. 15
( 18)


37 Cmnd 1406, para. 13(a).

The functions of compensation systems

17.1 Compensation
17.1.1 Some preliminary questions
So far in this book the word ˜compensation™ has been used loosely and in various
contexts. We must now consider more carefully what is meant by the term, and ask
why we compensate the victims of injury and disease. One possible answer to this
second question is that widely held notions of justice and fairness demand it.
Unfortunately, however, we have very little evidence concerning what people think
about compensation for death and personal injury. Attempts have sometimes been
made to ascertain common views by survey questionnaires, but the results are not
particularly helpful. One writer concludes that ˜there seems to be rather little evi-
dence that when asked, people actually do express consensus support for a fault-
based compensation system™.1 Several (now rather old) surveys2 found that an
overwhelming proportion of those questioned were in favour of damages being
awarded for pain and su¬ering; but when it is appreciated that those questioned
were themselves recent victims of road accidents, what is surprising is not the
majority of a¬rmative replies, but the substantial minority who did not favour
such awards “ some 20% in one and around 30% in another. One US survey
devoted to pain and su¬ering found widespread misunderstanding about the way
damages are calculated and about the likelihood of receiving damages for pain and
su¬ering.3 In England, a former trade union claims o¬cial once said that people
injured in industrial accidents sometimes ˜can hardly be convinced™ that they
are entitled to claim disablement bene¬t when they have su¬ered no loss of earn-
ings.4 On the other hand, debates following disasters, such as ¬res in public
places and rail crashes, often reveal a considerable demand for increased damages

1 S. Lloyd-Bostock, ˜Fault and Liability for Accidents: the Accident Victim™s Perspective™ in Harris
1984 Survey, 143, referring to US and Canadian surveys; see also Ison, The Forensic Lottery, 217
2 Conard, Automobile Accident Costs and Payments, 265; Osgoode Hall Study, ch. VIII, section E. See
also J. O™Connell and R.J. Simon, Payment for Pain and Su¬ering “ Who Wants What, When and
Why? (Champaign-Urbana, Ill., 1972).
3 O™Connell and Simon, Payment for Pain and Su¬ering.
4 J. Bell, How to Get Industrial Injuries Bene¬ts (London, 1966), 78. But attitudes may have changed.

The functions of compensation systems 409

for non-pecuniary loss and even for exemplary damages. However, if people were
asked whether they would be willing to pay premiums su¬cient to cover increased
compensation for pain and su¬ering, it is not clear that they would answer
A major di¬culty, in basing any justi¬cation for compensation on widespread
beliefs, is that most people simply know too little about the way compensation
systems operate in detail to have any relevant views on the matter. Compensation
systems are extremely complex sets of rules, institutions and practices, and only
those with expert knowledge of the way they work are likely to have considered
views about them. Public debates about compensation are often characterized by
widespread, basic ignorance not only of the way compensation systems work and
interact, but also of the vigorous debates, about their strengths and weaknesses and
about options for reform, that have taken place around the world in the past forty-
odd years. Another di¬culty arises from the fact that to the extent that ordinary
people have views about the justice of compensation systems, they are probably
conditioned by their contact with such systems, whether directly as claimants or
defendants or indirectly through the media or the experiences of colleagues and
friends. There is evidence that the attitudes of personal injury victims to questions
of fault, responsibility and compensation are heavily in¬‚uenced by what they know
of the relevant legal rules.5 More generally, there is probably a good deal of inter-
action between the law, on the one hand, and ideas of justice and equity held by
members of the community at large, on the other.
Even when Members of Parliament (who might be thought to represent com-
mon views in some sense) discuss compensation systems, the discussion is usually
limited to one compensation system or one aspect of a system, and does not
concern itself with the complex network of systems we have surveyed in this book.
The debates on the CICS illustrate the defect of this sort of selectivity only too well.
Asked to approve a scheme for compensating the victims of crimes of violence,
both Houses gave almost unanimous approval; but the question of whether the
claims of victims of violent crimes rank higher (say) than those of accident victims
who receive socia1 security bene¬ts was not discussed. This problem of narrowness
of view is vividly illustrated by the fact that with the exception of o¬-setting of
social security payments against tort damages and recoupment of NHS costs from
tortfeasors, the wide-ranging investigation conducted by the Law Commission in
the 1990s into various aspects of compensation law has focused on the tort system
to the exclusion of the larger picture of interrelated systems of provision for the
injured and disabled. Discussion of no-fault compensation schemes tends to be
limited to areas such as medical mishaps, without consideration of larger questions
about the interrelationship of such schemes with either the tort system or the social
security system.
5 Lloyd-Bostock, ˜Fault and Liability for Accidents™ in Harris 1984 Survey; see also H. Genn, ˜Who
Claims Compensation: Factors Associated with Claiming and Obtaining Damages™ in the same
volume, esp 65“70.
410 Chapter 17

Even if we could say that the web of compensation systems we have conforms to
widespread ideas of justice, we might still want to ask whether the whole structure
operates e¬ciently, and whether the various parts of it ¬t together in a consistent
and coherent way. We can do this only when we have decided what purposes are
served by individual compensation systems and by the structure as a whole. The
¬rst thing to note in this regard is that paying compensation is a two-sided process:
money is transferred from one person or group of people to another. There are
three important questions we can ask about this process, namely who should
be paid?, what should they be paid? and who should pay? These questions are inter-
related. Take exemplary (or ˜punitive™) damages, for example. These are rarely
awarded in a tort action, and their main aim is to penalize the tortfeasor™s conduct
and to deter such conduct in the future; they do not make good any loss su¬ered
by the injured person. So why should they be paid to the claimant? If they are, will
not the claimant enjoy an undeserved windfall? And who should pay such damages?
For example, should exemplary damages ever be awarded against a person who is
vicariously liable for the tort that attracts the damages?6 Should a person be allowed
to insure against liability for punitive damages so that he or she does not pay them
personally? If the aim of such damages is primarily to punish, the proper answer
would seem to be negative in both cases; and if it is to deter, then the answer should
be negative in the case of insurance, and also in the case of vicarious liability, unless
it is thought that the vicariously liable person ought to exercise some control over
the conduct of the tortfeasor so as to prevent the tortious conduct.7
Or take the case of damages for non-pecuniary loss (pain and su¬ering and loss
of amenities): if we ask whether injury victims ought to be paid such damages, the
answer may very well depend on who will pay them. If they will be paid by a person
who was seriously to blame for the accident, we are more likely to feel that such
damages should be awarded than if they will be paid by the taxpayer or by some
large group of persons in no way responsible for the accident (such as insur-
ance premium payers). For example, the Working Party on Compensation for the
Victims of Crimes of Violence8 expressed grave doubts about the idea of the State
paying compensation for non-pecuniary loss except as part of a contributory
When we turn to damages for ¬nancial loss, we may not be too concerned about
who will pay; our main concern may be that the victim is actually paid by someone
and that what is paid represents all, or a substantial proportion of, the losses
su¬ered. In this context, once we know that any compensation will be ¬nanced by
quite small payments from each of a large group of people (taxpayers or insurance
premium payers) we may become even less concerned about who pays and focus

6 P.S. Atiyah, Vicarious Liability in the Law of Torts (London, 1967), ch. 39.
7 For a recent discussion of all of these issues see Law Com. No. 247, Aggravated, Exemplary and
Restitutionary Damages (1997), Parts IV“VI. The Commission answered both of the questions
posed in the text a¬rmatively.
8 Cmnd 1406 (1961), para. 48.
The functions of compensation systems 411

our attention almost entirely on who and what is paid. This has happened to a large
extent in the social security system, but in the tort system, much attention is still
given to the question of who pays: the complex rules of tort liability ensure this.
In answering the question of what will be paid, there can be no doubt that the
types of misfortune which a society is willing to treat as compensatable, and the
amount of compensation it is willing to award, will depend to a signi¬cant extent on
the wealth of that society. A very poor society in which people die of starvation or
malnutrition and many live in conditions of great need is unlikely to be willing or
able to devote much of its resources to compensation for illness and injury. Even in
wealthy societies, the amount we are prepared to spend on compensation systems is
limited. For this reason, if for none other, it is of vital importance to ensure that
money available for compensation is spent in the best possible way.

17.1.2 The meaning of ˜compensation™
The legal notion of ˜compensation™ is a complex one.9 Lawyers usually talk about
compensation ˜for loss™; but as we will see, not all forms of compensation are con-
cerned with ˜loss™ in any common sense of that word. To compensate a person is to
make good an undesirable aspect of their circumstances or situation in life which
falls below some pre-determined benchmark of acceptability. A useful distinction
can be drawn between two di¬erent types of compensation according to the bench-
mark they use. We shall call these respectively ˜corrective compensation™ and ˜redis-
tributive compensation™. Corrective compensation
Corrective compensation takes as its benchmark the situation in which the person
to be compensated was at some earlier stage of their life. This is the sort of com-
pensation tort law provides. As we saw in chapter 6, tort law compares the position
the injured person was in immediately before the tort occurred with their position
after the tort, and aims to restore them to that earlier position. Corrective compen-
sation is essentially backward-looking. It seeks to protect people from the e¬ects of
adverse changes in their circumstances and to maintain continuity and stability in
their lives. Tort law provides corrective compensation because its concern is to hold
people responsible for injuries caused by their acts and omissions. Tort compensa-
tion is backward-looking because tort liability is backward-looking. Redistributive compensation
By contrast, the benchmark used by redistributive compensation is not some
position which the person to be compensated formerly occupied, but rather the
position that other people now occupy. Redistributive compensation does not
compare earlier and later positions of the same person, but contemporaneous posi-
tions of di¬erent people. This is the main sort of compensation provided by the

9 R.E. Goodin, Utilitarianism as a Public Philosophy (Cambridge, 1995), chs. 11“13.
412 Chapter 17

social security and social welfare systems.10 The basic aim of these systems is to
reduce di¬erences between individuals by transferring resources from the more to
the less a¬„uent, from the healthy to the sick, from the fortunate to the unfortunate.
Redistributive compensation is essentially forward-looking. It is not concerned
with making up for the past but with improving people™s lives in the future. The
social security system, for instance, is not concerned with how a person came to be
in the position they are in but with whether fairness and humanity demand ame-
lioration of that position.
Another useful distinction can be drawn between what we might call ˜equivalent
compensation™, on the one hand, and ˜substitute/solace compensation™, on the other.
This distinction is based on the fact that the compensation provided by the tort
system is in the form of monetary payments, but that not all of the adverse changes
for which tort law compensates are ¬nancial in nature. When tort law gives mone-
tary compensation for adverse ¬nancial changes in a person™s life, we can say that
the compensation is equivalent to that which is being compensated for; but this is
not the case when those changes are not ¬nancial. Equivalent compensation
There are at least three di¬erent types of equivalent compensation. First, a person
may be compensated for having been deprived of money or some other asset that
can be fully replaced with money. A person loses wages when away from work as a
result of an accident; or a person™s car is wrecked in an accident and they need to
buy a replacement; or a house is destroyed by ¬re and the owner must rebuild or
buy a new house. In such cases, monetary compensation can give the person
back exactly what they have ˜lost™ or enable them to acquire an exactly similar
The second type of equivalent compensation is designed to meet costs incurred
by a victim of injury or damage. Such costs may take various forms from medical
expenses to the cost of hospital visits or the cost of modifying a house to make it
easier for the victim to live in. Here again, money can make good the adverse change
in the claimant™s position, namely the need to incur expense which was formerly
The third type equivalent compensation is concerned with lost expectations,
chie¬‚y the expectation of being able to earn in the future. This is compensation for
loss of the capacity to earn as opposed to compensation for earnings lost in the past.
We might ask why a person should be compensated for loss of earnings when that
person will never render the services for which the earnings are payment. One good
answer is that the person has been deprived of the choice whether or not to exercise
their earning capacity. In many cases the best evidence available of the value of this
capacity is what the person was earning before being incapacitated. But in some

10 However, there are pockets of corrective compensation in cause-based schemes such as the IIS and
the CICS.
The functions of compensation systems 413

cases (e.g. children) the court has to speculate about what the victim™s capacity
would have enabled them to earn had they not been injured. If an injured person
has never worked, and it is clear that they would never have worked even if they had
not been deprived of the capacity, no compensation for loss of future earnings
would be awarded. Another good reason for compensating for loss of future earn-
ings is to provide stability and continuity in people™s lives. This partly explains why
tort law adopts the earnings-related principle for the assessment of damages: it
aims to protect people from changes in their life circumstances.
Although equivalent compensation is similar in nature to that which is com-
pensated for, it is not necessarily equal in ¬nancial value. Both the tort system and
the social security system provide equivalent compensation; in other words,
equivalent compensation may be either corrective or redistributive. But whereas
the tort system is committed to the ˜full compensation™ principle in relation to loss
of income (for instance), the social security system is not. Compensation as substitute and solace
Tort law compensates not only for ¬nancial changes in a person™s life, but also for
adversities such as pain and su¬ering and loss of amenities which cannot, in any
meaningful sense, be valued in money. In relation to such adversities, the compen-
sation the law provides is obviously di¬erent in nature from that which is being
compensated for. The object of such compensation is to enable the injured person
to obtain a substitute source of satisfaction or pleasure (where some ˜amenity™ has
been lost), or alternatively to comfort the victim or provide him or her with solace
for what has happened (as in the case of pain and su¬ering). This type of compen-
sation is most commonly awarded in cases of personal injury (although damages
for inconvenience and mental distress are increasingly being awarded in cases
involving claims arising out of property damage or ¬nancial loss).
There is only one situation in which tort law explicitly awards damages as a solace,
namely when it awards damages for bereavement under the Fatal Accidents Act 1976.
But in some other types of case “ where, for example, a person loses their sense of
smell “ it is di¬cult to think of anything that would count as a substitute. Even where
substitute pleasures can be found, they are almost bound to be only partial. So com-
pensation for lost amenities is often wholly or partly solace for what has been lost.
Damages for pain and su¬ering can really only be understood as providing solace.
Few compensation systems provide substitute/solace compensation. For example,
cover under personal accident insurance policies is usually limited to medical
expenses or income losses; and though disability payments are often made under
comprehensive road accident insurance policies, they are usually very small. Apart
from the tort system and the CICS, the only system that gives signi¬cant substitute
or solace compensation is the industrial injuries scheme; and this is the sole instance
of such compensation in the social security system.
It might be thought that substitute compensation requires much greater
justi¬cation than equivalent compensation. There is, for one thing, the di¬culty of
414 Chapter 17

¬xing the level of compensation. What is a reasonable substitute for a pleasure
foregone? How can we measure the amount of pleasure or happiness a person
derives from this or that activity? Should we make some estimate in money of the
subjective value to the victim of various forms of activity, or should we look at
objective costs? Should we deduct, from the value of the pleasure foregone, the cost
of obtaining it?
Solace compensation, it might be thought, is even harder to justify than substi-
tute compensation. If such compensation were actually paid by the tortfeasor
personally, it might be supportable on grounds of fairness. But since such com-
pensation will usually be paid out of insurance premiums by a substantial section
of the public, one is forced to ask whether there are not other claims on society™s
resources that deserve priority. This point applies to both substitute and solace
compensation, but more so to the latter. It is hard to justify compensation for
mental distress and deprivation of pleasure when many disabled people receive
little or no compensation even for income losses. On the other hand, it might be
thought that physical pain does deserve legal recognition in its own right, at least if
it is severe.

17.1.3 Compensation and compensation systems
There are, perhaps, three main criteria for judging the success of a system as a
mechanism for compensating for personal injuries and death. First, how many
of those who are entitled to compensation according to the rules of the
system actually receive compensation and in the amount to which they are enti-
tled? Secondly, how many people, by fraud or abuse of the system, receive com-
pensation to which they are not entitled? Thirdly, what are the administrative
costs of the system? As regards this last criterion, we have seen that the adminis-
trative costs of the tort system are very high relative to the costs of the social secu-
rity system and the CICS. As for the second criterion, the reader should refer to
the discussion in 13.8.
Concerning the ¬rst criterion, it is di¬cult to judge how successful the tort
system, the CICS or the social security system is because we lack reliable informa-
tion about how many potential recipients of compensation from the various
systems make no claim or about how many good claims are rejected. There is
reason, however, to think that all of these systems fail to reach a signi¬cant pro-
portion of those entitled under them. As for the amount of compensation, we have
seen that one result of the process of settling tort claims is that very many claims
are settled for less than they are worth according to the rules of the system, and that
a signi¬cant proportion are probably settled for more than they are worth. By con-
trast, the CICS and the social security system are, subject to administrative errors,
much more likely to deliver to claimants the ˜correct™ amount of compensation. On
the other hand, it might be thought that, in theory at least, the tort system is ˜better™
than the CICS or the social security system because unlike those systems, it aims to
provide ˜full compensation™ for death and personal injuries.
The functions of compensation systems 415

17.2 Distribution of losses
17.2.1 What should be distributed?
Compensation is, by de¬nition, one of the functions of compensation systems. But
they may have others. A common suggestion is that the distribution of losses is an
important function of compensation systems. But to say this begs at least three
important questions. The ¬rst is that of how we are to de¬ne ˜loss™. A ˜loss™ is not
something that exists or occurs outside the law and for which the law simply pro-
vides compensation. It is the law itself that de¬nes what is meant by ˜loss™. In some
respects, the law™s use of the word ˜loss™ is odd. Outside the law, for instance, we
would probably describe the su¬ering of pain as ˜harm™ rather than ˜loss™. In tort law,
˜loss™ refers broadly to adverse changes in a person™s life circumstances. However,
the legal meaning of ˜loss™ is not simply a question of proper linguistic usage but
also of policy. For example, it is only relatively recently that English common law
has recognized that a domestic carer who is rendered incapable of providing such
care has su¬ered a loss for which damages may be awarded; again, the loss caused
by bereavement was recognized for the ¬rst time by English law in 1982. A striking
example is provided by the need of an injured person to be nursed: if a relative or
friend does the nursing gratuitously, the law treats the victim as having su¬ered a
loss (assessed as the reasonable value of the nursing services) for which damages
can be awarded; but if the victim is nursed for free in an NHS institution, no
damages are awarded for the value of the nursing. It is, therefore, not possible to
de¬ne what is meant by ˜loss™ “ all one can do is to describe the losses which the law
recognizes as proper subjects of damages awards.
A second important question begged by speaking of the goal of the law as loss
distribution is how we are to value losses. We have seen, for example, that tort law
adopts the ˜full-compensation™ and ˜hundred-per cent™ principles in relation to
pecuniary losses, and a tari¬ system for the assessment of non-pecuniary losses.
On the other hand, for example, social security systems never compensate for
income losses in full. A di¬erent type of valuation issue arises in relation to com-
pensation for loss of ability to keep house or for gratuitous nursing services.
Suppose the domestic carer or nurse gives up a job or foregoes the opportunity of
working in order to work in the home or to nurse. Should the services be valued at
what it would cost to employ someone to perform them, or at what the person
doing them could have earned at work? Tort law has not explicitly committed itself
to either of these measures of value but awards what, in the particular case, the
court considers to be the ˜reasonable value™ of the services.
The third and most fundamental question begged by speaking of loss distribu-
tion as a goal of a compensation system is whether it is ˜losses™ with which the
system should be concerned. The idea that losses should be our concern is to some
extent a corollary of the fault principle and of the individualistic nature of tort law.
Social security systems tend to be concerned essentially with meeting basic
¬nancial ˜needs™ by means of ¬‚at-rate bene¬ts. The idea that losses should not
416 Chapter 17

always be the law™s focus has been the subject of some debate in the tort system: in
one case Lord Denning argued unsuccessfully that an injured person in a state of
(almost) total incapacity who has no dependants should not be awarded damages
for loss of income in addition to adequate damages for the cost of care because, in
e¬ect, the person did not need and could not use them, and they would be simply
a windfall to her relatives.11

17.2.2 How should it be distributed?
Compensation for loss can be said to involve shifting the loss from one person to
another. But where the loss is not of money or money™s worth (e.g. pain and
su¬ering), the process of making some person compensate the victim for the mis-
fortune is very inaccurately described as ˜shifting the loss™. Losses of this kind cannot
be shifted from one person to another in any meaningful sense. It may be possible
to minimize pain and su¬ering, for example, by medical treatment, and it may be
possible to make someone else pay for this medical treatment; and this may,
perhaps, be regarded as ˜shifting a loss™. But when all has been done to minimize the
pain and su¬ering by medical means, any residual pain and su¬ering cannot be
shifted: it remains with the victim, no matter what compensation is paid to that
person by others.
The shifting of a loss “ or making one person compensate another for some mis-
fortune “ involves an alteration of the status quo; and so it involves administrative
expense. Therefore (it is usually asserted), the onus is on those who wish to shift a
loss to justify the shift. Unless there is some good reason for shifting a loss, it should
be left to lie where it falls. Tort law mostly attempts to justify the shifting of losses
by reference to the fault principle, but this has become increasingly unattractive to
many people. Tort lawyers have searched for something to put in its place, and some
have found their answer in the idea of ˜loss distribution™. The e¬ect of the tort
system is not, in general, merely to shift a loss from one person to another; the loss
is normally distributed over a large number of people, and over some period of
time. It is true that this distribution is not normally achieved by rules and practices
which are regarded as part of tort law; but the combined practical e¬ect of tort law,
liability insurance and the operation of the market is, in practice, to distribute
losses among a large group of people and over a period of time.
There is no doubt that the distribution of losses is an important and usually
desirable result of successful tort claims. Losses that may be crushing if imposed on
an individual can be borne easily when distributed amongst a large group of
people; this is clearly in the interests of those who are themselves at risk, as well as
of society as a whole. There is also a great gain in security and peace of mind when
the fear of crushing losses from sudden disaster is displaced by knowledge that the
loss will be spread. But to advocate loss distribution as an end in itself invites as an
obvious retort: loss distribution among whom? There are many ways in which

11 Lim Poh Choo v. Camden AHA [1979] 1 All ER 332.
The functions of compensation systems 417

losses can be distributed; and “ which is merely saying the same thing in another
way “ there are many ways in which money can be raised from members of the
public. Tort law, interwoven as it is with liability insurance, provides one way. First-
party insurance provides another way. Social security provides still another way of
distributing losses, ¬nanced as it is by a mixture of insurance and taxation.
These are not the only ways of distributing losses. All sorts of permutations and
combinations are possible. For example, compensation for road accidents could be
¬nanced by a fund contributed to solely by motorists, but by ¬‚at-rate contributions
instead of by variable insurance premiums as at present; or it could be ¬nanced by
a special tax on petrol, so that (in general) motorists who used the roads more
would pay more; or the fund could be contributed to by all road users, including
cyclists and pedestrians “ though this large group is so nearly co-incident with the
entire population that it would be tantamount to an ordinary tax. Or again, a
special levy could be imposed on motor manufacturers and perhaps also on
highway authorities, which could be paid into the fund. Much would no doubt
depend on how compensation was to be assessed. If it involved paying variable
compensation for income losses, it might be thought equitable to make people pay
contributions to the fund according to their income. Alternatively the whole
process could be ¬nanced out of general taxation, and no special tax imposed for
the purpose. Because there are so many ways of distributing losses, it cannot be said
that loss distribution as such is a rational and desirable goal of the law. Any loss dis-
tribution system must be judged according to the way it distributes particular
losses. For example, few would ¬nd acceptable a system which distributed the costs
of road accidents entirely amongst non-motorists or the costs of smoking-induced
cancer entirely amongst non-smokers.
Two things about the tort system seem clear, however. The ¬rst is that loss dis-
tribution cannot be the sole purpose or justi¬cation of the tort system, because
both in theory and in practice tort law allows recovery for only a small proportion
of personal injury losses.12 Secondly, it seems clear that there are much cheaper and
more e¬cient ways of distributing the losses the tort system does deal with than the
present combination of liability rules and liability insurance.
How, then, are we to choose between methods of loss distribution? Or, to put the
question di¬erently, how ought a system of loss compensation to be funded? There
seem to be three broad options: ¬rst-party loss insurance (whether private or State-
run (National Insurance)), which spreads the loss amongst potential victims; third-
party liability insurance, which spreads the loss amongst those likely to in¬‚ict it;
and general taxation. Each of these mechanisms can be used alone or in combina-
tion with other methods. Just as important as the method of funding is the ques-
tion of whether the compensation paid out, on the one hand, and the contributions
to the fund, on the other, are to be the same for all bene¬ciaries or contributors (i.e.
12 For the same reason, it makes no sense to say that compensation for death and personal injury is
the goal of the tort system because even in theory, tort law only compensates for death and per-
sonal injury which comes about in certain ways.
418 Chapter 17

¬‚at-rate) or variable in some way (e.g. income-related).13 A third important ques-
tion, which cuts across the ¬rst two, is whether losses ought to be distributed by free
market mechanisms (which would make people pay for what they enjoy and for the
losses they cause), or by State-run schemes or by a mixture of the two. State-run
schemes can be used to achieve other ends in addition to loss distribution, such as
income redistribution in favour of the poor.
Another criterion relevant to choosing between methods of distribution is
e¬ciency. There are several aspects to the question of e¬ciency. One is the admin-
istrative cost of the process; and as we have seen, the administrative costs of
the social security system and of the CICS are much lower than those of the tort
system. The comparisons are slightly misleading because a signi¬cant part of the
cost of the tort system is attributable to the cost of collecting the money used to pay
compensation, while the cost of collecting social security payments and criminal
injuries compensation is borne partly by employers (in the case of National
Insurance contributions and income tax) and partly by the tax system, and so it
does not ¬gure in the administrative costs of providing bene¬ts or compensation.
But even making allowance for this, it is clear that, on cost alone, the tort system is
extremely ine¬cient.

17.3 The allocation of risks
Another commonly suggested purpose of the law is to allocate the risk of the occur-
rence of certain events between various parties rather than to decide whether one
person has caused injury to another by fault or whether one person should com-
pensate another for a wrong. The idea of risk allocation seems particularly appro-
priate to cases of strict liability in tort law and to many areas of the law of
compensation outside the tort system, such as the industrial injuries scheme. For
example, the basis of the principle in Rylands v. Fletcher is that a person who col-
lects dangerous substances on land should bear the risk of their escape whether or
not the escape was that person™s fault.
It was in the ¬eld of workers™ compensation that the idea of risk allocation ¬rst
made a powerful impact. It was felt to be unjust that the whole burden of accidents
should lie on the injured worker, and that it was immaterial that the accidents were
or were not caused by fault. These risks were felt to be risks of the business.14
Similarly, the vicarious liability of an employer for the wrong of its employees has
often been justi¬ed by invoking the idea that the employer should take the risk of
its employees causing injury or damage.15 This is why strict liability is often referred
to as ˜enterprise liability™.

13 It should be noted that ¬‚at-rate levies are ˜regressive™ in the sense that they place a greater rela-
tive burden on those who are poorer; by contrast, income-related levies bear less heavily on
the poor.
14 P.S. Atiyah, Vicarious Liability in the Law of Torts (London, 1967), 22“4.
15 Ibid.
The functions of compensation systems 419

Of course, saying only that the law is concerned with allocating risks does not
answer the question of to whom particular risks ought to be allocated or why
certain risks are allocated in a particular way. This question alerts us to the fact that
although the language of risk allocation is often associated with strict liability in
tort law, the di¬erence between fault-based and strict liability is not that the latter
allocates risks of injury but the former does not, but that they allocate risks of harm
in di¬erent ways. When, for instance, a pedestrian is injured on the road without
fault on anyone™s part, why should the risk of such injury rest on the pedestrian
rather than on motorists? Or if a person is injured as a result of an explosion of gas
escaping from a fractured main, should the risk of such injury be borne by the gas
undertaker whether or not negligence can be proved?16 Underlying many judg-
ments about how risks should be allocated lies the notion that a person who
˜creates™ a risk should be made to bear the cost of the risk; and the idea of ˜creating™
a risk is often based on causal concepts. A gas board whose leaking pipes result in
an explosion that destroys a house is thought of as having ˜caused™ the destruction
of the house, and so to have ˜created™ the risk of such damage. But in the absence of
fault, there may be no greater reason for saying that the destruction of the house
was ˜caused™ by the presence of the gas pipes than by the presence of the house itself.
And if the house was built after the pipes were laid, there may be no more reason
to say that the risk was ˜created™ by the gas board than by the house owner. If we are
to justify allocation of risk to one party or the other, we need some better criterion.
In tort law, the rule in Rylands v. Fletcher, for instance, is based on the idea that
people who carry on very dangerous activities should bear the risks of those activ-
ities; and vicarious liability rests on the idea that people who carry on activities for
pro¬t should bear the risks of those activities. Another possible approach is to
impose liability on the party in the better position to minimize the risk, so as to give
that party an incentive to do so. A di¬erent approach would be to ask which party
would be in a better position to distribute the loss if the risk materialized.
So, as with loss distribution as a goal of the law, risk allocation as such makes no
sense as an objective of a compensation system. A scheme of risk allocation must
be judged according to how risks are allocated.

17.4 Punishment
Punishment of wrongdoers (and by this means expressing disapproval of what they
have done) is widely accepted as a legitimate function of the criminal law; but the
˜conventional wisdom™ is that punishing tortfeasors is not the reason why they are
obliged to pay damages for injury and damage in¬‚icted by their torts.17 The House
of Lords has, therefore, de¬ned narrowly the circumstances in which ˜punitive
damages™ (i.e. damages designed to punish the tortfeasor rather than to compensate

16 See Dunne v. NW Gas Board [1964] 2 QB 806.
17 Rookes v. Barnard [1964] AC 1129; Broome v. Cassell [1972] AC 1027.
420 Chapter 17

the injured person) are recoverable in a tort action.18 Unless there is express statu-
tory authority, punitive damages can be awarded in only two situations: ¬rst, where
a person or body exercising governmental powers has been guilty of arbitrary,
oppressive or unconstitutional conduct; and, secondly, where the wrongdoer has
sought to make a pro¬t out of the tort. These two grounds for awarding punitive
damages were not seen by the Law Lords as based on a rational foundation; indeed,
the judges would probably rather have held that punitive damages were never avail-
able in a tort action, but they felt compelled by earlier cases to reach the decision
they did.
Not all would agree with this restrictive approach to punitive damages. In the
USA, for instance, punitive damages are sometimes awarded against corporate
defendants, especially in unfair competition, product liability and environmental
pollution cases; and some in this country have urged that punitive damages should
be more widely available in personal injury actions, particularly those in which cor-
porate or governmental defendants have been responsible for what is seen as seri-
ously culpable inattention to the safety of members of the public. The main
arguments used against punitive damages are that they amount to a criminal ¬ne,
but one which is imposed without the procedural safeguards for defendants which
are built into the criminal process; and that damages which do not represent any
loss su¬ered by the claimant are a ˜windfall™, and so unjusti¬able.
Apart from the two grounds already mentioned on which punitive damages
can be awarded, there are certain other situations in which the law of torts might
be seen as performing a punitive function. First, courts are sometimes prepared
to award what are called ˜aggravated damages™ in cases where the tortfeasor™s con-
duct towards the victim was particularly outrageous or humiliating; such dam-
ages are said to be ˜compensatory™, not punitive, but they are awarded over and
above ordinary compensation and are really indistinguishable from punitive
damages.19 Such damages may be awarded even if the tort victim su¬ers no com-
pensatable loss, but only humiliation, outrage or indignity. It is doubtful whether
wilfully or intentionally in¬‚icting outrage or indignity on a person is tortious in
itself (although it is in the USA);20 but even in England such conduct would be
actionable where the defendant has been guilty of conduct falling under one
of the traditional heads of tort liability such as fraud, battery or libel. So, for exam-
ple, if a person commits a battery by spitting at another, it is inconceivable that a
judge would send the victim away with nothing but nominal damages. Another
part of the law of torts that may be seen as performing a partly retributive func-
tion is the tort of defamation, although in theory this tort is designed to com-
pensate for ˜loss of reputation™.
The Law Commission has recently given detailed consideration to the law of
punitive damages. Its main recommendation was that punitive damages should be
18 Rookes v. Barnard [1964] AC 1129.
19 But see Law Com. No. 247, Aggravated, Exemplary and Restitutionary Damages (1997), Part II.
20 Prosser and Keeton on the Law of Torts, 5th edn (St Paul, Minn., 1984), 57¬.
The functions of compensation systems 421

available in cases where the wrongdoer has shown ˜deliberate and outrageous dis-
regard of the claimant™s rights™. Such a case could, in theory anyway, be a case of
death or personal injury; but it is likely that the proposal would have no impact on
the vast majority of personal injury claims. On the other hand, it would make pun-
itive damages available in some types of case in which they cannot currently be
awarded: where, perhaps, an employer deliberately cuts corners on safety for the
sake of pro¬t.

17.5 Corrective justice
As we have seen, punishment can rarely be a feature of a personal injury action.
Because damages for personal injuries aim to be compensatory we may, perhaps,
describe the function of the law in terms of corrective justice “ by awarding com-
pensation the law aims to restore and redress the balance of fairness or justice which
the tortfeasor has upset by negligence or by creating a risk of injury. There is a huge
modern literature on corrective justice in tort law, and it contains many di¬erent
de¬nitions of the concept. However, the basic idea behind it is fundamental to
understanding tort law. Personal injury tort claims involve one individual seeking
damages from another individual on the basis that the latter is responsible in some
sense for the former™s injuries. In tort law, a person will be responsible for another™s
injuries only if they ˜caused™ those injuries in the sense of ˜cause™ adopted by tort law.
Normally, too, responsibility for personal injuries in tort law rests on fault. The real-
ization of whatever other functions we may attribute to tort law or look to it to
achieve is necessarily constrained by the fact that tort liability for personal injuries
depends on responsibility for those injuries. For instance, tort law compensates
people for losses, but only if responsibility for those losses can be pinned on some
other individual. Again, we may hope that tort law will reduce accident levels; but
its ability to do so is limited by the fact that it is only concerned with accidents that
result in injuries for which someone other than the victim is responsible. And even
if we accept punishment as a proper function of tort law, performance of that func-
tion is necessarily restricted to people responsible for causing injury to others.
In short, tort law is based on ideas of personal responsibility for the adverse
e¬ects of a person™s conduct on others; and one of its functions is to express and
give e¬ect to such ideas. In this respect, it is quite di¬erent from the social security
system (as well as from ¬rst-party insurance). Social security bene¬ts are not
claimed from individuals but from the State; and the basis of entitlement is not that
the State is responsible for the claimant™s plight. Nor is the obligation of citizens to
˜contribute™ to social security funds based on such responsibility. One might say
that the social security system involves compensation without responsibility. A
major theme of this book has been that the requirement of responsibility in tort
law is one of the main sources of the gross ine¬ciency of the tort system of com-
pensating for personal injuries. In this context, at least, the cost of corrective justice
may be unacceptably high. Ironically, at the same time the development of liability
422 Chapter 17

insurance has seriously undermined the ability of tort to operate as a corrective
justice system.21
Further doubt is cast on the capacity of the tort system to achieve the corrective
justice aspirations of tort law by the fact that only a small proportion of accident
victims secure compensation through the system and the fact that there is no direct
relationship between accident victims™ decisions to make tort claims and notions of
responsibility.22 In short, while the rules and principles of tort law are based on
ideas of corrective justice, ˜the notion that the tort system adjusts the relationship
between victim and harm-doer is a ¬ction™.23

17.6 Vindication or satisfaction
Tort victims (and their relatives and friends) are often angry and resentful. Com-
pensation can go some way to removing such feelings, as can punishment of the
wrongdoer (whether by an award of aggravated or punitive damages, or by some
other penalty). Often such feelings arise out of ignorance about what happened,
which may be accompanied by refusal on the part of the tortfeasor to admit any
responsibility.24 In such circumstances a formal or informal inquiry aimed at
¬nding a convincing explanation of what went wrong can often do much to assuage
feelings of anger and resentment. It may be a great satisfaction to a personal injury
victim (and to other interested parties) to be able to demand an ˜o¬cial™ inquiry
into what happened. Such demands may be met by the setting up of a public
inquiry. Inquiries of this sort may take place even when legal liability is not con-
tested, so as to expose the wrongdoer to public scrutiny,25 to apportion blame (and
praise) formally and to explore ways of preventing similar occurrences in the
future.26 Such public inquiries are designed to ascertain facts and, sometimes, to
ascribe responsibility for them. They do not award damages or grant any other legal
remedy, although the ¬ndings of public inquiries often greatly assist the process of
evidence-gathering for the purposes of the settlement or trial of legal claims.
Indeed, tort claimants whose claims are investigated by a public inquiry may be put
in a much stronger bargaining position than other claimants, and may have many
of the costs of constructing their claims met by the taxpayer.
Ordinary litigation, which is usually conducted in public and which may attract
a certain amount of media attention, can also satisfy the desire that wrongdoers be

21 For further development of this argument see P. Cane, The Anatomy of Tort Law (Oxford, 1997),
ch. 7.
22 Lloyd-Bostock, ˜Fault and Liability for Accidents™ in Harris 1984 Survey, 160.
23 Ibid.
24 A. Simanowitz, ˜Accountability™ in C. Vincent, M. Ennis and R.J. Audley eds., Medical Accidents
(Oxford, 1993), ch. 14.
25 Otherwise the wrongdoer may settle the claims with a minimum of publicity and without having
to consider their future conduct.
26 For a discussion of public railway inquiries see B. Hutter,˜Public Inquiries: the Case of the Railway
Inspectorate™ (1992) 70 J. of Public Administration 177.
The functions of compensation systems 423

made publicly accountable. Indeed, it has been argued that tort law can play a role
as a public grievance mechanism similar to an ombudsman, especially in cases
against public authorities or large corporations (such as drug companies or trans-
port undertakings) whose actions have caused widespread damage or injury to
many people.27 In such cases a tort action may serve as much to establish respon-
sibility and to vindicate feelings of outrage and grief as to obtain compensation.
As compared with a public inquiry, tort has the attraction that the citizen can set
the system in motion and does not have to wait for the government to act. If the
tort system of compensation for personal injuries were ever abolished entirely, it
might be thought desirable to institute some procedure for citizen-initiated
inquiries of this type.
However, the potential of the tort system for a¬ording public vindication is
limited. Most tort actions attract very little publicity even if they go to trial; and
the vast majority are settled out of court by private agreement. The real defendant
is usually an insurance company, not a tortfeasor, and the purpose of most tort
claims is simply to unlock the door to an insurance fund. It is not uncommon for
settlement agreements to contain clauses by which the claimant agrees not to pub-
licize the grounds or terms of the settlement.28 Furthermore, settlements are typ-
ically made ˜without admission of liability™ on the part of the tortfeasor. The low
value which the tort system law places on vindication is further shown by the fact
that if the defendant o¬ers su¬cient compensation to the claimant by way of set-
tlement of the claim but the claimant rejects the o¬er and insists on a trial out of
a desire for public vindication, they will normally have to pay the costs of the
hearing (10.4). So far as the legal system is concerned, settlement out of court is
better than trial in court. Even so, there is empirical evidence that some people
view the very making of a tort claim as an aggressive act, and this may partly
explain why tort claims are very rarely made, for instance, against friends, rela-
tives and even employers.29 From this point of view, whatever vindicatory capac-
ity the tort system has does not support but interferes with its compensatory
Public vindication is not provided at all by compensation systems, such as
personal insurance and social security, in which entitlement to compensation does
not depend on establishing legal wrongdoing, where the compensation is sought
from and paid by persons in no way responsible for the loss su¬ered, and where
entitlement to compensation is determined by an administrative process con-
ducted in private. Appeal hearings before tribunals and Criminal Injuries Com-
pensation Appeals Panel adjudicators normally attract no publicity at all; indeed,
criminal injuries appeal hearings are held in private.

27 A. Linden, Canadian Tort Law, 5th edn (Toronto, 1993), 20“7.
28 But if a case is settled after a draft judgment has been sent by the court to the parties, they cannot
prevent its being delivered in open court: Prudential Assurance Co. Ltd v. McBains Cooper [2000]
1 WLR 2000.
29 Lloyd-Bostock, ˜Fault and Liability for Accidents™ in Harris 1984 Survey, 155.
424 Chapter 17

17.7 Deterrence and prevention
One of the most important of the suggested functions of personal injuries com-
pensation law is deterrence of potentially injury-causing conduct and the preven-
tion of injury-causing incidents such as accidents. A distinction is sometimes
drawn between speci¬c and general deterrence. Speci¬c deterrence involves the
express prohibition or regulation of dangerous conduct or activities by means of
statutes or regulations, typically backed up by criminal sanctions for non-
compliance, in order to reduce the number of injuries and injury-causing inci-
dents. General deterrence involves the use of compensation rules to provide
indirect incentives to people to behave safely “ the basic idea is that the prospect of
having to pay damages for injuries caused by particular conduct will deter people
from engaging in conduct of that type.
This idea of general deterrence can be given a more or less precise interpreta-
tion. The less precise version says that by establishing rules and standards of con-
duct and by attaching the sanction of damages (or, in the case of victims, a
reduction of damages) for failure to satisfy those standards and rules, the law can
provide incentives to safe conduct. This version also sees the law as performing an
educational function. The more precise interpretation, which is based on economic
principles, says, broadly, that if the cost of injuries in¬‚icted by an activity is required
by law to be paid by those who engage in that activity, they will take precautions to
prevent in¬‚icting injuries, provided the cost of precautions is less than the cost of
the injuries (i.e. the damages they have to pay); and that in this way the optimum
or e¬cient level of precautions (and, conversely, of injuries) will be reached. If the
cost of precautions exceeds the cost of the injuries, so the theory goes, precautions
will not be taken; but because those who engage in the activity are required to bear
the cost of the injuries, the activity will become more costly. This will deter people
from engaging in it, at least on the same scale as before, and will, consequently,
reduce the number of accidents or injuries caused by it. This more precise version
is sometimes called ˜market deterrence™, but more commonly ˜general deterrence™;
and in what follows the term ˜general deterrence™ will normally be used to refer to
the more precise version.
In this book we are primarily concerned with the two versions of general deter-
rence because they are about the way compensation systems can be used to promote
safety. It should be noted that compensation systems are not the only, or even the
most important, means of deterrence and injury prevention. In the two main ¬elds
in which personal injuries are compensatable by common law damages, that is road
accidents and workplace injuries, other methods are at least as important. On the
road, reliance is placed on a combination of criminal penalties “ ¬nes for reckless,
careless, dangerous and drunken driving and disquali¬cation from driving; and
improved road and vehicle design. In the case of workplace injuries, reliance is placed
on the criminal law and on education and publicity. On the other hand, it may be
that compensation law is more e¬ective as a deterrent in the latter case than in the
The functions of compensation systems 425

former, because in these cases the potential tortfeasor will very often be a business
concern, and business concerns are probably more sensitive to the incentives that tort
law provides than are individuals. This will be discussed in greater detail later.

17.7.1 Rules and standards of behaviour
The less precise version of general deterrence has two aspects: ¬rst, that of deter-
ring people from conduct which may injure others;30 and secondly that of deter-
ring people from conduct which may injure themselves. Causing injury to others
If we make people pay for damage or injury they cause to others, they may try to
cause less damage or injury. The e¬ectiveness of the tort system as a deterrent
depends crucially on the ability of the potential tortfeasor to take steps in advance
to prevent the damage or injury occurring. There are several aspects to this. One,
as we saw earlier (7.6), is that by allowing victims free choice as to who is and is not
sued and by concentrating on the parties to the action, tort law tends to ignore
other factors which may be responsible for accidents in many cases, such as the state
of roads. In order to reduce accidents and injuries it is necessary to study their
causes very carefully.
If we try to ¬nd out why a particular accident occurred, we can seldom pinpoint a single
cause. Nearly always, it might have been prevented if any one of a variety of things
about the road, the vehicles, or the people involved had been di¬erent.31

When the causes of road accidents are carefully investigated, it may be found that
deterring drivers from negligent conduct is not the easiest or cheapest way to avoid
a certain type of accident. For example, the most e¬ective way of preventing people
driving into the car in front of them at night may be to require the rear end of vehi-
cles to be better lit. Clearly the deterrent value of tort law is thrown into doubt if
courts regularly ignore important causes of injuries. It is also the case that the less we
know about the causes of particular injuries, the less will we be able to use tort law
as a means of establishing standards of safety. This problem is particularly acute in
respect of diseases.32 Furthermore, the standard-setting function of tort law is

30 See generally D. Dewees, D. Du¬ and M. Trebilcock, Exploring the Domain of Accident Law: Taking
the Facts Seriously (New York, 1996).
31 Road Safety “ A Fresh Approach (Cmnd 3339, 1967), para. 7. See also J.T. Reason, ˜The Human
Factors in Medical Accidents™ in Vincent, Ennis and Audley, Medical Accidents, ch. 1; C. Vincent, M.
Ennis and R.J. Audley ˜Safety in Medicine™ in ibid., ch. 15. But see Pearson Report, vol. 2, table 42
summarizing ¬ndings of a Road Research Laboratory study that found 65% of road accidents may
be the result of human error alone. See also C.J. Bruce ˜The Deterrent E¬ects of Automobile
Insurance and Tort Law: A Survey of the Empirical Literature™ (1984) 6 Law and Policy 67, 68“73.
A survey by a ¬rm of insurance brokers in 1989 showed that 80% of tra¬c accidents occur in ¬ne
weather, 72% in good visibility and 70% on dry roads: Financial Times, 5 September 1989.
32 J. Stapleton, Disease and the Compensation Debate (Oxford, 1986), ch. 4; D.N. Dewees and
R. Daniels, ˜Prevention and Compensation of Industrial Disease™ (1988) 8 International R. of Law
and Economics 51.
426 Chapter 17

unlikely to be of much importance in relation to the design of new products or drugs,
where manufacturers and designers are operating at the edge of known technology.33
Another aspect of this point is that it is easier to deter deliberate conduct by
the threat of a tort action for damages than it is to deter merely negligent conduct.34
The risk of tort liability is more likely, for instance, to deter a surgeon from operat-
ing on a patient without his or her consent than to deter the surgeon from making
a negligent mistake in the course of the operation. Again, the risk of liability is no
doubt a factor in deterring newspapers from publishing material that they realise
might be defamatory;35 and there is some US evidence that the risk of tort liability
may discourage vendors of alcohol from serving obviously intoxicated customers.36
Road and workplace accidents, with which the tort system is in practice chie¬‚y con-
cerned, are often caused by inadvertent failure of observation and perception, by
faulty judgment, by lack of basic skills and other factors which the threat of liabil-
ity is unlikely to deter.37 For example, the fear of liability will not induce a motorist
to stop at a stop sign he or she has not seen “ even if the driver ought to have seen
it. Furthermore, given that negligent drivers put themselves as well as others at risk
by their carelessness, it is perhaps unlikely that they will be deterred from careless-
ness by the prospect of having to pay a ¬ne or damages if they are not deterred by
fear for their own safety.38
It does not follow, however, that we can never hope to deter people from negligent
conduct by threats of liability. For instance, we may not be able to deter a drunk from
dangerous driving once behind the wheel, but we may be able to deter a person from
getting drunk before driving. We cannot induce a driver to stop at a stop sign he or
she has not seen, but we can try to bring pressure on drivers to look for stop signs
rather more carefully.39 Drivers can be discouraged from speeding,40 drinking before

33 In this context, disclosure to consumers of product-risk information and enforcement by regula-
tory bodies of minimum safety standards are likely to be important: P. Burrows,˜Products Liability
and the Control of Product Risk in the European Community™ (1994) 10 Oxford Review of
Economic Policy 68, 78“82.
34 For a general theoretical discussion of this issue see H. Latin, ˜Problem-Solving Behaviour and
Theories of Tort Liability™ (1985) 73 California LR 677. The empirical literature on the causation
of accidents distinguishes between (unintended) ˜errors™ and (deliberate) ˜violations™.
35 E. Barendt et al., Libel and the Media: The Chilling E¬ect (Oxford, 1997); U. Cheer, ˜Myths and
Realities About the Chilling E¬ect: The New Zealand Media™s Experience of Defamation Law™
(2005) 13 Torts LJ 259.
36 F.A. Sloan et al., ˜Liability, Risk Perceptions and Precautions at Bars™ (2000) 43 J. of Law and
Economics 473.
37 The idea that some people are accident prone plays a part here. See C. McManus and C. Vincent,
˜Selecting and Educating Safer Doctors™ in Vincent, Ennis and Audley, Medical Accidents, 80“5.
38 There is evidence that people do learn from accidents independently of whether their accident-
causing conduct was penalized: Transport and Road Research Laboratory (TRRL), Survey Report
750 (1982); A. Quimby and K. Watts, Human Factors and Driving Performance (TRRL, Laboratory
Report 1004, 1981).
39 By seeing accidents or being involved in them people may learn how they are caused and how to
avoid them. Simulated accidents and defensive driving courses are thought to be helpful: TRRL,
Survey Report 750 (1982).
40 On the strong correlation between speeding and accidents see D.J. Finch et al., ˜Speed, Speed
Limits and Accidents™, Transport Research Laboratory (TRL), Project Report 58 (1994). On the
The functions of compensation systems 427

driving and driving when overtired41 (all major causes of road accidents) and from
consciously taking other unreasonable risks.42 Pedestrians can be discouraged from
crossing against red lights and going out on the roads when drunk.43 Tort liability
may also have some deterrent e¬ect where the potential defendant is a business
concern or an institution, such as a health authority, which will weigh the relative
costs of paying damages and preventing injuries; and where the injuries are caused
by defective products or premises or by unsafe working conditions which can be
made safer by conscious design and which may be unsafe exactly because of a con-
scious decision to skimp on safety.44
Even so, there are serious limitations on the e¬ectiveness of the risk of tort lia-
bility as a means of preventing injuries. Trying to deter people from dangerous
conduct by bringing pressure to bear before they get themselves into the dangerous
situation is only likely to be e¬ective to the extent that people recognize dangerous
situations. For example, one of the reasons why so many people drive while drunk
is because, despite the introduction of drink-driving laws45 and extensive educa-
tional campaigns, they still do not fully appreciate the risks involved.46 Another
reason why the risk of tort liability is likely to be a relatively ine¬ective deterrent is
that it probably does not present itself as a very serious possibility to many people.
This may be partly a result of the same ignorance of the law we noted when dis-
cussing claims consciousness,47 and partly because of the prevalence of liability
insurance. People are more likely to be deterred from careless driving (for instance)
by fear for their own safety or by a signi¬cant risk of incurring a ¬ne or driving
disquali¬cation48 than by the risk of tort liability. Important, too, are social attitudes
to particular types of careless behaviour. For example, it has been found that
social disapproval of drunk driving plays an important part in reducing its
relationship between personality and speeding accidents see R. West, J. Elander and D. French,
˜Decision Making, Personality and Driving Style as Correlates of Individual Accident Risk™ (TRL
Contractor Report 309, 1992).
41 It has been estimated that tiredness is a contributory factor in about 10% of car accidents:
G. Maycock, ˜Tiredness as a Factor in Car and HG Accidents™, (TRL Report 169, 1995); and of
15“20% on ˜monotonous roads, especially motorways™: J.A. Horne and L.A. Reyner,˜Sleep Related
Vehicle Accidents™ (1995) 310 British Medical Journal 565. But hard evidence is lacking: J. Connor
et al., ˜The Role of Driver Sleepiness in Car Crashes: A Systematic Review of Epidemiological
Studies™ (2001) 33 Accident Analysis and Prevention 31.
42 See also Bruce, ˜The Deterrent E¬ect of Automobile Insurance and Tort Law™, 68“73, 78“81;
Dewees, Du¬ and Trebilcock, Exploring the Domain of Accident Law, 20“1.
43 A third of pedestrians killed on UK roads are under the in¬‚uence of alcohol.
44 Some businesses may be more susceptible to external incentives than others: H. Genn, ˜Business
Responses to Regulation of Health and Safety in England™ (1993) 15 Law and Policy 219.
45 On the e¬ect of such laws see J.R. Snortum, ˜Another Look at “The Scandinavian Myth” ™ (1984)
6 Law and Policy 5. One researcher found that the proportion of injured road users with blood
alcohol concentrations above the legal limit was highest in groups not subject to the constraint of
such laws: J.T. Everest, ˜Drinking and Driving™ in Papers on Vehicle Safety, Tra¬c Safety and Road
User Safety Research (TRRL, 1991).
46 Drink Driving: The E¬ects of Enforcement (Home O¬ce Research Study 121, London, 1991), 33.
About 20% of drivers and motorcycle riders killed in road accidents are over the legal alcohol
limit: TRRL, Research Report 266 (1990).
47 8.2.3.
48 Drink Driving: The E¬ects of Enforcement, 46.
428 Chapter 17

incidence;49 and it has been suggested that seat-belt laws are more e¬ective than
drunk-driving laws because they are ˜morally neutral™.50
One of the reasons why ¬nes (etc.) for driving o¬ences are more likely to achieve
deterrent e¬ects than tort liability is that the standards of conduct prescribed by the
criminal law (in the form of tra¬c and industrial safety rules and regulations) are
usually much more precise than those laid down by the law of tort. If we are to take
deterrence seriously, we must give people detailed guidance as to how to behave.51
For example, because people like to think they are careful drivers, general exhorta-
tions to drive with care or at a reasonable speed are less likely to deter carelessness
than speci¬c rules such as that a car must not cross a double unbroken centre line
or must not be driven at more than a stated speed.52 It also appears to be important
that rules should be clear and unambiguous:53 it is better to establish a clear rule
for right of way at roundabouts than to tell driver to take care at roundabouts.54
The common law of tort does not typically give detailed guidance but only requires
people to take reasonable care according to all the circumstances of the situation
they are in. What is reasonable care will only be decided by a court after an injury
has occurred, and courts usually refuse to give speci¬c guidance for the future on
the ground that the exact circumstances of each particular case are crucial in decid-
ing how the defendant ought reasonably to have behaved. The courts fear that if
they lay down detailed rules of conduct, such rules may turn out to be ˜under-
inclusive™ by failing to deal with a particular set of facts which the court did not con-
template but which it thinks gives rise to a good claim for compensation.55 The
disadvantage of this approach is that many decisions that a particular defendant
was negligent do not have much deterrent potential.
On the other hand, if a case deals with a frequently recurring situation, and if
the nature of the negligence is identi¬ed in some detail, even court decisions may
in¬‚uence future conduct.56 For example, in one case it was decided that a local
authority had been negligent in using untoughened glass only one-eighth of an
inch thick in school doors instead of toughened glass, because of the risk of

49 Ibid., 46“7.
50 J.S. Legge Jr, ˜Reforming Public Safety: An Evaluation of the 1983 British Seat Belt Law™ (1987) 9
Law and Policy 17, 33.
51 Also, this guidance needs to be well publicized through the media or educational programmes.
52 On the other hand, excessively detailed rules and regulations may be counter-productive in safety
terms: Robens Committee Report, paras. 28“9. The deterrent e¬ect even of speci¬c rules is
a¬ected by how rigorously they are enforced. For instance, over a 6-year period in Victoria, a cam-
paign of rigorous enforcement of speed limits reduced the proportion of drivers exceeding the
limit from 23% to 1.8%; and rigorous enforcement of drink-driving laws reduced drink-related
fatal accidents by 50%: The Times, 11 January 1999.
53 D.R. Harris, ˜Evaluating the Goals of Personal Injury Law: Some Empirical Evidence™ in P. Cane
and J. Stapleton eds., Essays for Patrick Atiyah (Oxford, 1991), 304.
54 M. Austin, Accident Black Spot (Harmondsworth, 1966), 68“9.
55 The courts are more concerned to maximize the e¬cacy of tort law as a compensation mechanism
than as a deterrence mechanism.
56 Data about settled cases is bound to be much less in¬‚uential because the nature of the negligence
will often not be clearly identi¬ed, and records of settled claims may not provide the required
information. Also, of course, the fact that a negligence claim is settled favourably to the claimant
The functions of compensation systems 429

accidents to pupils.57 It would not be surprising if this decision was treated as laying
down a rule that glass in all school doors should be toughened, and if local author-
ities proceeded to act on the decision by replacing glass which did not conform to
the rule. Even relatively non-speci¬c rules can have some e¬ect. For example, it has
been suggested that replacing a rule that doctors need give patients only such infor-
mation about their treatment as doctors think reasonable with a rule that doctors
must, as a general rule, give patients all ˜material™ information and answer all their
questions, can have an e¬ect on doctors™ behaviour, even though no two patients
and no two situations are likely to be identical in all respects.58 It is still probably
true to say, however, that other things being equal, the more speci¬c a rule of
conduct, the greater its deterrent potential is likely to be.
We might summarize the discussion so far by saying that there are reasons to
doubt the e¬ectiveness of tort law as a deterrent to negligent conduct and as an
e¬ective mechanism for reducing accidents and injuries. But there is also some
reason to think that fear of tort liability may sometimes actually be counter-
productive in terms of accident prevention. A suggestion sometimes made is that by
focusing attention on accidents which generate compensation claims, the tort system
diverts attention away from the majority of accidents which do not, and so discour-
ages the formation of systematic and thorough accident-prevention strategies.59 In
another direction, however, it is sometimes said that the fear of liability may make
potential defendants unwilling to investigate or discuss injury-causing incidents lest
evidence of tortious conduct should emerge. Even worse, they may tamper with
documentary or other evidence in order to protect themselves. In the workplace
context, the injured person™s fellow workers, too, may be reluctant to discuss what
happened for fear of prejudicing their colleague™s claim for damages. Moreover,
potential defendants may be unwilling to take remedial measures before any claim is
settled for fear that this may be interpreted as an admission that adequate precau-
tions were not taken before.60
It has often been suggested that the fear of tort liability can also be counter-
productive by encouraging excessive caution and unnecessary precautions on the
part of potential defendants; in other words, it may be said that tort law over-deters.
This suggestion has been made in a variety of contexts, but most commonly in

does not mean that the defendant™s conduct was tortious. Indeed, claims are typically settled
without admission of liability. For a discussion of the di¬culties of using data about medical
negligence claims as a basis for risk management programmes see R. Dingwall and P. Fenn in
R. Dingwall and P. Fenn eds., Quality and Regulation in Health Care (London, 1992), ch. 1.
57 Re¬ell v. Surrey CC [1964] 1 WLR 358.
58 For a medical practitioner™s view see W.A. Ollbourne, ˜The In¬‚uence of Rogers v. Whitaker on the
Practice of Cosmetic Plastic Surgery™ (1998) 5 J. of Law and Medicine 334.
59 E.g. M. Brazier, ˜NHS Indemnity: The Implications For Medical Litigation™ [1990] Professional
Negligence 88, 90; H. Genn and S. Lloyd-Bostock, ˜Medical Negligence “ Major New Research in
Progress™ [1990] J. of Medical Defence Union 42, 43. Concerning shortcomings of studies of liti-
gated medical mishaps as aids to accident prevention see C. Vincent, ˜The Study of Errors and
Accidents in Medicine™ in Vincent, Ennis and Audley, Medical Accidents, 21“3.
60 E.A. Webb, Industrial Injuries: A New Approach (London, 1974), 11.
430 Chapter 17

relation to medical treatment.61 For instance, it is widely said that in the USA, where
the number of medical malpractice actions is thought to be relatively high, doctors
are encouraged to practise ˜defensive medicine™, that is to prescribe or refuse to
provide treatment, procedures (such as Caesarian sections) or medical tests not
because this is in the patient™s best interest but merely to safeguard themselves from
possible legal liability. Other suggestions are that fear of being sued leads people to
leave or not to enter high-risk specialties (such as obstetrics) and may cause certain
medical services to be unavailable.
There are several problems with such suggestions. First, such empirical evidence
as there is does not really establish any clear connection between the incidence of
tort liability and the various e¬ects ascribed to fear of being sued.62 One explan-
ation may be (as we saw earlier: 8.3.3) that apparently only a small proportion of
events of medical negligence results in tort claims.63 Furthermore, evidence sug-
gests a signi¬cant lack of ¬t between success or failure in a tort claim for medical
negligence and presence or absence of negligence (respectively).64 Secondly, the
concepts of ˜over-deterrence™ and ˜defensive medicine™ are themselves problematic
because the amount of care which it is appropriate to take is, to some extent at least,
a value-judgment.65 Thirdly, in England, as a general rule, a doctor will be held to
have been negligent only if no body of reputable medical opinion (regardless of
how small the minority may be which holds the opinion) can be found to support

61 For a careful discussion see R. Dingwall, P. Fenn and L. Quam, Medical Negligence: a Review and
Bibliography (Oxford, 1991), 41“56. See also M.A. Jones and A.E. Morris, ˜Defensive Medicine:
Myths and Facts™ (1989) 5 Journal of Medical Defence Union 40; D. Tribe and G Korgaonkar, ˜The
Impact of Litigation on Patient Care: An Enquiry into Defensive Medical Practices™ [1991]
Professional Negligence 2; Dewees, Du¬ and Trebilcock, Exploring the Domain of Accident Law,
96“112; B. Dickens, ˜The E¬ects of Legal Liability on Physicians™ Services™ (1991) 41 U. of Toronto
LJ 168; Factors In¬‚uencing Clinical Decisions in General Practice (O¬ce of Health Economics,
London, 1991); M. Ennis, A. Clark and J.G. Grudzinskas,˜Change in obstetric practice in response
to fear of litigation in the British Isles™ (1991) 338 The Lancet 616.
62 Dewees, Du¬ and Trebilcock, Exploring the Domain of Accident Law, 112. In Britain, a survey by the
Royal College of Obstetricians and Gynaecologists reported in 1992 showed that although 85% of
British obstetricians are or have been involved in litigation, this is less of a deterrent to recruitment
than long working hours, resident conditions and the job prospects of junior hospital doctors: P.
Saunders, ˜Recruitment in Obstetrics and Gynaecology: RCOG Sets Initiatives™ (1992) 99 Brit. J. of
Obstetrics and Gynaecology 538; see also F.A. Sloan et al., ˜Tort Liability and Obstetricians™ Care
Levels™ (1997) 17 International Rev. of Law and Economics 245. It is sometimes suggested that fear
of liability discourages doctors from acting as Good Samaritans; but see K. Williams, ˜Doctors as
Good Samaritans: Some Empirical Evidence Concerning Emergency Medical Treatment in Britain™
(2003) 30 J. of Law and Society 258. This is not to say that there is no evidence of defensive medical
practices. See e.g. D. Kessler and M. McClellan, ˜Do Doctors Practice Defensive Medicine?™ [1996]
The Quarterly Journal of Economics 353; K. Clark, ˜Litigation: A Threat to Obstetric Practice?™
(2002) 9 J. of Law and Medicine 303. But there is little evidence that it contributes signi¬cantly to
health spending: G.F. Anderson et al., ˜Health Spending in the United States and the Rest of the
Industrialized World™ (2005) 24 Health A¬airs 903.
63 M.M. Mello and T.A. Brennan, ˜Deterrence of Medical Errors: Theory and Evidence for
Malpractice Reform™ (2002) 80 Texas LR 1595.
64 Ibid., 1618“20.
65 So some would argue that exposure to tort liability might have a bene¬cial e¬ect on medical prac-
tice: R. Bowles and P. Jones, ˜Medical Negligence and the Allocation of Health Resources™ [1988]
Professional Negligence 111.
The functions of compensation systems 431

what the doctor did as reasonable. This means that a doctor is unlikely to be held
negligent for doing what other doctors do or for failing to do what other doctors
do not do.66
There is no doubt that the prevalence of liability insurance greatly reduces the
deterrent potentiality of tort law. For example, a motorist who is not deterred from
doing something foolish by fear for his or her own safety or that of passengers in
the car, nor by fear of the criminal law, nor by fear of being disquali¬ed from
driving, is not likely to be deterred by the fear of being sued in a tort action in which
the damages will be paid by an insurance company. There is no real evidence that
fear of the loss of a no-claims bonus or of having to pay the excess under a policy
has any signi¬cant e¬ect on the incidence of accidents.67
It has sometimes been suggested that motorists should not be permitted to
protect themselves by unlimited liability insurance but that they should, for
instance, be required to pay a certain amount of any claim personally.68 This might
jeopardize the accident victim™s chances of receiving full compensation, and it
would be a high price to pay for the additional deterrent value of such a scheme.
A variant, which would overcome this problem, would allow or require the insurer
to recover the designated sum from the insured after it had been paid to the victim.
But the administrative cost of such a scheme would probably be unacceptably
high. If the threat of having to pay large sums of money is a valuable deterrent
against dangerous driving, a more satisfactory approach would be to introduce
much larger criminal ¬nes for driving o¬ences. This would not jeopardize acci-
dent victims™ compensation and it would have the desirable e¬ect of giving incen-
tives to be careful to o¬ending motorists who do not cause accidents as well as to
those who do.
In workplace-accident cases the problem is similar. Here also, liability insurance
is compulsory and this means that the immediate ¬nancial consequences of an
accident are not felt by the employer.69 It is true that employers shoulder consider-
able indirect or incidental costs arising from accidents, such as disturbance of pro-
duction, consumption of management time and e¬ort in dealing with the accident

66 There is US evidence that doctors greatly overestimate the risk of being sued: M. Ennis and
C. Vincent, ˜The E¬ects of Medical Accidents and Litigation on Doctors and Patients™ (1994) 16
Law and Policy 97, 100. If this is true, the better (partial) explanation of defensive medicine may
be unrealistic fear of incurring tort liability rather than the incidence of tort liability as such. The
way to cope with such fear is better education of doctors about the law. Assessing the impact of
(the fear of) tort liability is further complicated by the fact that a doctor who has a medical mishap
may face disciplinary proceedings or, in extreme cases where the patient dies, prosecution for
manslaughter.˜Seriously negligent treatment™ can amount to ˜serious professional misconduct™ for
which a doctor can be disciplined: McCandless v. General Medical Council [1996] 1 WLR 167.
There is some evidence that defensive practice may result from medical accidents even in the
absence of actual or threatened litigation: Ennis and Vincent, ibid., 101“3. On the impact of acci-
dents more generally see B.M. Hutter and S. Lloyd-Bostock, ˜The Power of Accidents™ (1995) 30
Brit. J. of Criminology 409.
67 But it may have an e¬ect on claims under ¬rst-party policies.
68 J.J. Leeming, Road Accidents: Prevent or Punish? (London, 1969), 210“11.
69 The e¬ectiveness of liability insurance in accident prevention is examined later (17.7.2).
432 Chapter 17

and claims arising from it and so on. Such incidental costs represent a signi¬cant
proportion of total accident costs. These are not normally covered by insurance,70
but they will be passed on to the extent that the employer is able to do this; so they
will not hit any individual su¬ciently hard to provide an e¬ective incentive to take
accident avoidance measures. Even in the industrial ¬eld, the role of tort law as a
deterrent will be limited so long as the primary aim of the law remains the com-
pensation of the injured. Avoiding injury to oneself
It seems unlikely that tort law provides people with signi¬cant incentives to take
care for their own safety. The law attempts to do this through the doctrine of con-
tributory negligence; but (as was argued above in the context of causing injury to
others) if a person™s instinct for self-preservation does not deter them from dan-
gerous conduct, it is unlikely that a denial of monetary or other assistance will do
so. If people are not deterred from smoking by knowledge of the health risks, it is
unlikely that they would be deterred by a refusal of treatment under the NHS for
disease caused by smoking, or by the possibility of being met by a defence of con-
tributory negligence or volenti non ¬t injuria in a tort action against the cigarette
manufacturer. Similarly, it seems unlikely that a person with an irrational fear of,
or a religious objection to, receiving certain medical treatment which is medically
advisable following an accident, will be induced to have it because, as a result of the
doctrine of mitigation of damage, they may be deprived of damages for refusing
treatment. This is not to say that a serious risk of incurring a ¬ne or other criminal
penalty might not have some deterrent e¬ect, or to deny the importance of social
pressures not to engage in dangerous conduct; but it does cast doubt on the deter-
rent impact of tort law.
The case of failure to wear a car seat-belt illustrates this. The common law treats
such failure as contributory negligence, and a claimant™s damages may be reduced
on this ground. However, research has shown that during the period 1973“80 only
about 30% of drivers and front-seat passengers wore seat-belts despite the risk of
receiving reduced damages if injured while not wearing a seat-belt. In 1983 when,
subject to limited exceptions, it became a criminal o¬ence for a driver or front seat
passenger not to wear a seat-belt if provided, the compliance rate rose to 95% for
cars and over 80% for vans.71 This dramatic change is perhaps partly attributable
to the greater deterrent e¬ectiveness of the criminal over the civil law; partly to the
fact that the new provision was much better publicized and known than the
common law rule; and partly because of the con¬rmed e¬cacy of seat-belts in
reducing injuries “ it is estimated that there are now some 20“25% fewer fatal and
serious injuries to drivers and front-seat passengers than there would have been

70 It has been estimated that the ratio of insured to uninsured losses resulting from work accidents
is between 1:8 and 1:36 depending on the type of industrial operation in question: The Costs of
Accidents at Work (HMSO, 1993).
71 Road Accidents Great Britain 1983 (HMSO, 1984), table 11.
The functions of compensation systems 433

without compulsory belt-wearing.72 Social attitudes may also be important. Like
smoking in public, failure to wear a seat-belt is widely frowned upon. By contrast,
attempts to reduce speed-related accident rates may be hampered by ˜the ethos of
speed™ that ˜permeates discourse on driving™ and ˜has in¬ltrated all aspects of
modern life™.73

17.7.2 Accident prevention via insurance
Because insurance (whether third party or ¬rst party) weakens incentives to
prevent the loss insured against, insurers use various techniques to overcome what
they call the ˜moral hazard™ generated by insurance.74 There are two main types of
technique. One is to encourage or require insured parties to take loss prevention
measures; and the other is the use of risk-related premiums, excess payments (or
˜deductibles™), ceilings on cover, and so on. Encouraging or requiring loss prevention measures
Because insurance involves the pooling of a large number of risks, the insurer may
¬nd it worthwhile to take steps to reduce or eliminate risks that no individual risk-
bearer would take. For example, ¬re brigades in England were originally established
and maintained by insurers; not until 1865 was responsibility for ¬re-¬ghting
transferred to local authorities. No individual householder would ¬nd it worth-
while to maintain adequate ¬re-¬ghting facilities, because the value of any partic-
ular house multiplied by the risk of its being burned down is far less than the cost
of such facilities. But if an insurer sells ¬re policies to 100,000 house-owners, the
risk the insurer bears may make it pro¬table for it to maintain a ¬re brigade.
Similarly, insurers employ inspectors to survey plant, equipment and premises and
to advise insured parties how to minimize risks and avoid losses. In the USA, for
. . . regulation of the design safety of aircraft was initiated by insurers . . . Underwriters
Laboratories, at the request of the National Aircraft Underwriters™ Association, formed
an aviation department in 1920 to certify the airworthiness of aircraft and develop

72 Ibid., para. 3.10. See also Legge,˜Reforming Public Safety™, 31“2. For research showing the overall
pattern of reductions and increases in particular types of injuries following the introduction of
compulsory seat-belts see TRRL, Research Report 239 (1989). On the other hand, it has been
suggested that seat-belts (and other aspects of car design provided as ˜safety features™) give
drivers a false sense of security which may lead them to take risks with their own and other
people™s safety which they might not otherwise have taken. For evidence that improvements in
road design may have similar e¬ects see R.B. Noland, ˜Tra¬c Fatalities and Injuries: The E¬ect
of Changes in Infrastructure and Other Trends™ (2003) 35 Accident Analysis and Prevention 599.
But for a deeply sceptical view of the validity of such research see R. Elvik, ˜To What Extent Can
Theory Account for the Findings of Road Safety Evaluation Studies?™ (2004) 36 Accident Analysis
and Prevention 841.
73 Department for Transport, ˜The E¬ects of Speed Cameras: How Drivers Respond (No 11)™, 8.
74 C. Parsons, ˜Moral Hazard in Liability Insurance™ (2003) 28 Geneva Papers on Insurance and
Risk 448.
75 J. Braithwaite and P. Drahos, Global Business Regulation (Cambridge, 2000), 454. This may be seen
as an example of the e¬cacy of liability-cum-liability-insurance as a ˜regulatory mechanism™; in
434 Chapter 17

airworthiness standards . . . a function ultimately taken over by the Federal Aviation

In this country insurers have traditionally played their largest loss minimization
role in connection with engineering and ¬re insurance.76 For example, they main-
tain (in conjunction with ¬re authorities) a joint ¬re research organization which
investigates ¬re causation and prevention. No insurer would take on any signi¬cant
¬re risk without surveying the premises. Employers™ liability insurers may also
survey premises and operations, and persuade or require the employer to take
various precautions before the risk is accepted.77 On the other hand, Stapleton
argues, in relation to the prevention of diseases and health hazards (as opposed to
traumatic accidents) that, ˜loss minimization by insurers via the direct action of
inspecting and advising policy holders is virtually impossible because of the much
greater scienti¬c expertise required in this area™.78
Potentially even more e¬cient than the e¬orts of insurers to prevent losses are
community-wide accident-prevention techniques. Fire brigades maintained by
insurers for the bene¬t of their policyholders are better than no ¬re brigades at all;
but ¬re brigades maintained by the State for the bene¬t of all are better still.
Similarly in the area of industrial safety, it is, in principle at least, more e¬cient for
the Health and Safety Commission and the various government safety inspec-
torates to monitor and promote safety in industry generally by inspection, research,
education and the enforcement of safety regulations,79 than for the various
employer™s liability insurers to attempt to do so. In any event, when the insurance
market is highly competitive, insurers have little incentive to spend money on loss-
prevention: it has to be paid for out of premium income, and an insurer who
chooses to spend little or nothing on loss prevention may be able to undercut insur-
ers who spend signi¬cant amounts trying to reduce losses. Unless insurers under-
take this loss-prevention activity jointly (as in the ¬re insurance industry) not
much can be achieved; but joint activity, in turn, may attract accusations of illegal

other words, as a means of in¬‚uencing human behaviour in order to promote socially desirable
objectives, such as safety. By contrast, it has been argued that in relation to nuclear power gener-
ation, tort law was prevented from performing a regulatory function because governments pro-
tected the nuclear power industry from full tort liability, which would have been uninsurable:
Braithwaite and Drahos, ibid., 308. On the relationship between tort law and regulation see
P. Cane, ˜Tort Law as Regulation™ (2002) 31 Common Law World Review 305; J. Stapleton,
˜Regulating Torts™ in C. Parker et al. eds., Regulating Law (Oxford, 2004).
76 Sansom (1965) 62 Journal of the Chartered Insurance Institute 97.
77 See the evidence of the Association of British Insurers to the Robens Committee, Robens
Committee Report, vol. 2, 43¬. One of the initiatives taken to address the recent ˜crisis™ in employ-
ers™ liability insurance was to encourage insurers to o¬er lower premiums to employers who could
demonstrate that they had devoted additional resources to safety. But as of mid-2005, employers™
organizations were complaining of lack of progress in this regard: e.g. R. Tyler, ˜Liability pledge a
waste of time™, Telegraph, 28 July 2005.
78 Stapleton, Disease and the Compensation Debate, 128.
79 Public regulatory techniques are beyond the scope of this book. For a discussion see R. Baldwin,
˜Health and Safety at Work: Consensus and Self-Regulation™ in R. Baldwin and C. McCrudden
eds., Regulation and Public Law (London, 1987), 132“58.
The functions of compensation systems 435

restrictive practices and price-¬xing.80 Even in the USA, where the insurance indus-
try has played an important role in promoting safety in industry, it has not done
much, if anything, to further accident prevention on the roads.81 One notable illus-
tration is the failure of the insurance industry to force motor manufacturers to
install seat-belts in cars once it became clear that they could play a major role in
reducing the gravity of injuries: action in this ¬eld had to be taken (both in the USA
and in this country) by legislation.82 Risk-related premiums and similar techniques
Excess (or ˜deductible™) provisions, under which the insured is obliged to pay the
¬rst £X of any claim, are designed both to encourage the insured to take measures
to minimize the possibility of insured events occurring and to discourage small
claims, which are disproportionately expensive to process. Deductibles are a
common feature of ¬rst-party insurance; and compulsory motor insurance poli-
cies in respect of third-party property damage were, until recently,83 not required
to cover the ¬rst £300 of loss.84 Exclusions from cover can also serve a deterrent
function. For example, some insurers will not meet ¬rst-party road accident claims
(or will seek to recover from the insured amounts paid out in respect of third-party
liability claims) which result from the insured driving with a blood alcohol content
above the legal limit.
Insurance companies can, in principle, set premiums in such a way as to encour-
age the insured to take loss-prevention measures and to reduce the severity of losses
that do occur. For example, ¬re insurers may set di¬erent premiums for the insur-
ance of buildings that have adequate sprinkler installations and those which do not.85
In some contexts, di¬erential premiums may prove unacceptable. For example, pre-
miums for professional liability insurance under non-pro¬t (˜mutual™) schemes run
by the professions themselves often subsidize younger practitioners at the expense of
the more experienced, even though the former might be expected to present a greater
risk. Again, the threat of the introduction of medical negligence liability premiums
related to the riskiness of di¬erent medical specialties, in place of ¬‚at-rate premiums,
was a major catalyst for the introduction of Crown indemnity.86 In the case of liabil-
ity insurance in respect of road accidents and workplace injuries, it has proved
di¬cult to devise premium structures that have a signi¬cant e¬ect on accident costs
and rates.

80 See e.g. Report of the Monopolies Commission on Fire Insurance (HMSO, 1972).
81 R.E. Keeton and J. O™Connell, After Cars Crash (1968), 95.
82 Insurers may be unduly sceptical of their ability to in¬‚uence vehicle design: Report of the British
Columbia Royal Commission on Automobile Insurance (1968), citing Swedish experience.
83 See 9.9 n. 107.
84 Ceilings on cover are also a common feature of liability insurance, including compulsory employ-
ers™ liability insurance (£5 million per incident) and compulsory motor insurance in respect of
third-party property damage (‚¬1 million). Ceilings are not designed to deter insured events but
to limit the insurer™s exposure and so keep premiums down.
85 Report of the Monopolies Commission on Fire Insurance (HMSO, 1972), para. 106.
86 16.1.
436 Chapter 17

So far as concerns road accidents, the general picture is well enough known.
Liability insurance premiums depend on a variety of factors which have, in the past,
been shown statistically to be associated with high claims experience; such factors
include age,87 sex,88 vehicle type and level of usage. ˜Feature-rating™ on the basis of
such factors does not take account of the claims record of the individual insured
but is based on the aggregate claims record of the group(s) to which the particular
insured belongs. On the other hand, the typical no-claims bonus system takes
account of each particular insured™s claims record (and so it may be called a system
of ˜experience rating™). A person with a history of accidents (or, more accurately, of
accidents which give rise to claims) will pay substantially more for liability insur-
ance than a person who has made few or no claims in recent years, regardless of
which risk-category each belongs to. Drivers with a certain number of points on
their licence may also be charged higher premiums.
Despite appearances, however, there is reason to doubt that the premium struc-
ture of motor liability insurance is based on sound scienti¬c or statistical princi-
ples. For example, probably only some 3 to 4% of drivers are involved in a personal
injury road accident in any one year, whereas six or seven times this number may
be involved in accidents resulting in property damage alone. As a result, the claims
record of most drivers is based largely on property damage accidents only; but this
claims record also a¬ects the size of that part of the premium which covers third-
party personal injury risks. Unfortunately, however, there is no evidence of a
signi¬cant correlation between involvement in property-damage accidents and
involvement in personal injury accidents.89 Moreover, except in relation to youth
and inexperience, there is no ¬rm evidence showing any signi¬cant correlation
between past accident experience and likely future involvement in accidents, all
other things being equal.90 Nor is it easy to believe that fear of loss of a no-claims
bonus plays a signi¬cant role in making drivers drive more carefully,91 although

87 For evidence that the risk of being involved in an accident is greatest amongst young drivers and
motorcycle riders see e.g. TRRL, Research Report 27 (1986); Research Report 135 (1988);
Contractor Report 146 (1990) and Research Report 315 (1991). There is a strong correlation
between lack of driving experience and accident involvement amongst young drivers: E. Forsyth,
G. Maycock and B. Sexton, ˜Cohort Study of Learner and Novice Drivers: Part 3, Accidents,
O¬ences and Driving Experience in the First Three Years of Driving™, TRL Project Report 111
(1995); D.D. Clarke, P. Ward and W. Truman, ˜In-Depth Accident Causation Study of Young
Drivers™, TRL 542 (2002).
88 For a discussion of the ethical basis of such classi¬cations see W.A. Wiegers, ˜The Use of Age, Sex
and Marital Status as Rating Variables in Automobile Insurance™ (1989) 39 U. of Toronto LJ 149.
Women typically pay lower premiums than men, especially amongst young drivers. Young women
have many fewer accidents than young men partly because they are ˜safer™ drivers and partly
because they drive less.
89 Although there is some evidence that drivers who engage in risky driving behaviour “ and who,
for that reason, might more likely to be involved in property-damage accidents “ are more likely
to su¬er personal injury while driving: S. Blows et al., ˜Risky Driving Habits and Motor Vehicle
Driver Injury™ (2005) 37 Accident Analysis and Prevention 619.
90 Pearson Report, vol. 2, para. 202.
91 Sansom, (1965) 62 Journal of the Chartered Insurance Institute 97, 107“8. But see Bruce, ˜The
Deterrent E¬ects of Automobile Insurance and Tort Law™, 84“7.
The functions of compensation systems 437

it may well reduce willingness to claim (which is, indeed, a major purpose of the
no-claims bonus system).
With regard to industrial injuries, employers™ National Insurance contribu-
tions are not risk-related. By contrast, liability insurance covering employers™ tort
liability does operate according to market principles; accordingly, employers who
present higher risks may pay higher premiums.92 However, the great majority of
employers pay premium rates based on certain features of the insured™s business
and not by reference to the individual insured™s own claims experience. The main
reason for this is that an employer™s own pattern of claims is not a statistically reli-
able indicator of likely future claims unless the employer has quite a large work-
force. It is impractical to take account of a ¬rm™s own experience unless it employs
at least one hundred workers, and a ¬rm would not be fully rated on its own
experience unless it employed a great many more than this “ perhaps 500 employ-
ees or more. This rules out the great majority of employers, because 99% of all
¬rms93 in manufacturing industry have a workforce of fewer than 500. However,
the other 1% of ¬rms in manufacturing industry together employ about 50% of
the total manufacturing workforce, and for this group, experience rating may be
There are other problems with experience-rating which reduce its value for loss-
prevention purposes.94 First, it is not simple to de¬ne what is meant by the
˜experience™ of a ¬rm. Does this mean the number or the cost of the accidents and
injuries that have occurred? Should the experience take account of costs which are
not paid by the employer™s insurer; and if so, how are these to be estimated? What
is to be done about one very large claim that could distort the employer™s claims
experience for years? And so on. Secondly, there is a time-lag problem. Experience-
rating is always based on conditions which are to a greater or lesser extent out of
date. For example, a ¬rm™s premium in 2006 must be ¬xed some time in 2005, and
must therefore be based on the experience of 2004 and earlier years. One year™s
experience is unreliable statistically, so a ˜moving average™ of 3 years™ experience is
usually used, but this means that the premium for 2006 will depend in part on the
experience of 2002. During this period, many relevant factors may have altered
the reliability of that previous experience as a guide to likely future experience.
Thirdly, the total cost of an employer™s liability insurance represents only a very


. 15
( 18)