer”). The bare bones of this doctrine are that an employer will be
held to answer for the unintentional torts of an employee if the employee was “acting
within the scope of [his or her] employment.” To illustrate, an employer tells an employee
never to drive the company’s truck faster than the speed limit. The employee
speeds, and the truck has an accident. The employer is liable.
Does respondeat superior induce efficient behavior by employers and employees?
The rule creates an incentive for the employer to take care in selecting employees, in
assigning them various tasks, in deciding with which tools to equip them, in training
them, in monitoring them, and more. This is efficient if it is the case—as it generally
would seem to be—that employers are better placed than are employees to make precautionary
decisions.20
In discussing tort liability, we often distinguish between two rules: strict liability
and negligence. Our analysis of the difference applies to vicarious liability. Under a rule
of strict vicarious liability, the employer is liable for harms caused by an employee.
Under a rule of negligent vicarious liability, the employer is liable for harms caused by
negligent supervision of an employee. A switch from negligence to strict liability lightens
the plaintiff’s burden of proof. To illustrate, a careless nurse employed by a hospital
harms a patient. To recover damages from the hospital under a rule of strict vicarious
liability, the patient must prove that someone in the hospital caused the harm, which is
relatively easy. To recover damages from the hospital under a rule of negligent vicarious
liability, the patient must prove that the hospital negligently supervised the nurse, which
20
For a full discussion of the economics of this issue, see Alan Sykes, The Economics of Vicarious Liability,
93 YALE L. J. 1231 (1984).
I. Extending the Economic Model 245
is relatively hard. You already encountered this argument in favor of strict liability when
we discussed consumer product injuries.
We have given an information-cost argument for favoring a rule of strict vicarious
liability for employers rather than a rule of negligent vicarious liability. Another argument
goes in the opposite direction. To illustrate, a sailor on a tanker might negligently
discharge oil onto a public beach at night. Informing the authorities quickly about the
accident will reduce the resulting harm and the cost of the cleanup. The captain of the
ship might be the only person besides the sailor who knows that the harm occurred or
who can prove that pollution came from its ship. Strict vicarious liability gives the captain
an incentive to remain silent in the hope of escaping detection. In contrast, a rule
of negligent vicarious liability gives the captain an incentive to reveal the harm to the
authorities immediately. As long as the captain can prove that he carefully monitored
the sailor, the rule of negligent vicarious liability allows the captain to escape liability.
As compared to a rule of strict vicarious liability, a rule of negligent vicarious liability
encourages employers to report more wrongdoing by employees.21
QUESTION 7.5: What if an accident has occurred because an employee was
performing a job for which he was not qualified after the employee had falsely
told the employer that he was qualified? Should the employer still be liable for
the victim’s losses under respondeat superior?
QUESTION 7.6: The common law did not hold parents liable for their children’s
unintentional torts unless the parents’ negligent supervision led directly
to the tort. But the common law did hold husbands vicariously liable for their
wives’ torts (a rule since abrogated by statute). Can you provide an efficiency
explanation for these common law rules?
QUESTION 7.7: In many states, a bartender (under so-called “dram shop
laws”), friend, party host, or other person who serves liquor to an alreadyintoxicated
person may be held vicariously liable for any damages that person
subsequently inflicts on other people or their property. Does this form of vicarious
liability make economic sense?
2. Joint and Several Liability With and Without Contribution When
several parties cause harm to someone, a question arises concerning whom the victim
can sue and how damages should be allocated among them. To illustrate, suppose that
you suffer a loss of $100 in an accident caused by two people called A and B. They are
jointly liable if you can sue both of them at once, naming A and B as codefendants and
receiving a judgment of $100 against them. They are severally liable if you can sue
either A or B separately, naming each of them as a defendant in a distinct trial and recovering
$100 from each.
21
Jennifer Arlen & Reinier Kraakman, Controlling Corporate Misconduct: An Analysis of Corporate
Liability Regimes, 72 NYU L. REV. 687 (1997).
246 C H A P T E R 7 Topics in the Economics of Tort Liability
Suppose that the plaintiff chooses to recover from only one of several injurers. May
that defendant then force the other injurers to contribute to paying the damages? At common
law for unintentional torts, the defendant did not generally have a right to
contribution, as this is called, from other joint tortfeasors. This harsh rule against contribution
has been abrogated, usually by statute but sometimes by judicial decision, in almost
all the states.22
The law usually subtracts the contribution of one party from the compensation
owed by the other. For example, if A and B jointly cause you harm equal to $100
and you settle with A for $40, then the upper limit on a trial judgment against B is $60.
Defendants are said to be “jointly and severally liable” if each of them is liable for
all the victim’s full losses, not just a portion of them. The plaintiff may proceed jointly
against all his injurers or may elect to recover all damages from only some of them or
only one of them. (Typically, the plaintiff proceeds against the defendant or defendants
who have “deep pockets,” that is, the resources to compensate him.) In the United
States, the actual law in most cases involving multiple injurers is “joint and several liability
with contribution.” Because liability is “joint and several,” the plaintiff can sue
the injurers jointly or separately, as he prefers. Because the law allows “contribution,”
the recovery is limited to 100 percent of the value of the harm. (The common law recognized
two circumstances in which joint and several liability would hold.)23
There are several economic reasons for joint and several liability. One is that it relieves
the victim of the potentially high costs of proving who caused her harm. The
doctrine allows the victim to assert that one of these people, and perhaps many of them,
caused her injuries without incurring the special costs of showing which one or more
of them were responsible and in what proportion. In essence, the doctrine shifts the
costs of establishing exactly what happened to the defendants. Imagine a situation in
which a patient is anesthetized and taken into surgery. During the operation someone
injures the patient. Later she sues all those who were in the operating room, but for obvious
reasons she cannot tell who precisely caused her injury.
Another economic reason for joint and several liability is that it makes the victim’s
recovery more certain by allowing him to proceed against the defendant or defendants
who have the most assets with which to pay damages. Suppose that an uninsured motorist
is going at high speed, strikes a pothole in the road, loses control of his car, and
hits another passenger car, seriously injuring its driver. Assume for the sake of argument
that 90 percent of the fault is attributable to the speeding driver, and 10 percent of the
fault is attributable to the city government for not filling the pothole. The victim will
have difficulty recovering anything from the speeding driver because he lacks insurance
and may have limited resources. However, if the law allows the victim to hold the
22
This is true only for unintentional torts; for intentional torts, such as a violation of the antitrust statutes,
there is still no right of contribution among joint tortfeasors.
23
First, the defendants acted together to cause the victim’s harm, as when two cars driven by A and B are
racing down a street and one of them hits C, a pedestrian. The two drivers acted together by illegally racing
each other. Second, the defendants acted separately, but victim’s harm was indivisible between them.
Thus, two hunters using identical ammunition fire at a pheasant and both of them accidentally hit a third
person. They acted separately to cause the harm, but no one can disentangle the harm caused by one of
them from the harm caused by the other.
I. Extending the Economic Model 247
motorist and the city jointly and severally liable, and if the victim can prove that the city
was negligent in maintaining the road, then the victim can recover 100 percent of his
losses from the city. The city then faces the hopeless task of trying to recover 90 percent
of the damages it paid from the speeding driver who lacks insurance and resources.
Another economic issue concerns contribution and efficiency. Is a rule of contribution
or no contribution more efficient? The no-contribution rule makes all defendants
internalize the cost of accidents, thus creating incentives for optimal precaution by each
of them. In contrast, the rule of contribution causes each defendant to internalize part
of the cost of accidents and to externalize part of the cost. Because costs are partly externalized,
the rule of contribution may not create incentives for optimal precaution by
each defendant. To illustrate, in the example where A and B race their cars and strike
C, optimal incentives require A to bear the full cost of the accident and B to bear the
full cost of the accident.24
Although the rule of no-contribution creates efficient incentives for precaution by
joint injurers, it can also, as we have noted, change reluctant victims into eager victims.
For example, if C receives perfectly compensatory damages from A, then C is indifferent
between no accident and an accident. If C receives perfectly compensatory damages
from A and also from B, then C prefers an accident to no accident. Perhaps the phenomenon
of eager victims explains why law favors contribution rather than no contribution.
Many states have reformed joint and several liability to reduce the ability of the
victim to recover all the costs of the injurer from the injurer with the deep pockets.
After these reforms, the victim must recover some damages from injurers with shallow
pockets. A careful econometric study found that these reforms caused reductions in the
accidental death rate in the United States. The authors’ explanation is that reforms increase
the incentives for precaution by injurers with shallow pockets, and their precaution
reduces accidents the most.25
Web Note 7.3
When there are multiple defendants, it sometimes happens that one or more of the
defendants make an agreement to settle their claims with the plaintiff and then
keep the existence of that agreement secret from the other defendants. Such agreements
are called “Mary Carter agreements” after the case in which they first arose.
See our website for a history and economic analysis of Mary Carter agreements.
Web Note 7.4
We describe some recent empirical literature on “high-low agreements”
between plaintiffs and defendants.
24
See Landes & Posner, Multiple Tortfeasors: An Economic Analysis, 9 J. LEGAL STUD. 517 (1980); and
Polinsky & Shavell, Contribution and Claim Reduction Among Antitrust Defendants: An Economic
Analysis, 33 STAN. L. REV. 447 (1981).
25
W. Bentley MacLeod, Janet Curry, & Daniel Carvell, Accidental Death and the Rule of Joint and Several
Liability, BERKELEY LAW AND ECONOMICS WORKSHOP (October 29, 2009).
248 C H A P T E R 7 Topics in the Economics of Tort Liability
3. Evidentiary Uncertainty and Comparative Negligence26
In the previous
chapter we discussed the several forms of the negligence rule: simple negligence,
negligence with contributory negligence, and comparative negligence. For most of the
last 200 years, negligence with contributory negligence has been not only the dominant
form of the negligence rule but the dominant tort liability rule in the common law countries.
However, within the last 40 years all this has changed. Today, all but a handful of
the states in the United States have altered their law of accidents so that the prevailing
liability standard is one of comparative negligence for non-product-related torts. The
change has been effected principally by statute, with a minority of states adopting the
rule by judicial decision. In most civil law jurisdictions of Europe, the principle of
comparative negligence was adopted long before the United States made this change.
In this section we shall explain briefly how the comparative-negligence rule works and
how it differs from the rule of negligence with contributory negligence. Then we shall
show how something called “evidentiary uncertainty” can give rise to an efficiency argument
for comparative negligence.
The simple reason for the rise of comparative negligence is an increasing dissatisfaction
with the rule of contributory negligence. Recall that a contributorily negligent
plaintiff could not recover anything from the defendant, even from a negligent defendant.
This rule struck most people as exceedingly harsh. To see why, imagine that an
automobile accident has occurred; both the plaintiff and the defendant were driving.
Suppose that violation of the speed limit constitutes negligence and that the evidence
shows that the plaintiff was going 35 miles per hour in the 30 mile-per-hour zone and
that the defendant was going 65 miles per hour in that same zone. Under the rule that
bars recovery for a contributorily negligent plaintiff, the plaintiff will not be able to recover.
This seems harsh in that the plaintiff’s negligence was trivial in comparison to
the defendant’s.
To avoid this sort of harsh result, most jurisdictions found a means of limiting the
scope of the rule of contributory negligence—for example, by means of the last-clearchance
doctrine.27
But eventually these limitations on the application of the rule of contributory
negligence gave way to comparative negligence.
The principal difference between comparative negligence and the rule of negligence
with contributory negligence is that under comparative negligence the plaintiff’s
contributory fault is a partial but not a complete bar to recovery from a negligent defendant.
Thus, under comparative negligence the negligent injurer usually owes something,
but not full compensation, to the negligent victim.28
The equitable argument is the principal justification for the switch to comparative
negligence. However, there are economic efficiency arguments that can be made
on behalf of comparative negligence. To make these arguments requires relaxing at
26
The material in this section draws on Cooter & Ulen, An Economic Case for Comparative Negligence, 61
NYU L. REV. 1067 (1986).
27
Each of these limitations allowed an otherwise contributorily negligent plaintiff to recover all losses. Note
how this differs from the result under comparative negligence described below.
28
There are three different forms of comparative negligence: pure, modified, and slight-gross. These are
extremely interesting but are not central to our economic analysis.
I. Extending the Economic Model 249
least one of the core assumptions that we made in the previous chapter. Recall that
the basic economic theory of tort liability of Chapter 6 showed that all forms of the
negligence rule (simple, contributory, and comparative negligence) were equally efficient.
The only way we can draw efficiency distinctions among them is to relax one
of the core assumptions. Suppose that, in a negligence case, we assume that litigation
is costly in the sense that it is not certain how the court will evaluate the evidence developed
at trial. Thus, neither the plaintiff nor the defendant can be certain whether
the court will determine that their precautionary behavior was sufficient to absolve
one of them of fault. It is possible, for example, that the court will determine that the
precaution of one of the parties was insufficient, even though that party thought that
he or she had complied with the relevant duty to take due care. Or the court may find
one of the parties nonnegligent when in fact the party was violating the legal standard
of care. We may call this condition “evidentiary uncertainty.”
These possibilities of error may influence the precautionary decisions of a potential
injurer. We assume, as seems realistic, that the probability that potential injurers
will be found not liable increases as their precaution increases. Figure 7.2
shows the impact of this fact on expected costs. The effect of evidentiary uncertainty
is to smooth the discontinuity in expected liability at the (presumed) legal standard
that we developed in Chapter 6. Smoothing occurs because injurers’ expected costs
are a weighted average of their costs when liable and their costs when not liable,
with the weights given by the probability that they will be found liable. The effect is
indicated in Figure 7.2 by the sloping curve that connects the expected-cost curve
and precautionary-cost line. Uncertainty about the court’s assessment of a party’s
precautionary level with regard to the legal standard of care induces most injurers to
take more precaution than is prescribed by the legal standard of care. In effect, they
give themselves a margin of error to be sure that they avoid liability. This behavior
is represented in Figure 7.2, which illustrates the fact that an injurer’s costs are minimized
on the smoothed curve at , which is a higher level of precaution than the
legal standard x*.
x+
$
x0
x+x*
wx + p(x)A(x)
wx
p(x)A(x)
FIGURE 7.2
Evidentiary uncertainty smooths the
discontinuity at the legal standard of
care and induces extra precaution by
the potential injurer.
250 C H A P T E R 7 Topics in the Economics of Tort Liability
Evidentiary uncertainty causes potential injurers to go beyond the level of precaution
that might just barely exonerate them.29
That is, evidentiary uncertainty will cause
overprecaution relative to the efficient level of precaution.
This result is true under any form of the negligence rule. What efficiency advantage,
if any, does comparative negligence provide when there is evidentiary uncertainty? The
overprecaution caused by evidentiary uncertainty is less under comparative negligence
than it is under any other form of the negligence rule. The simple reason is that under
comparative negligence, if either party makes a mistake in choosing the level of precaution
that is necessary to satisfy the legal standard of care, the consequence of that mistake
is not visited entirely on the person who made it, as it would be under any other
form of the negligence rule, because, under comparative negligence, the losses are
shared between the two parties rather than being concentrated on one party.
One frequent criticism that is made of comparative negligence is that its administrative
costs are high. The rules to be used in apportioning fault are vague, it is said,
even when the parties are engaged in the same activity: No one is quite sure how to
apportion fault when A was going 45 in a 30 mile-per-hour zone and B was going 60.
But things are even worse when the parties are engaged in different activities. How,
for instance, would you have apportioned fault in Butterfield v. Forrester, the case in
Chapter 3 involving an obstruction in the road left by Forrester and a negligent horseman,
Butterfield, who crashed into the obstruction? Given this difficulty, it is alleged
that litigants and juries will spend inordinately large amounts of effort trying to
establish exact percentages of fault when such exactitude is impossible to achieve.
There may be some truth in the contention that comparative negligence has high
administrative costs. If so, there is a balance to be struck between the efficiency gains
of comparative negligence and these administrative costs. Until we can examine careful
empirical studies, we cannot say whether there is a net efficiency gain from moving
to this new liability standard.
Web Note 7.5
There has been considerable writing about the economics of comparative negligence.
We review that literature on our website.
QUESTION 7.8: Admiralty law—the law that deals with controversies arising
on navigable waters—used a rough-and-ready method of dealing with the
problem of the administrative costs of comparative negligence. Rather than try
to fine-tune the degrees of culpability between the contending parties, admiralty
law simply split the losses 50-50 whenever there was negligence on the
part of both parties. Comment on the efficiency of this method of reducing the
administrative costs of comparative fault. Would you recommend that the admiralty
rule be adopted in apportioning losses in, say, automobile accidents
where both parties are at fault? Why or why not?
29
This is an instance of the point we made earlier in this chapter—namely, that the prospect of costly litigation
will induce potential injurers to take more precaution than they would otherwise. A little extra precaution
makes an accident (and thus, a lawsuit) less likely.
I. Extending the Economic Model 251
4. Products Liability Fifty years ago products liability was a minor part of tort
law, but recently it has become a large and important specialty and the focus of much
of the public dissatisfaction with the entire tort liability system. The liability standard
in product-related accidents is called “strict products liability.”31
For a defendantmanufacturer
to be held liable under this standard, the product must be determined to
be defective. A defect can take three forms:
1. A defect in design, as would be the case if the design of automobile gas
tanks made them liable to rupture and explode (see our website for more
information);
2. A defect in manufacture, as would be the case if a bolt were left out of a
lawn mower during its assembly, causing a piece of the mower to fly off
and injure a user; and
3. A defect in warning, as when the manufacturer fails to warn consumers of
dangers in the use of the product.
What liability standard would economic theory recommend for product-related
accidents? Recall that our discussion of negligence and strict liability focused on whether
precaution for reducing the likelihood and severity of the accident is unilateral or
Incentives for Invisible Actors
How can authorities create incentives for someone who is invisible to them? The State of
Florida has a clever solution. Florida farmers fertilize their fields with phosphorus, which rain
carries into that massive, marvelous, fragile swamp called the Everglades. To control phosphorus,
regulators have adopted a novel incentive system. Beginning in 1995–1996, phosphorus
loadings are compared to a baseline derived from loadings recorded from 1979–1988. If
basin-wide reductions in nutrient load into the Everglades do not meet statutory targets, all
of the farmers in a designated area must pay the “Agricultural Privilege Tax.” The farmers can
escape the tax increase by exceeding an overall 25 percent basin-wide phosphorus reduction
goal. Under this system, each farmer’s abatement efforts reduce his own liability (and the liability
of every other farmer) by the resulting reduction in pollution. Each farmer, consequently,
internalizes the marginal benefits and costs of abatement, as required for efficiency.
This incentive system is remarkable in two ways. First, each farmer has incentives for efficient
abatement without the authorities knowing how much any one of them abates. The authorities
only need to know the total abatement by all farmers. Second, if the authorities have
chosen the phosphorus reduction goal correctly, the farmers will continue abating until they
reach it; so, none of the farmers will actually pay the tax. Theorists describe this approach as
a rule of total liability for excessive harm. “Total liability” refers to the fact that each actor’s
liability depends on the harm that all actors cause. “Excessive harm” refers to the fact that
liability applies to the amount by which the harm caused by all actors exceeds a baseline.30
30
Robert Cooter & Ariel Porat, Total Liability for Excessive Harm, 36 J. LEGAL STUD. 63 (2007).
31
RESTATEMENT (SECOND) OF TORTS (1965), §402A, published by the American Law Institute, lays out this
standard.
252 C H A P T E R 7 Topics in the Economics of Tort Liability
bilateral. If it is bilateral (that is, if both parties can take precautionary action to
reduce the probability and severity of an accident), then a form of the negligence
rule is the appropriate standard. If precaution is unilateral (that is, if only the injurer
can be looked to for actions to reduce the probability and severity of an accident),
then strict liability is the appropriate liability standard. Using this economic
analysis, which standard would modern products-liability law apply?
The more efficient standard would seem to be strict liability because in most instances
of product-related harms, for clarity precaution lies unilaterally with the manufacturers.
It is they who are in control of the design of the products and of the
manufacturing process and who are most likely to be aware of any special dangers that
their products present and, therefore, can most efficiently convey information about
those dangers through warnings.
However, on further reflection, one finds elements of bilateral precaution in the
product-accident situation. Users can also take precautions to reduce the probability
and severity of accidents. For example, they can pay heed to the warnings and use the
products only for their intended uses. There are stories about some consumers picking
up their gasoline- or electric-powered lawn mowers and turning them sideways in order
to trim their hedges, and being injured as a result. No manufacturer intends a lawn
mower to be used in that fashion.
Products-liability law can steer a middle course between the view that precaution is
unilateral (and, therefore, that strict liability is the appropriate standard) and the view that
precaution is bilateral (and, therefore, that negligence is the appropriate standard). It can
do so by holding defendant-manufacturers strictly liable for defective design, manufacture,
or warning but allowing them to escape liability if the victim voluntarily assumed
the risk of injury or misused the product. These defenses encourage the efficient allocation
of risk of loss from product-related injuries between the consumer and manufacturer.
If the lawn mower manufacturer could not exclude liability for consumer misuse
or for voluntarily assumed risk, it would be forced to insure each of its consumers. To
cover the cost of this insurance policy, the manufacturer would have to raise the product
price. The difficulty with this result is that all consumers must pay the higher price,
not just those who are careless. Consumers who are careful would prefer to pay a lower
price for the product and to purchase insurance against loss elsewhere.
We have discussed allowing manufacturers the defense of consumer misuse of defective
products. In practice, a more decisive consideration is often the information that
the court uses to decide whether a product has one of these defects. The court can either
use the information available when the product was manufactured or when the case
was tried. In the mean time, scientific progress uncovers much about risk that was previously
unknown. Asbestos was originally celebrated as a lifesaver that miraculously
protects people from flames, and later it was vilified as a deadly cause of lung cancer.
A company that manufactured asbestos insulation for ships in 1947 did not produce a
product known to be defective at the time, but the product was shown to be defective
by 1977 when a worker brought suit. In the United States, courts have found asbestos
to be defective by using the scientific knowledge available at the time of trial, thus
holding the manufacturer liable for harm that it could not have foreseen by using the
scientific knowledge available at the time of manufacture. Products liability is most
II. Computing Damages 253
“strict” when, besides disallowing manufacturers’ defenses of reasonable care, the law
also holds products to a standard of defect based on science available at trial and unavailable
at manufacture.
QUESTION 7.10: Some scholars discern a trend in modern products-liability
law toward absolute liability or what is sometimes called “enterprise liability.”
Under that theory, manufacturers would be held liable for almost every injury
resulting from the use of their outputs. Give an economic analysis of that liability
standard for product-related harms.
II. Computing Damages
In the previous chapter we noted that the ability of liability rules to induce efficient
precaution depends in part on the ability of the court to award truly compensatory damages
to the victims of a tort. These damages accomplish two things simultaneously:
First, they put the victim back onto the utility level or indifference curve occupied before
the tortious act, and second, they are the “price” that the injurer must pay for having
harmed the victim. In this section we elaborate the ways in which microeconomics
can help to determine the appropriate amount of damages. Additionally, we use microeconomics
to discuss the efficiency aspects of punitive damages in tort awards.
A. Hand Rule Damages
Compensatory damages are intended to “make the victim whole.” In some circumstances,
this is impossible. For example, when a child is killed in a tortious accident,
damages cannot be computed on the formula, “find a sum of money such that the parents
are indifferent between having the money and a dead child, and not having the
money and having their child alive.” The same difficulty arises in a more attenuated
form for irreparable physical injuries, such as those resulting from a crippling accident.
There are, in fact, two distinct concepts of compensatory damages in tort law. One
concept is the standard economic concept of indifference: Compensation is perfect
when the victim is indifferent between having the injury and the damages, and having
neither. Compensatory damages are thus perfect when the potential victim is indifferent
about whether there is no accident or an accident with compensation. This concept
is relevant for injuries in which a substitute for the lost good is available in the market.
When a substitute is available, the market price of the substitute measures the value of
the good to the plaintiff. This concept is also relevant for goods that are bought and sold
from time to time but for which there is no regular, organized market. For example, a
handwritten letter by James Joyce and a 1957 Chevy convertible are sold from time to
time, but these items are so rare that a regular market for them does not exist. The owners
of these rare goods usually have prices at which they are prepared to sell them, and
these prices measure perfectly compensatory damages.32
32
Economists use the term “reservation price” to refer to the minimum price at which the owner of a good is
willing to sell it. Determining the owner’s reservation price for a unique good is a difficult practical problem,
but it is not a problem conceptually.
254 C H A P T E R 7 Topics in the Economics of Tort Liability
This concept of perfect compensation, based on indifference, is fundamental to an
economic account of incentives. If potential injurers are liable for perfectly compensatory
damages, then they will internalize the external harm caused by accidents. And
this creates incentives for the potential injurers to take efficient precaution.
Compensation of this kind is most easily computed for those losses for which there is a
ready market substitute.
But for some tortious injuries there is no ready market substitute. For example,
there is no price at which a good parent would sell a child. The idea that a person could
be “indifferent” between a sum of money and a child is repugnant. And, for some people,
there may be no price at which they would sell an arm or a leg.33
So, for injuries
involving the loss of a child or a limb, compensation simply cannot be perfect. Courts
must, nevertheless, award damages for the wrongful death of a child or for grievous
personal injuries. Our task, then, is to provide a more satisfactory understanding of
their computation.
When U.S. courts award damages for incompensable losses, such as the death of a
child, juries usually set the amount. Unfortunately, judges provide juries with no coherent
instructions for how to compute damages. To illustrate, the recommended jury
instruction for Massachusetts reads:
Recovery for wrongful death represents damages to the survivors for the loss of value
of decedent’s life. . . . There is no special formula under the law to assess the plaintiff’s
damages. . . . It is your obligation to assess what is fair, adequate, and just. You
must use your wisdom and judgment and your sense of basic justice to translate into
dollars and cents the amount which will fully, fairly, and reasonably compensate the
next of kin for the death of the decedent. You must be guided by your common sense
and your conscience on the evidence of the case. . . .
It is common sense that money cannot compensate for a loved one’s death, so how is
common sense supposed to lead the jury to a dollar value? Rather than common sense,
the California jury instructions refer to “reasonableness”:
Also, you should award reasonable compensation for the loss of love, companionship,
comfort, affection, society, solace or moral support.
If no amount of money can compensate for loss of a loved one, then adding “reasonable”
to “compensation” deepens the puzzle rather than clearing it up.
Besides courts, regulators must assign value to loss of life for purposes of costbenefit
analysis. Unlike court practice, the regulators have some clear methods developed
by economists. We will describe such a method and explain its modification for
use by courts.
A necessary part of living is being exposed to the risk of death or serious injury.
For example, flying on an airplane or driving down the expressway involves such a risk.
These risks can often be reduced, but doing so is costly. To illustrate, we may note that
airplanes must be inspected and repaired at regular intervals, which is costly, but the
33
For some people, there may be an amount of money at which selling an arm is an attractive bargain, but
their concept of morality would not permit them to do it.
II. Computing Damages 255
shorter the intervals, the fewer the accidents. Similarly, heavy cars with special safety
features provide extra safety to passengers. But these cars are more expensive to produce
and, therefore, more costly to consumers. When a parent decides what features of
a car to buy or a commercial air carrier decides how frequently to inspect planes for
safety, a decision is being made that balances the cost of additional precaution against
reductions in the probability of injury.
A rational decision about these risks involves balancing the costs and benefits of
precaution. By reasoning in this way, it is possible to compute damages for the loss of
life. To illustrate, we may suppose that the probability of a fatal automobile accident
falls by 1/10,000 when an additional $100 is spent on automotive safety. If expenditures
on automotive safety are rational, then the reduction in the probability of a fatal
accident, multiplied by the value of fatal risk, equals the marginal cost of care:
or
which suggests that the value of fatal risk is $1,000,000.
This method of computing damages for wrongful death, which is called the “value
of a statistical life,” takes actual market purchases as a guide to how much the purchaser
values safety and, by implication, the value of being alive. For example, suppose
that a consumer may purchase a safety device, such as an air bag, by paying extra to
the retailer. If we know how much the safety device costs the consumer and by how
much that device reduces the likelihood of death, then we may infer the consumer’s
valuation of safety, which implies a value of fatal risk. Using the figures from the previous
paragraph, we may assume that the device costs $100 and that it reduces the likelihood
of death by 1/10,000. (Remember that this implies a $1,000,000 value on being
alive.) If consumers purchase the device, then they must value safety at a level that implies
that the value of fatal risk equals at least $1 million.
To apply this method in a legal dispute, the court should consider those situations
in which risk is “reasonable” and well known. In those circumstances, there will be
some value p for the probability of a fatal accident, and some value B for the burden of
precaution. Efficiency requires taking additional precaution until the burden equals the
change in probability p multiplied by the loss L, or (Notice that this is the
Hand rule.) Thus, the court would compute the value of fatal risk by solving the equation
for L, yielding
Notice that his method uses the Hand rule in an unusual way. In the usual way, the
court uses the Hand rule to determine whether the injurer’s precaution satisfied the
legal standard. In its unusual use, the decision maker uses the accepted legal standard
of care that an individual violated to determine his liability.
We have described two distinct methods for computing compensatory damages:
the indifference method and Hand rule damages. The first method is appropriate for
market goods—that is, for losses for which there is a market substitute; the second
method is appropriate when there are legal and moral barriers to such markets. Only
when the indifference method is appropriate can damages be perfectly compensatory.
L = B>p.
B = pL.
(value of fatal risk) = 100>11>10,0002,
(1>10,000)(value of fatal risk) = 100,
256 C H A P T E R 7 Topics in the Economics of Tort Liability
However, both methods, when applied without error, provide incentives for an efficient
level of precaution by potential injurers.
Empirical evidence suggests that Hand rule damages are several times higher than
the U.S. average for damages that courts award in automobile accident cases involving
loss of life.34
For example, the National Highway Traffic Safety Administration
(NHTSA) often values a traffic fatality at $2.5 million. Implementing Hand rule damages
would, consequently, cause a significant increase in damage awards and insurance
costs for some important kinds of accidents. Hand rule damages would also tend to
smooth large differences in damages in individualized cases.35
Besides bringing coherence
to legal doctrine, implementing Hand rule damages would provide incentives for
more rational safety expenditures and create a safer world.
The same accident results in larger damages in the United States than in Germany,
and larger damages in Germany than in Japan. Damages cannot be different in similar
countries for identical accidents and also be optimal in each country. Substantial reform
seems to be required somewhere. Our view is that substantial reform is required
everywhere. Creating a safer society by improving incentives for precaution begins by
using economics to think more clearly about the problem. Hand rule damages suggest
that damages for personal injuries are mostly too low to deter injurers, even in the
United States. This is especially true for automobile accidents and other harms caused
by ordinary people, as opposed to harms caused by corporations or governments, where
damages are higher. A substantial increase in damages for personal injuries involving
automobiles would increase insurance rates, which would reduce the amount of driving
and make drivers more cautious. The lack of systematic calculation for damages in personal
injury cases in the United States also means that damages vary randomly, which
causes liability disparity. Liability disparity is even greater when corporate defendants
are held liable for punitive damages, which is our next subject.
QUESTION 7.9: Victim V works at a job where he might be exposed accidentally
to a chemical that increases the probability from .01 to .02 of dying
from lung cancer in 20 years. V would pay $15,000 to avoid exposure to this
risk, or he would accept $15,000 to expose himself to this risk. No matter how
hard he tries, V cannot imagine any sum of money that he would accept in exchange
for certain death by lung cancer. V’s employer accidentally exposes
him to the chemical. The risk materializes after 20 years, and V dies abruptly
from lung cancer. How much are Hand rule damages for V’s heirs? After exposure
and before dying, V spent $1000 to move to another neighborhood
with better air quality. Should $1000 be added to Hand rule damages, or is it
already implicitly included?
34
See Robert Cooter, “Hand Rule Damages,” conference entitled “Theories of Compensation,” Institute for
Law and Philosophy, University of San Diego Law School, February 28, 2003.
35
It is interesting to note that one of the first recorded legal codes, the Code of Hammurabi, stipulates the
same amount of damages for the wrongful death of a free man or woman, whereas the individualized system
in the United States awards higher damages on average for the wrongful death of a man than a woman.
II. Computing Damages 257
Web Note 7.6
There is extensive literature on how regulators and other legal decision makers
should place a value on a lost life—referred to as the “value of a statistical
life” or VSL. We review that literature, with examples, on our website.
Web Note 7.7
Another difficult category of damage to evaluate for the purposes of compensation
is “pain and suffering.” (We saw earlier in this chapter that there is a
principled argument to be made for not awarding pain-and-suffering damages.
But most jurisdictions do so.) On our website we report on some recent empirical
evidence that seeks to understand how jurors make decisions about
how much money to award for “pain and suffering.”
B. Punitive Damages
In 1984 Getty Oil allegedly agreed to sell itself to Pennzoil, but the Texaco oil company
encroached on the deal and bought Getty. In a lawsuit a Texas jury awarded $7.53
billion in compensatory damages to Pennzoil and $3 billion in punitive damages. (In the
end, the plaintiff settled out of court because the full judgment would have bankrupted
the defendant.) Commentators on the case do not agree as to whether the defendant actually
committed the wrong, which is unusual and has the name “tortious inducement to
breach a contract.” In any case, the award of punitive damages of $3 billion, which broke
previous records, was unforeseeable. Earlier we mentioned the problem of liability disparity
that arises when like cases result in different judgments. Punitive damages are a
“Fortunately for My Client, the Victim Died.”
Would you rather be dead or crippled? In most tortious accidents, victims and their families
prefer the person alive and crippled rather than dead. It is, consequently, worse to cause
someone’s death in a tortious accident than to cause him or her to be crippled.
Yet, the death of the victim can be fortunate for the injurer, because the damages
awarded by courts are often greater when the victim of a tortious accident is crippled than
they are when he or she dies. Someone who is injured severely but has a relatively long life
still ahead will require extraordinary compensation. The income that the victim can no longer
enjoy must be replaced, and the fact that he or she may require constant, expensive medical
attention every day must be taken into account in the assessment of damages.
By contrast, if the victim is killed, the family (or other dependents) will receive only what
they would have received from the victim if he or she had been alive. Thus, if the decedent
would have made $100,000 per year for the next 20 years and would have given his or her
dependents two-thirds of that income each year, then the dependents are entitled to receive
the two-thirds of $100,000 for 20 years, discounted to present value.
258 C H A P T E R 7 Topics in the Economics of Tort Liability
significant source of liability disparity. They cause much uncertainty and fear among
corporate and government defendants. We will analyze punitive damages in the hope of
understanding them better and seeing how to improve them.
Punitive damages are, by definition, damages given to the plaintiff as a way of
punishing the defendant. We must begin our economic analysis of punitive damages by
answering two questions:
1. Under what conditions might punitive damages be awarded?
2. How is the amount of punitive damages computed?
In most states there is a statute describing the conditions under which punitive
damages may be awarded. These are usually attempts to state the common law practices
actually followed by the courts. According to the usual formulation, punitive damages
can be awarded when the defendant’s behavior is malicious, oppressive, gross,
willful and wanton, or fraudulent.36
These statutes merely provide guidelines for
awarding punitive damages. Because the guidelines have not been formulated into exact
rules, there is much uncertainty about when punitive damages can be awarded.
Studies in cognitive psychology demonstrate conclusively that people can order acts
consistently according to how bad they are, but people cannot attach consistent numbers
to the appropriate level of punishment.37
Because moral orderings do not map consistently
into dollar sanctions, the law must devise rules for computing dollar sanctions
to avoid arbitrary disparity in the treatment of people.
There is much uncertainty concerning how to compute punitive damages under
current laws. Statutes typically contain no specific instructions for computing punitive
damages. Punitive damages are supposed to bear a reasonable relationship to compensatory
damages and to the ability of the defendant to pay, but the courts have not specified
what “reasonable” or “ability to pay” mean in this context. It is uncertain, for
example, whether punitive damages may be only double the amount of compensatory
damages or up to 1000 times compensatory damages. In several recent cases, described
in Web Note 7.9, the United States Supreme Court held that punitive damages that are
a double-digit multiple of compensatory damages will attract close scrutiny as possibly
being unconstitutionally excessive. Judges apparently have an idea of how much is
enough, and jury awards have often been reduced by judges, but there are no rules regarding
the computation of punitive damages.38
There is a compelling need in torts for
a more coherent account of punitive damages, and economic analysis can provide
guidelines for the development of this account.
36
The following is the section on “Exemplary Damages” (which is another name for punitive damages) from
the CALIFORNIA CIVIL CODE, §3294:
“For Oppression, Fraud or Malice.
(a) In an action for the breach of an obligation not arising from contract, where the defendant has
been guilty of oppression, fraud, or malice, the plaintiff, in addition to the actual damages,
may recover damages for the sake of the example and by way of punishing the defendant.”
(b) That is not much detail to govern actions on which millions of dollars turn. Notice that
nothing is said about how to compute punitive damages.
37
Daniel Kahneman, Cass Sunstein, et al., Assessing Punitive Damages, 107 YALE L. J. 2071 (1998).
38
See State Farm Mutual Automobile Ins. Co. v. Campbell, 123 S. Ct. 1513 (2003).
II. Computing Damages 259
Organizations as Victims
Economists routinely impute utility functions to individual consumers and workers. But what
if the victim seeking compensation is an organization, such as a partnership, a corporation, a
government, or a club? Like individuals, organizations can be regarded as decision makers,
and their choices can be regarded as revealing organizational preferences. Like those of an individual,
the preferences of a rational organization can usually be represented by a well-ordered
utility function. So the question arises, “Can the utility analysis of the idea of compensation
be extended to organizations?”
In applied welfare economics, benefits or harms to institutions are traced to individuals,
at least in principle. For example, the loss in profits suffered by a business is traced
back to a loss in income to the business’s owners. A common practice is to assume a oneto-one
relationship between the loss in profits to the organization and the loss in income
to individual owners, and to assume that the owners are interested in the business only for
the sake of profits. Under these assumptions, the company’s profits “stand in” for the utility
of affected persons. Because the changes in profits to the business equal changes in income
to its owners, compensating the organization is equivalent to compensating its
owners.
In the case of business firms, the conventional assumption in economics is that they maximize
profits. Thus, when a utility function is imputed to a business, it has a simple form:
Profits are the only thing that the business cares about. For a business, the fall from a higher
indifference curve to a lower indifference curve corresponds to a fall in profits that can be
compensated for, when the fall results from another’s wrongdoing, through an award of
damages to the business equal to the lost profits.
In general, when there is a one-to-one relationship between the loss as measured by
the institution’s preferences and the losses to individuals, the institutional preferences can
be used as a surrogate for the welfare of affected individuals. However, the extension of
the utility analysis of compensation to organizations that are not profit-seeking, such as
governments, clubs, and nonprofit corporations, is problematic because there is less
agreement about the behavioral theories used in describing them. In the absence of an
accepted behavioral theory, there cannot be agreement about how to trace the consequences
of harm suffered by these organizations back to its effects on the welfare of
individuals.
To begin the economic analysis of punitive damages, let us supply some numbers
to the situation described in Example 3 of the preceding chapter.
FACTS: A manufacturer of a fuel additive for automobile engines is
keeping a careful eye on costs. He can set quality control at a high or a low
level. High-level quality control costs $9000 per year and guarantees that the
fuel additive is pure and never causes damage to automobile engines. Lowlevel
quality control is costless (thus saving $9000) but results in some batches
of the fuel additive’s being flawed. A few of the cars using the flawed batch
will be harmed; specifically, the expected damage to cars is $10,000 per year
($1000 in expected damages to each of 10 cars).
260 C H A P T E R 7 Topics in the Economics of Tort Liability
39
Implicit in this argument is the assumption that the rate at which consumers successfully bring suit against
the manufacturer does not change when punitive damages are added to compensatory damages. This is a
strong and unrealistic assumption. When damage awards are high, victims and their attorneys have
stronger incentives to bring action against those who have injured them. This then causes a second-round
effect, in that, as the number of actions increases, the enforcement error falls, and therefore the punitive
multiple should fall. It is an open question whether the existence of a punitive multiple can increase the
number of actions just enough to correct for the inefficiency caused by enforcement error or whether it
leads to overenforcement.
From an economic viewpoint, efficiency requires the manufacturer to make
the quality-control expenditures because the company can expect to save consumers
$10,000 per year by spending $9000 per year on quality control.
Will making the manufacturer strictly liable for compensatory damages produce
this result? The answer is “yes” if the tort liability system is perfect, but “no” if it is imperfect.
Suppose that the tort liability system is perfect in the sense that disputes between
the manufacturer and consumers can be resolved without cost and error, and
damages are perfectly compensatory. With a perfect tort liability system and a rule of
strict liability, every car owner harmed by the product will recover from the manufacturer
without having to spend anything to resolve the dispute. The manufacturer thus
faces $10,000 in expected liability if he does not take precautions costing $9000. A rational
manufacturer maximizes profits net of expected tort liability, so our manufacturer
will set quality control at the high level.
But suppose we make the more realistic assumption that the tort liability system
works imperfectly. Specifically, let us suppose that for every two consumers whose cars
suffer damage, only one actually brings suit and recovers. The other consumer does not
sue because she does not know that the fuel additive caused the harm, or she knows and
cannot prove it. Call the ratio of compensated victims to total victims, which is 1/2 in
this example, the “enforcement error.” Given an enforcement error of 1/2 and assuming
the successful plaintiff only receives compensatory damages, the manufacturer’s
expected liability will be $5000 if he adopts the low level of quality control. He can,
however, save $9000 by reducing his quality control from high to low. So, enforcement
error in this example creates a situation in which a profit-maximizing manufacturer,
whose expected liability is limited to compensatory damages, will choose low-level
quality control, which is inefficient.
The efficiency loss due to enforcement error can be offset by augmenting compensatory
damages with punitive damages. Suppose, as above, that the actual damages
are $1000 per car but that the court doubles this compensatory amount so that
total damages are $2000 per car. If we call the amount in excess of compensation
“punitive damages,” then the punitive damages are $1000 per car. We might also
refer to the multiplicative factor by which we adjusted the compensatory damages
in order to offset the enforcement error as the “punitive multiple.” In our example,
a punitive multiple of two exactly offsets the enforcement error of 1/2 and restores
the manufacturer’s liability to the level that would have prevailed under perfect
enforcement.39
III. An Empirical Assessment of the U.S. Tort Liability System 261
We can state this method of computing punitive damages more abstractly by using
some notation. Without punitive damages, decision maker’s liability L is limited to
compensatory damages, A, which are imposed with enforcement error e in the event of
an accident. Thus,
To offset the error, impose the punitive multiple m, so that liability is given by the equation
By mathematical definition, the “reciprocal” of any value x equals 1/x. Set the punitive
multiple equal to the reciprocal of the enforcement error: m = 1/e. Thus, the punitive
multiple exactly offsets the enforcement error, and the decision maker’s liability
reduces to L = A.
The law might adopt as a rule that, when punitive damages are awarded, the punitive
multiple should equal the inverse of the enforcement error. If such a rule were written
into the law, either by statute or by judges, juries would have some guidance in
setting the punitive multiple. For example, if there were proof that an injurer had failed
to take the appropriate amount of care, because she suspected that only a fraction of
those injured would bring an action against her, the court could impose punitive damages
in an amount determined by application of a punitive multiple equal to the inverse
of the enforcement error.40
Web Note 7.8
In addition to the Kahneman, Sunstein, et al. piece cited in footnote 36 on
page 258, there has been much additional interesting literature on the economics
of punitive damages, including an important article by Polinsky and
Shavell. We review that literature on our website.
Web Note 7.9
The United States Supreme Court has handed down several important recent
decisions on the constitutionality of punitive damages. We describe these
holdings and relate them to the material of this section on our website.
III. An Empirical Assessment of the U.S. Tort
Liability System
How well does the tort liability system achieve its economic goal of minimizing the
social costs of accidents? Many people—including some legislators and other leading decision
makers—believe that the U.S. tort liability system is chaotic, unfair, and inefficient.
L = Aem.
L = Ae.
40
A recent opinion by the United States Supreme Court—Philip Morris v. Williams (2007)—seems to suggest
that the instrumental use of punitive damages that we suggest in this section may be unconstitutional.
See Web Note 7.9.
262 C H A P T E R 7 Topics in the Economics of Tort Liability
Their evidence is largely anecdotal, not systematic, but those anecdotes are striking.
For example, many people are aware that in the mid-1990s a woman who was scalded
when she spilled coffee from a drive-through window at McDonald’s recovered
$640,000 in compensatory damages and $2.9 million in punitive damages for her in-
juries.41
And many municipalities are said to have removed play structures and swing
sets from their public parks and diving boards from the public swimming pools because
of the fear of liability. Congress and many state legislatures consider numerous tort reform
bills each year, all motivated by a desire to reduce high liability insurance costs
(triggered, it is said, by adverse liability judgments).
Anecdotes should not be the basis for assessing something as complex as the
U.S. tort liability system. Such an assessment requires quantitative empirical evidence,
which is quickly improving but still remains at an early stage of development.
We will review the evidence and show that the U.S. tort liability system performs better
than its harshest critics claim. We begin with some descriptive number and then
briefly consider the empirical literature on products liability, medical malpractice,
and mass torts.42
A. Some General Facts About the U.S. Tort
Liability System
Over the past 150 years the most numerous controversies in federal and state
courts arose under contract law. However, some time in the mid-1990s tort cases
became the most common form of adjudicated controversy.43
In the United States, tort law, like contract and property law, is largely state
law. And as we have indicated at several points in the text, there are some significant
differences among the states in these substantive areas of law. In one of the
most recent assessments of the civil justice system, there were, in 1994, slightly
more than 41,000 tort cases resolved in federal district courts (some of which
were in federal court on the ground of diversity but were resolved by the federal
court by the application of state law). During the same time period there were
more than 378,000 tort cases resolved by state courts in the largest 75 counties in
the United States.
Of these state and federal tort cases, 94 percent involve an individual plaintiff.
This is in clear contrast to contract cases, where a significant fraction involve multiple
plaintiffs.
Slightly more than 60 percent of the tort cases in the 75 largest counties in the
United States deal with accidents involving automobiles. The next most common
type of tort dispute (accounting in 1992 for just over 17 percent of all tort cases) is
41
See Libeck v. McDonald’s Restaurants, P.T.S., Inc., No. D-202 CV-93-02419, 1995 WL 360309 (Bernalillo
County, N.M. Dist. Ct. Aug. 18, 1994).
42
We have relied on Daniel Kessler & Daniel Rubinfeld, Empirical Study of the Civil Justice System, in A.
MITCHELL POLINSKY & STEVEN SHAVELL, EDS., HANDBOOK OF LAW AND ECONOMICS, V. 1 (2007). Also see
our website for some additional material on the empirical assessment of the tort liability system.
43
See our discussion of other trends in litigation in the last part of Chapter 11.
III. An Empirical Assessment of the U.S. Tort Liability System 263
that arising from “premises liability” for, say, slips and falls or other injuries at residences,
governmental offices, or commercial establishments. The third most common
form of tort case is medical malpractice, accounting for just under 5 percent of all
torts. And the fourth is products liability, accounting for 3.4 percent of the total.44
Critics of the tort liability system in the United States contend that juries award punitive
damages too often and too liberally and that judges do little to restrain these punitive
awards. However, punitive damages are extremely rare. In all product-liability cases between
1965 and 1990 there were only 353 punitive awards, and those averaged $625,000
(in 1990 dollars). Appellate panels reduced many of these punitive awards so that, after
appeal, the average fell to $135,000. More than 25 percent of those 353 awards involved
asbestos. Over the entire period there was an average of 11 punitive-damages awards per
year in product-liability cases in all state and federal courts. A careful study of punitive
damages in product-liability cases found that at the trial level the ratio of punitive to compensatory
damages was 1.2 to 1; in more than one-third of the cases in which punitives
were awarded, compensatory damages were larger than the punitives.45
More than half
the states prohibit or cap punitive damages or raise the evidence standard that must be
met before they can be awarded.46
Recall that the usual standard in civil actions is “preponderance
of the evidence,” which is generally taken to mean 51 percent believability.
The “clear and convincing” evidence standard is more demanding, but not as demanding
as the criminal law’s standard of “beyond a reasonable doubt.”
The theme of much of the empirical literature is that the tort liability system
(perhaps in conjunction with the administrative agency regulatory system) works reasonably
well at deterring accidents. In most situations in which accidents might happen,
the recent trend in the United States has been toward fewer and less severe
accidents. For instance, the number of motor vehicle deaths and injuries peaked
around 1970 and has declined ever since. The death and injury rate per capita has
shown a dramatic drop.
Web Note 7.10
A central issue for the economic analysis of tort law is the extent to which exposure
to tort liability induces parties to behave in an efficient manner. In the last
few years there has been an outpouring of scholarship designed to explore this
matter with careful empirical studies. We review this literature on our website.
44
We will see in Chapter 11 that there is evidence to suggest that the total number of trials of all kinds has
been declining in the United States. We will also consider some additional facts about litigation, such as
the success rate of plaintiffs in different kinds of actions. To the extent that there is specific information
about success in torts cases, we will describe it in the next two sections of this chapter.
45
Michael Rustad, Demystifying Punitive Damages in Products Liability Cases: A Survey of a Quarter
Century of Verdicts, The Roscoe Pound Foundation (1991).
46
Twelve states require a “clear and convincing” evidence standard for punitive damages but do not limit the
amount. Another twelve states cap the amount of damages and require the “clear and convincing evidence”
standard. Seven states require a portion of the punitive award to be paid to the state. Four states prohibit
punitive-damages awards.
264 C H A P T E R 7 Topics in the Economics of Tort Liability
B. Medical Malpractice
Even though disputes about iatrogenic injuries (those arising in the course of health
care delivery) are a relatively minor portion of all tort cases (accounting for about 5 percent
of the total), there has been a great deal of concern about medical malpractice litigation.
And with some justification. The Institute of Medicine reported in 2000 that
medical errors are the “leading cause of accidental death in the United States.”47
Exact
figures are hard to pin down, but estimates made in 1997 range from 44,000 to almost
100,000 deaths per year. “Medication errors alone account for approximately 7000
deaths per year, exceeding the number of deaths due to workplace injuries.”48
A comprehensive
study of hospital admissions in New York State during 1 year in the 1980s
found that 1 percent of admissions involved serious injury due to negligent care.49
The most careful studies of medical malpractice litigation indicate that the “number
of malpractice claims per physician and the award paid per claim increased rapidly
in the United States from the 1960s to the 1980s. Claim frequency increased at more
than 10 percent per year, reaching a peak of 17 claims per 100 physicians in 1980s.
Awards paid per claim increased at roughly twice the rate of inflation.”50
There is some
evidence to suggest that in at least some jurisdictions the rate of increase of both claims
against physicians and award levels ceased or slowed significantly in the 1990s and remained
at those lower levels through the early 2000s, perhaps as a result of statutory
reform, which we will discuss at the end of this section.
The tort liability system should provide an incentive for physicians and other health
care professionals to take precautions against injuries. Are the incentives currently provided
by the system deficient, efficient, or excessive? With current evidence, we can
only speculate. Physicians have monetary incentives to be careful when they bear the
cost of the accidents that they cause. Liability insurance transfers the cost of accidents
caused by physicians to the insurer. The insurance rates paid by individual physicians
respond only weakly to the history of tort claims against them, so physicians may bear
only a fraction of the costs of patient injury.51
Furthermore, a study of New York
State hospital patients by Weiler found that only about 10 percent of those who were
injured—even seriously—filed a complaint against their health care provider. These
facts suggest that monetary incentives for care by physicians are deficient.
Evidence about “defensive medicine” can be interpreted as implying the opposite conclusion.
This phrase refers to procedures and treatments motivated by reducing liability
47
JANET CORRIGAN, LINDA T. KOHN, & MOLLA S. DAVIDSON, TO ERR IS HUMAN: BUILDING A SAFER HEALTH
SYSTEM (2000).
48
See Daniel Kessler in the Winter 2000 edition of the NBER Reporter, available at http://www.nber.org/
reporter/winter00/kessler.html.
49
See PAUL C. WEILER, HOWARD HIATT, JOSEPH P. NEWHOUSE, WILLIAM G. JOHNSON, TROYEN A. BRENNAN, &
LUCIAN L. LEAPE, A MEASURE OF MALPRACTICE: MEDICAL INJURY, MALPRACTICE LITIGATION, AND PATIENT
COMPENSATION (1993).
50
Patricia Danzon, Liability for Medical Malpractice, in A. J. CULYER & JOSEPH P. NEWHOUSE, EDS.,
HANDBOOK OF HEALTH ECONOMICS, V. 1B (2000).
51
See Frank Sloan, Experience Rating: Does It Make Sense for Medical Malpractice Insurance?, 80 AM.
ECON. REV. 128 (1990).
III. An Empirical Assessment of the U.S. Tort Liability System 265
more than by medical needs. Thus, doctors and hospitals take too much care in the hope
of forestalling injury or demonstrating in later litigation that they did “everything possible”
to prevent harm. Patients have little reason to resist unnecessary procedures that
do no harm so long as insurance companies pay the bill. In 2005, U.S. health care
spending was 16 percent of gross domestic product, or $6697 per person.52
Plausible
estimates suggest that defensive medicine accounts for, at most, 5 percent of total
health care costs. If these figures are right order of magnitude, then defensive medicine
costs each American around $300 per year.
Societal concerns about the cost and availability of health care and the possible
link between medical malpractice and those concerns have motivated many states to reform
their medical malpractice systems.53
Those reforms have taken two particular
forms—limitations on the total amount and kind of damages available in medical malpractice
actions and abrogation of the collateral source rule in medical malpractice.
These reforms may sometimes have perverse results. In the mid-1980s, Indiana
capped medical-malpractice awards at a maximum of $500,000 for all damages and instituted
a professionally administered patient-compensation fund to decide all losses
above $100,000. The unexpected result was that malpractice awards in Indiana became
one-third higher than those in Michigan and Ohio, which had kept the traditional
method of compensation. Perhaps the reason for the Indiana result was that the professional
administrators were better able than lay jurors to calculate damages and, therefore,
came closer to the “true,” higher losses of the victims.
The intended effect of the limitation on damages for medical malpractice is, we
hope, obvious. Some states limited the total amount that could be recovered in any tort
action, while others capped noneconomic damages, such as those on pain and suffering.
Some state supreme courts, such as that in Illinois, have struck down those limitations.
But most caps have survived litigation and attempts at legislative reform.54
52
Aaron Catlin, Cathy Cowan, Stephen Heffler, & Benjamin Washington (the National Health Expenditure
Accounts Team), National Health Spending in 2005: The Slowdown Continues, 26 HEALTH AFFAIRS, 142
(2007).
53
Health care costs in the United States in 2009 account for almost 18 percent of the $13 trillion Gross
Domestic Product, and in late 2009 there were approximately 47 million uninsured people (of a total population
slightly greater than 310 million), two-thirds of whom are low income.
54
At least one state, Florida, sought to reduce malpractice litigation by abrogating the American rule in favor
of the English rule for attorney’s fees in some malpractice actions. (The American rule, which we will discuss
in Chapter 11, calls for each party to pay its own attorney’s fees. The English rule calls for the losing
party to pay not only for its attorney but also for the winning party’s attorney.) The thought was that there
would be less incentive to bring a speculative cause of action under the English rule. Florida abandoned the
experiment after only five years, 1980–1985. See James W. Hughes & Edward A. Snyder, Litigation and
Settlement under the English and American Rules: Theory and Evidence, 38 J. LAW & ECON. 225 (1995)
(which found that plaintiff success rates, average jury awards, and the size of out-of-court settlements all increased
under the English rule, perhaps because the average quality of those claims brought forward
increased under the English rule), and A. Mitchell Polinsky & Daniel Rubinfeld, Does the English Rule
Discourage Low-Probability-of-Prevailing Plaintiffs?, 32 J. LEGAL STUD. 517 (1998) (which argues that,
taking the settlement process into account, the English rule encourages more litigation by low-probabilityof-prevailing
plaintiffs).
266 C H A P T E R 7 Topics in the Economics of Tort Liability
In an earlier discussion of insurance, we explained that an accident victim’s
compensation from an insurance company does not reduce the tort damages owed
by the injurer. Suppose that the plaintiff seeks $100,000 in damages from the defendantinjurer
and has already been paid $80,000 from her insurer for her injuries.
According to the legal principle called the “collateral source rule,” the plaintiff does
not have to reduce the amount she seeks from the defendant—to, say, $20,000—by
deducting the collateral benefits. Some states changed this rule, mandating deduction
of collateral benefits for injuries received in the course of health care delivery.
The thought was that the collateral source rule created an incentive for plaintiffs to
litigate on the theory that they could recover more than their actual losses. So, removing
the rule would, all other things equal, reduce the incentive to bring medical
malpractice complaints.
Both types of reform seek to reduce health care costs by reducing medical malpractice
awards. Were the reforms responsible for the slowdown in the number of
malpractice actions in the 1990s and early 2000s and the cessation in the rate of increase
of the average malpractice award? Or were there other factors—such as an
increase in the technology of treatment and levels of precaution—that explain these
effects? The evidence is still not clear. There is some evidence to suggest that reforms
in the 1980s reduced defensive medicine expenditures by 5 to 9 percent and
that the supply of physicians was about 12 percent greater in those states with caps
on noneconomic damages, by comparison to the supply in states without those
caps.55
But other evidence suggests that the medical malpractice problems observable
in the United States are not specific to the structure of the U.S. civil justice
system but are happening worldwide, perhaps because the great advances in medical
technology have made a much wider range—a perhaps a riskier range—of
medical interventions possible.
Web Note 7.11
Some prominent legal scholars have been using a remarkable data set from the
State of Texas to explore many issues in medical malpractice. We summarize
their work on our website.
Web Note 7.12
Product liability might be triggered by a design defect, a manufacturing defect,
or a failure to warn consumers (and intermediaries) of risk of harm. On
our website we discuss some economics of the duty to warn and some recent
empirical evidence on the effectiveness of that duty.
55
See Kessler & Rubinfeld, supra n. 42, for citations.
III. An Empirical Assessment of the U.S. Tort Liability System 267
C. Reforming Products Liability
Products liability law is the focus of much of the public dissatisfaction with the entire
tort liability system. A survey of chief executive officers by the Conference Board (a
business interest group) found that liability concerns caused 47 percent of those surveyed
to drop one or more product lines, 25 percent to stop some research and development,
and 39 percent to cancel plans for a new product. In some instances, insurers have
decided that the products liability area is so uncertain that they have withdrawn from the
market entirely. Some of the manufacturers and others who have been left without insurance
coverage have decided to stop making their products, or they have raised prices to
cover the cost of additional risks of liability. Since the early 1980s in the United States
there has been a powerful political interest in reforming products-liability law both at
the federal and state level. But until very recently no reform occurred.
Manufacturers have long argued for reform at the federal level for two reasons.
First, they contend that a uniform federal products-liability law would save costs, with
consequent savings to consumers. Second, many manufacturers believe that the
products-liability law that has become the norm in the states is seriously flawed.
Specifically, they believe that plaintiffs win too easily, and that juries are overly generous
to successful plaintiffs (as evidenced, they believe, by the example of the award
against McDonald’s for a hot coffee spill). The argument that the manufacturers make
is that these inefficiencies could be corrected by Congress enacting a sensible uniform
federal products-liability law. Although reform measures have been introduced in
Congress for many years, they have never been passed by both houses.
At the state level, there was a spate of reforms in the mid-1980s and a second round in
the mid-1990s. The reform movement has not revived in the early 2000s. State reform has
typically been limited to putting a cap or upper limit on the amount and kind of damages
that victims can recover. Sometimes the states place this cap only on what is perceived to be
the offending element in damage awards, such as pain-and-suffering or punitive damages.
For example, Illinois’ 1995 Civil Justice Reform Act put a cap on noneconomic damages of
$500,000 and limited punitive damages to three times compensatory damages.56
We should note, in light of the material at the beginning of this chapter, that federal regulatory
agencies do a significant job of promoting product safety. For example, the
Consumer Product Safety Commission, the National Highway Transportation Safety
Administration, the Occupational Health and Safety Administration, the Federal Aviation
Administration, the Food and Drug Administration, and the Environmental Protection
Agency issue and enforce product safety regulations for a wide variety of products.57
An
important issue is whether this joint system of ex ante safety regulation and ex post liability
exposure achieves the socially optimal amount of care. Or does it do too little or too much?
As was the case with medical malpractice, the effects are unclear. The empirical
literature on the effects of products liability, regulations, and the reforms instituted in
56
The Illinois Supreme Court found this act to be unconstitutional in 1997.
57
See CONGRESSIONAL BUDGET OFFICE, THE ECONOMICS OF U.S. TORT LIABILITY: A PRIMER (2003), available
at http://www.cbo.gov/showdoc.cfm?index=4641&sequence=0. Recall, too, that in Web Note 7.2 we
described the story of liability for harms arising from cigarette use.
268 C H A P T E R 7 Topics in the Economics of Tort Liability
the 1980s and 1990s is still young and has not yet reached a consensus. There is some
evidence to suggest that the reforms eased the liability pressure on manufacturers, and
thereby caused liability insurance premiums to stop their long pattern of increase. And
there is some evidence that accidental injuries and deaths have declined in the early
2000s, although how much of that decline is due to law is not clear.
QUESTION 7.11: Analyze caps and limitations on litigation awards using
the analysis of rent control in Chapter 2.
QUESTION 7.12: Use the graphical analysis of liability of the previous chapter
to show the effect on the precautionary decisions of a potential injurer when
the amount of compensatory damages that a victim may receive is capped.
QUESTION 7.13: Suppose that any punitive damages awarded to the plaintiff
were to be paid, not to the plaintiff, but rather to, say, a charity designated
by the plaintiff. How might plaintiffs’ incentives to seek punitive damages be
affected by such a scheme? How might the jury’s disposition to award punitive
damages be affected?
Notwithstanding specific problems, there are other indications that the system is
working reasonably well. Products-liability actions in the United States increased in the
mid and late 1980s, but the vast majority of those cases involved asbestos. If we exclude
asbestos claims, the number of products-liability cases in the federal courts between
1985 and 1991 decreased by 40 percent and has remained at that low level through the
early 2000s. Another interesting recent change regards plaintiff success rates. Between
1981 and 1987 the defendant won 51 percent of the verdicts in products-liability cases.
Between 1988 and 1994 defendants won 64 percent of the cases. The best recent figures
for the early 2000s suggest a retreat to a roughly 50 percent success rate for defendants.
Finally, products-liability insurance costs amount to one-quarter of one cent for each
dollar of product purchase price—an insignificantly small amount.58
Web Note 7.13
In a recent article in the Harvard Law Review, Professors Polinsky and
Shavell discuss the “Uneasy Case for Product Liability.” We discuss that article
and several comments on it in this web note.
D. Mass Torts
A “mass tort” is not a formal legal term but rather a term used to describe a situation
in which a large number of tort claims arise from a single incident or use of the
same product.59
An example might be the Bhopal disaster in 1984 in India in which a
58
See James A. Henderson & Theodore Eisenberg, The Quiet Revolution in Products Liability: An Empirical
Study of Legal Change, 37 UCLA L. REV. 479 (1990). The figures cited in this study are for actions filed
in federal courts. Most tort actions are filed in state courts, but the authors feel that the federal statistics
also reflect trends in state courts.
59
Mass torts are related to but distinct from class actions, which may be an administratively tractable method
of dealing with mass torts. We will discuss the economics of class actions in Chapter 11.
III. An Empirical Assessment of the U.S. Tort Liability System 269
cloud of a highly toxic chemical escaped from a Union Carbide plant and killed between
15,000 and 20,000 people and injured thousands more. Dealing with an incident
of this sort may overwhelm the normal institutions and practices of the tort liability system.
As a result, the law has increasingly tried to deal with the problems of mass torts
through novel arrangements.
Consider the problems arising from asbestos. Asbestos has remarkable fire-retardant
properties, which made it a valuable construction material. It was used extensively in the
United States from the 1930s through 1979, when use virtually ceased. The cessation occurred
because it became increasingly obvious that asbestos could be extremely dangerous
to one’s health, including killing some of those who inhaled asbestos fibers. Inhaling
asbestos causes cancer, but the gap in time between exposure to asbestos and the appearance
of cancer can be 20 years. During the 50 years of asbestos use and the decade or so
during which removal of asbestos was a common practice (before being suspended as
causing more harm than good), more than 25 million U.S. workers were exposed to ambient
asbestos fibers and could, therefore, contract a debilitating disease or die. It is estimated
that more than 225,000 premature deaths occurred between 1985 and 2000
because of exposure to asbestos fibers. And estimates are that an additional 10,000 people
will die each year for the next decade or more because of exposure to asbestos. Much
larger numbers of people have been injured but not killed by exposure to the fiber.
Naturally, when these health risks became widely known, a large number of claimants
stepped forward to seek compensation from the asbestos manufacturers and others. Indeed,
600,000 claimants had come forward by the year 2000 to proceed against 6000 defendants
representing 75 out of 83 possible industries (suggesting that this problem touches almost
every branch of the U.S. economy). Several important firms have filed for bankruptcy because
of earlier or reasonably anticipated adverse judgments involving asbestos.
A large number—perhaps a majority of claimants—may never develop an asbestosrelated
disease. Liability law does not compensate for exposure to risk, as opposed to
the realization of a risk. A person exposed to asbestos must wait until she actually develops
an asbestos-related disease before suing. However, the victims who develop
such a disease must assert their claims without delay. Every jurisdiction in the United
States has a statute of limitations that requires all those who seek compensation to
come forward within a relatively short time to assert their claim against others.60
Failing to do so may cause the victim’s cause of action to lapse. In light of these
statutes, claimants who think that they may develop an asbestos-related disease come
forward as soon as they can. They sue when X-rays are consistent with the early stages
of disease, even though disease may not develop fully. This uncertainty makes the
claimants extremely uneasy but also creates an incentive for the defendants to contest
liability (and to delay settlement as long as possible in the hope that asbestos-related
diseases will never become manifest).
60
Every legal system has a similar set of rules for encouraging those with legal claims to come forward. The
period during which a victim must assert her claim or lose it is called a “prescriptive period” in the civil
law systems. The law also uses the phrases “statute of repose” and laches to describe situations similar to
those covered by a statute of limitation.
270 C H A P T E R 7 Topics in the Economics of Tort Liability
Dealing with this litigation has been extremely expensive. One estimate suggests
that more than $50 billion has been spent on asbestos litigation. More than half of that
amount has, it is alleged, gone to pay for transaction costs, rather than redounding to
the benefit of victims. There are plausible estimates that total litigation costs have
reached more than $250 billion, with almost all asbestos cases resolved, as of late 2009.
To deal with mass torts, the courts and legislatures have been willing to entertain
novel practices. One reason for these novelties is a fear that relying upon standard
tools of tort liability might lead to injustices. The slow development of asbestos-related
diseases creates a conflict between the timeliness of claims required by statutes of limitation
and the need to get compensation to deserving plaintiffs. To address this and
other perceived problems with resolving the large number of asbestos claims through
private litigation, Congress has proposed (in “The Fairness in Asbestos Injury
Resolution Act” (2005)) for several years (but not enacted, as of July, 2010) the creation
of a trust fund to provide limited compensation to victims of asbestos-related
diseases and to limit liability of defendants.
Consider one more example—problems of proving causation arising in a mass tort
having to do with the drug DES (diethylstilbestrol). DES was administered to pregnant
women in the 1950s to prevent miscarriages. However, the drug caused genital diseases, including
cervical cancer, in some of the adult women whose mothers had taken DES 20 or
more years ago. By the time the connection between the adult diseases and the DES was
discovered, it was all but impossible for the plaintiffs to produce evidence about which manufacturer
had produced the DES taken by their mothers 20 years or more before. Standard
theories of causation in tort required the plaintiff to demonstrate by a preponderance of the
evidence that the defendant was responsible for the plaintiff’s harm. In these instances,
plaintiffs had been harmed by one of the manufacturers of DES, but they could not demonstrate
which one or ones. Rather than allow the plaintiffs to leave the court empty-handed,
the California Supreme Court fashioned a novel theory of liability—“market share
liability”—according to which all the manufacturers who might have been selling DES to
the plaintiff’s mother would share liability for the plaintiff’s damages in proportion to their
market shares in the market for DES at the time of the mother’s having taken the drug.61
Web Note 7.14
The tragic events of September 11, 2001 gave rise to mass torts. See our website
for a discussion of the methods by which the federal government put together
an administered compensation package for those who lost relatives and
others in the tragedy.
QUESTION 7.14: Explain how a potential victim’s waiving a future claim
(that is, an employee’s agreeing not to seek compensation from his employer
if he is injured on the job) is like a transaction in a UTC (which we explain in
the box on pp. 272–73).
61
See Sindell v. Abbott Laboratories, 26 CAL. 3D 588, 607 P.2D 924, 163 CAL. RPTR. 132, cert. denied, 101
S. CT. 285 (1980).
III. An Empirical Assessment of the U.S. Tort Liability System 271
Vaccines and Products Liability
Many recent products-liability cases involve the duty of pharmaceutical manufacturers and
doctors to warn those taking drugs of the potential risks involved.
One such case involved two polio vaccines. The first vaccine against this crippling disease
was the Salk vaccine or IPV, which is a so-called killed-virus vaccine. The killed-virus vaccine
prevents polio in the person who receives it without presenting the risk that the recipient will
contract polio. The second vaccine was the Sabin vaccine or OPV, a “live-virus” vaccine. The
recipient retains the live virus in his or her system and can pass it to others, who are themselves
immunized against polio. This external benefit is so considerable that public-health authorities
strongly recommended that young children take the Sabin vaccine instead of the
older Salk vaccine. When only the Salk vaccine was available, there were 2500 cases of polio
a year. After the development of the live-virus vaccine, polio virtually disappeared.
However, the live-virus presents a risk.62
Approximately one of every 4 million people who
take the vaccine or come in close contact with those who have taken OPV contracts polio.
The law should require vaccine manufacturers to warn recipients of the risk from the livevirus
vaccine. That is precisely what the U.S. Court of Appeals for the Fifth Circuit held in
Reyes v. Wyeth Laboratories, 498 F.2d 1264 (1974). After Reyes, it became standard practice
for vaccine manufacturers to include package inserts warning of the risks of the OPV vaccine.
However, that resolution was only temporary. The more general trend toward absolute
or enterprise liability for product-related harms has been felt in this market, too. In many recent
cases, children, whose parents had been warned in accordance with Reyes but who,
nonetheless, took the live-virus vaccine and developed polio, sued the manufacturers and received
large awards. Without the defense of assumption of the risk after an adequate warning,
the manufacturer cannot avoid liability. Therefore, the company must build this higher
expected-liability cost into the costs of production.
Pharmaceutical manufacturers are so fearful of products-liability awards that they have
become reluctant to manufacture and distribute beneficial drugs. In 1976, after an outbreak
of swine flu, a dangerous illness, manufacturers of a swine flu vaccine refused to market it
because private insurers, fearful of the product-liability consequences of 100 million or more
injections, would not issue liability insurance. The companies offered the inoculations only after
the federal government agreed to be the exclusive defendant in any actions for harms arising
from the vaccine.63
The DPT vaccine against whooping cough is in short supply in this
country because the largest manufacturer, Eli Lilly & Company, has stopped producing the
drug due to its fear of adverse products-liability judgments. Currently the following vaccines
that were once manufactured by a number of firms in the United States are now produced by
a single firm: measles, mumps, Sabin polio, Salk polio, and rabies. Worse still, the threat of
product-liability suits may reduce the incentive of pharmaceutical companies to invest in research
and development of potentially beneficial new drugs.
63
The vaccine’s manufacturers proved particularly astute in this matter. The vaccine seems to have caused a
potentially paralyzing or fatal disease called Guillain-Barré syndrome in a small fraction of those who
were inoculated. Numerous plaintiffs brought actions against the federal government, as the sole defendant,
on a theory of inadequate warning. The federal government relatively quickly stopped the program
of inoculation for swine flu.
62
See Edmund Kitch, Vaccines and Product Liability: A Case of Contagious Litigation, REGULATION
(May/June 1985).
272 C H A P T E R 7 Topics in the Economics of Tort Liability
64
The material in this box is based on Robert D. Cooter, Towards a Market in Unmatured Tort Claims,
75 VA. L. REV. 383 (1989). See also PAUL RUBIN, TORT REFORM BY CONTRACT (1993).
Contractual Solutions to the Tort Liability Crisis64
A victim’s right to compensation for accidental harm is a form of insurance. Victims buy these
rights from insurance companies by contract and the tort system gives these rights to potential
victims by law. The tort system, however, gives people far more rights to compensation than
they buy from insurers. Critics of the tort liability system complain that it gives people insurance
that they do not buy when they have to pay for it themselves. Thus, parents seldom insure their
children’s lives, and few people buy insurance against emotional distress or pain and suffering.
It is easy to see why people do not buy insurance for some kinds of harm that tort law
compensates. Insurance transfers money from the uninjured state (the premiums paid by the
customers for insurance policies) to the injured state (the claims made by the injured policyholders).
For some tort claims, the cost of the transfer is very high. Thus, claims for nonpecuniary
damages are very costly to assess and administer, and negligence is costly to prove in,
say, medical malpractice cases. For other tort claims, the transfer is inappropriate because
money is not more useful in the injured state than in the uninjured state. Thus, the death of a
dependent child reduces the parents’ need for money.
In principle, allowing potential victims to sell unwanted tort rights can solve the problem of
unwanted insurance caused by tort liability. Here is how sales would work. A potential victim’s
right to damages in the event of a future accident is an “unmatured tort claim” (UTC). Imagine a
market for UTCs. Potential tort victims could sell their right to recover and could include in the sale
whichever of their tort rights they chose to sell and retain others for their own use. For example, a
victim might sell the right to recover her nonpecuniary losses in an automobile accident but retain
her right to recover her major pecuniary losses. Or she might sell the right to recover in the event
of medical malpractice but keep the right to recover in the event of a product-related injury. If
someone had sold her tort claims to a third party and was later injured, she could not recover from
the injurer; she could, however, recover from an insurance company if she had bought insurance.
A market for UTCs could be extremely flexible. Consider, for example, how a regime of
no-fault automobile insurance could result from a market in UTCs. Suppose that drivers sell
some of their rights to recover for tortious injuries in automobile accidents to their own insurance
companies. Their own insurers might then waive these rights in exchange for payment
from the insurance companies of other drivers.
If it were legal to sell and buy UTCs, potential victims would probably substitute firstparty
insurance for the current method of compensation through the tort liability system. This
first-party insurance would probably be a cheaper means of compensating victims than is the
tort liability system. But what about the deterrence function of tort law? How will the creation
of a market for UTCs induce potential victims and injurers to take care? Interestingly,
there might be no significant difference in deterrence from the current system, and there
might be an improvement. There will, after all, be someone proceeding against the injurer for
recovery in the event of an accident; it just might not be the victim. Indeed, the deterrence
effect under UTCs may be better than under the current system: Third parties who have purchased
UTCs may have a strong incentive to monitor the behavior of potential victims and injurers
for optimal precaution.
III. An Empirical Assessment of the U.S. Tort Liability System 273
Current American law prohibits victims from selling tort claims to lawyers. Thus, the
plaintiff cannot contract with her lawyer to receive a fixed fee before the trial in exchange for
giving the lawyer all, or almost all, of the damages eventually awarded by the court. Current
law, however, does not prohibit nonlawyers from buying some matured tort claims, and a
small market has already developed on the Internet.
Workers’ Compensation
The most prominent form of no-fault liability in the United States is the system for dealing
with employee accidents that occur on the job, which are common. In any given year approximately
3 percent of all industrial workers will be injured while on the job sufficiently to result
in lost work time. This risk of on-the-job accidents is slightly higher than the risk of accidents
off the job—for example, in the home.
Through the late nineteenth century, the common law of job-related accidents made
it extremely difficult for plaintiff-employees who had been injured on the job to recover
from their employers or from anyone else.65
Early in the twentieth century, most industrialized
countries, including the United States, adopted an alternative to tort liability for
dealing with on-the-job accidents—namely, a system of compulsory compensation of injured
employees without regard to fault. Today all but three states in the United States
have a compulsory system for nearly all workers, and more than 90 percent of the United
States labor force is covered by workers’ compensation systems. The systems typically
work in the following way. Employers contribute sums to the state workers’ compensation
system based on the dollar amount of their payroll. When an employee is injured, he
65
There were three defenses available to the employer: (1) common employment (also known as the “fellow
servant rule”), under which the employer could escape liability by claiming that the proximate cause of the
plaintiff’s harm was the negligence of another employee; (2) assumption of the risk, under which the employer
could argue that the employee willingly assumed the risk of a job-related injury (on the economics of jobrelated
risks, see KIP VISCUSI, RISK BY CHOICE (1983); and (3) contributory negligence, under which the employer
could escape liability by showing that the employee’s own negligence had contributed to the harm.
QUESTION 7.15: Imagine a system of contractual or elective no-fault with
respect to product-related injuries. Manufacturers would offer with their products
schedules of benefits that they would pay if consumers should be injured
while using the products. In the event of an injury, there would be no inquiry
into the product’s defect or the user’s fault; benefits would simply be paid to
the injured consumer according to the contractual schedule. Pain and suffering
would not be compensable; collateral benefits would be deducted; and a
few other restrictions would apply. Those manufacturers who chose not to offer
elective no-fault would still be strictly liable for product-related injuries
under the current system. Explore the efficiency of this elective no-fault system.
(See J. O’CONNELL, ENDING INSULT TO INJURY (1975).)
(Continued)
274 C H A P T E R 7 Topics in the Economics of Tort Liability
66
For a report on work in progress, see Alison Morantz, Rethinking the Great Compromise: What Happens
When Large Companies Opt Out of Workers Compensation?, BERKELEY LAW AND ECONOMICS WORKSHOP,
14 April 2008.
files a claim with the state governmental agency that administers the system. If the
agency determines that the harm is job related, then it awards the employee compensation
according to a statutory schedule of benefits. For example, the compensation for a
lost finger may be $5000. For other injuries, the benefits may be determined in a relatively
brief evidentiary hearing. For example, the victim may be awarded two-thirds of the
lost wages and full compensation of medical and rehabilitation expenses. In the event of
a dispute between the employee and the workers’ compensation commission, a process
of appeal and adjudication is available.
In the workers’ compensation system, workers relinquished their right to sue their employers
in tort law for on-the-job injuries and, in exchange, employers assumed strict liability
for injuries suffered on the job. Is this an improvement? Almost all states in the United States
compel employers to participate in the workers’ compensation system, but Texas allows employers
to opt out. Texas generates some data to compare a workers’ compensation system
and a system of tort liability. When large Texas firms opt out, they create their own system for
compensating injured workers. If injured workers do not want to accept compensation offered
under the employer’s plan, then they must sue the employer in tort law and prove that
the employer’s negligence caused the injury.
Morantz investigated the claims for on-the-job injuries by employees of two large companies.
These two companies opt out of workers’ compensation for the retail outlets inside
Texas, and they must remain inside this system for their retail outlets outside Texas. Morantz
found a lower frequency of indemnity claims in Texas than outside of it, and she also found
lower average per-claim costs. While she demonstrated cost savings to employers, the welfare
effects on workers remains to be determined by future research.66
Conclusion
The tort liability system plays a significant role in reducing the frequency with
which we accidentally lose our property, health, and lives. By allocating the cost of
accidents, the tort liability system provides incentives for precaution, much as markets
allocate costs and provide incentives for production. Improving the efficiency of the
tort liability system can make the world safer at no more cost. Observers note various
signs of inefficiency in tort liability law, such as significant differences in the level of
compensatory damages for the same injury in different countries of similar wealth,
unpredictable decisions about liability and damages from one case to another (“liability
disparity”), defensive medicine, and vaccine shortages. Moving beyond anecdotes
requires careful statistical studies that remain in short supply. The statistical knowledge
that we do possess at this time, which we reviewed, is enough to debunk some myths
about the tort liability system.
Suggested Readings 275
Suggested Readings
BAKER, TOM, THE MEDICAL MALPRACTICE MYTH (2005).
Bar-Gill, Oren, & Omri Ben-Shahar, The Uneasy Care for Comparative Negligence, 5 AM.
LAW & ECON. REV. 433 (2003).
Black, Bernard, Charles Silver, David A. Hyman, & William M. Sage, Stability, Not Crisis: Medical
Malpractice Claim Outcomes in Texas, 1988-2002, 2 J. EMP. LEGAL. STUD. 207 (2005).
DEWEES, DONALD, DONALD DUFF, & MICHAEL J. TREBILCOCK, EXPLORING THE DOMAIN OF
ACCIDENT LAW (1996).
GAWANDE, ATUL, COMPLICATIONS: A SURGEON’S NOTES ON AN IMPERFECT SCIENCE (2002).
Hyman, David A., Medical Malpractice: What Do We Know and What (If Anything) Should We
Do About It?, 80 TEX. L. REV. 1639 (2002).
PORAT, ARIEL, & ALEX STEIN, TORT LAW UNDER UNCERTAINTY (2002).
Schwartz, Gary T., The Beginning and the Possible End of the Rise of Modern American Tort
Law, 26 GA. L. REV. 601 (1992).
Shavell, Steven, Liability for Accidents, in A. MITCHELL POLINSKY & STEVEN SHAVELL, EDS.,
HANDBOOK OF LAW AND ECONOMICS, v. 1 (2007).
SUNSTEIN, CASS R., REID HASTIE, JOHN W. PAYNE, DAVID A. SCHKADE, & W. KIP VISCUSI,
PUNITIVE DAMAGES: HOW JURIES DECIDE (2002).