. 14
( 18)


receive more than is necessary to achieve the purpose that either of the payments
is designed to serve. For example, if both payments are designed to replace lost
income, a person who receives a payment from both sources may receive more in
compensation than has been lost in earnings. Such over-compensation, as we might
call it, appears, prima facie, to be a wasteful use of resources. Although there are
some circumstances “ discussed below “ in which over-compensation may appear
unobjectionable, it seems a reasonable starting point to assume that money spent
on over-compensating people could be better spent on other things. This general
approach was endorsed by the Pearson Commission.3
Compensation systems usually try to avoid over-compensation within their
own con¬nes. For example, the tort system will not give damages for the same
injury more than once. Thus, if a person is injured in an accident caused by the
negligence of two or more tortfeasors, that person may recover any damages
awarded from the various tortfeasors in any proportions possible, but is not
allowed to recover in total more than the amount of the award. Again, where a tort
claimant is awarded damages both for lost earnings and for the cost of care in a

1 Even more complicated issues might arise where more than two sources were available.
2 E.g. R. Lewis,˜Deducting Collateral Bene¬ts from Damages: Principle and Policy™ (1998) 18 LS 15.
3 Pearson Report, vol. 1, ch. 13.

378 Chapter 15

residential institution, the former will be reduced to take account of savings in
living expenses as a result of living in the institution. As we have seen,4 the social
security system also sets its face against overcompensation within its own con¬nes.
A person can claim both incapacity bene¬t and disablement bene¬t, but this is
because the former is designed to replace lost income while the latter is compen-
sation for the disability as such and is not intended to be for lost income.
First-party insurance which is designed to replace something with measurable
¬nancial value “ ˜indemnity insurance™ “ only provides cover for the loss actually
su¬ered, whatever that may be. For instance, a person may be covered by two sep-
arate policies, each covering the full loss, but the insured may only recover once.
Prima facie each insurer will have to bear half the loss.5 However, the position is
di¬erent in the case of insurance against losses which cannot be measured in
money terms. For instance, a person who buys more than one life insurance policy,
or more than one policy of personal accident insurance against disability alone, can
recover under both because it is not really possible to say, in respect of a loss which
cannot be measured in money terms, that receiving £2X rather than £X is over-
compensation. By contrast, tort law does not allow a claimant to recover full
damages for non-pecuniary loss from several tortfeasors merely because the loss is
non-pecuniary. This is partly because the courts pretend that there is a ˜proper™ sum
to award as damages for disabilities. Perhaps a better reason for distinguishing
between ¬rst-party insurance and tort liability in this respect is that tort damages
are not paid for by the claimant, whereas ¬rst-party insurance is paid for by the
insured: if a person chooses to buy more than one policy to cover the same loss,
there is no good reason why the insured should not be allowed to keep all the insur-
ance bene¬ts paid for.
Our main concern in this chapter is with the way in which di¬erent compensa-
tion systems ¬t together, rather than the way in which each system treats the
problem of over-compensation within its own con¬nes.

15.2 The choice of compensation system
Because the compensation systems we have discussed have developed in a haphaz-
ard way, there is no coherent set of rules or principles governing the relationships
between them. Many social security bene¬ts are payable regardless of whether the
claimant has access to other sources of compensation. For example, statutory sick
pay is payable even if the employee is entitled to occupational sick pay, and it is for
the employer to decide whether to reduce the latter to take account of the former.
Again, social security bene¬ts are payable to recipients of criminal injuries
compensation or tort damages; it is the criminal injuries compensation or tort
damages which are reduced to take account of social security payments and not vice

5 Gale v. Motor Union Insurance Co. Ltd [1928] 1 KB 359.
A plethora of systems 379

versa.6 On the other hand, in assessing entitlement to means-tested social security
bene¬ts, account is taken of criminal injuries and tort compensation, for instance.
It is probably the case that most employers deduct statutory sick pay from occu-
pational sick pay. It is, however, less usual to take account of tort damages when sick
pay is granted. This is partly because it will usually be months or years before any
tort damages are received. Also, a claim for tort damages is probably not a su¬-
ciently common event to make it worthwhile to insert special provisions dealing
with it in sick-pay schemes. But it is apparently permissible to make provision for
the employer to ˜advance™ wages as a ˜loan™ to a person injured in an accident giving
rise to a tort claim, the injured person having an obligation to repay the employer
if and when the employee recovers damages for lost earnings.7 This is a device to
place the cost of wages for an injured employee onto the tortfeasor (and hence lia-
bility insurers) rather than the employer. The main issue here is whether employ-
ers or motorists should pay for lost wages arising from road accidents. In general
the cost is borne by motorists to the extent that fault can be proved, and it might
therefore be thought that there was no reason why an employer who organizes a
sick-pay scheme should not be entitled to insist that wages lost through negligent
driving should be borne by motorists rather than by the employer. But it is doubt-
ful if this is sound policy because the advantage to the employer must be small: it
can only be in a small proportion of cases that sick pay can be recouped in this way,
while the administrative cost of recoupment is likely to be high, because it involves
reliance on the tort system.
First-party insurers are free to try to shift their losses onto other compensation
systems by the terms of their insurance policies. For example, a personal accident
policy could stipulate that there is to be no liability if compensation is obtained
from other sources. In fact insurance companies do not generally attempt to do
this. They are mainly concerned with preventing the insured recovering for the
same loss under more than one policy. First-party insurers also have various sub-
rogation rights, which we will examine shortly.
The lack of clear rules of priority between compensation systems may create
di¬culties where one system refuses to meet a claim on the basis that another will
do so. For example, in one case8 the MIB sought to argue that a claimant, injured
by a defendant who deliberately used his car (which was uninsured) to run him
down, ought to have sought compensation from the CICB rather than from the
MIB, even though the latter was, as the House of Lords held, liable under its agree-
ment with the government in cases of deliberate running down. In this type of case,
the CICB does not normally award compensation precisely because the MIB agree-
ment covers it. The House of Lords, without adverting to this fact, held that the
CICS and the MIB scheme were not mutually exclusive, and that the claimant could

6 As we will see later, in some cases the value of social security bene¬ts paid to a recipient of tort
damages is recoverable by the Compensation Recovery Unit from the tortfeasor.
7 IRC v. Hambrook [1956] 2 QB 641, 656“7.
8 Gardner v. Moore [1984] AC 548.
380 Chapter 15

choose to proceed under whichever scheme was more favourable to him. Thus
stalemate was avoided, but the basic issue of which compensation scheme ought to
have borne the loss was not tackled. Like the CICA, the MIB sets o¬ sums received
from other sources against the claimant™s compensation.
Less satisfactory is a case in which the claimant™s leg was rendered sti¬ as a result
of a car accident due to the fault of D; he was later shot in the same leg by X and had
to have it amputated.9 The House of Lords held that D was liable for all the claimant™s
loss except that caused by the amputation of the leg. In a later case,10 doubt was cast
on this result and it was suggested that D ought to have been liable only for loss
su¬ered up until the time the claimant was shot, partly on the ground that C could
have made a claim to the CICB in respect of the shooting. This criticism of the earlier
decision ignores the fact that, in general, the CICB will only give compensation for
losses that are compensatable according to common law principles.
The mere fact of having di¬erent systems generates the need to spend time
deciding which is the appropriate system to bear the loss. For example, entitlement
to damages requires proof of negligence; entitlement to industrial injury bene¬ts
requires proof that the accident arose ˜out of and in the course of employment™;
entitlement to criminal injuries compensation requires (with a few exceptions)
proof that a crime of violence has been committed. Each of these (and other)
criteria of entitlement tend to produce troublesome borderline cases; and the allo-
cation of such cases to the ˜appropriate™ compensation system absorbs a dispro-
portionate amount of administrative e¬ort and cost. The more systems there are,
the more borderline cases there will be; and the more demarcation disputes will
have to be solved. Such disputes appear even more futile when it is remembered
that the burden of most compensation systems is ultimately borne by (large sec-
tions of) the public. One of the undoubted attractions of an integrated approach
to the question of compensation is that it would eliminate many of these demar-
cation disputes.

15.3 Subrogation and recoupment
Where two persons are legally liable to compensate a third party for some loss, and
as between these two, one is under a ˜primary™ liability and the other is only ˜secon-
darily™ liable, the latter is normally entitled, on paying the compensation, to be ˜sub-
rogated™ to the rights of the third party against the person primarily liable. In the
tort system, tort liability is still treated as primary liability while insurance is still
treated as an ancillary or secondary feature of the system. Therefore, a tortfeasor
can sometimes be sued by an insurer who has indemni¬ed the victim against the
loss. For example, if a person insures a house against ¬re and the house is burned
down as a result of the negligence of a tortfeasor, the owner of the house will prob-

9 Baker v. Willoughby [1970] AC 467.
10 Jobling v. Associated Dairies Ltd [1982] AC 794.
A plethora of systems 381

ably claim from the insurer who is then subrogated to the owner™s claim against the
tortfeasor. In practice, it would be worthwhile for an insurer to make a claim in this
sort of situation only if the tortfeasor was insured against the liability, as in the
familiar case of two motorists, both comprehensively insured, who collide as a
result of their combined negligence. Here each motorist can claim against their own
insurer for the damage to the vehicle; each insurer is then subrogated to the claims
of its insured against the other driver and is entitled to pass on to the other insurer
the cost of the claim by its own insured.11
Another instance of subrogation is to be found in the case of an employee whose
negligent conduct has caused the employer to be held vicariously liable to a third
party; if the employer was insured, the ˜primary™ liability is treated as resting on the
employee and the insurer is subrogated to the employer™s claim against the servant.12
The government has statutory rights analogous to subrogation rights under
the Social Security (Recovery of Bene¬ts) Act 1997, which is discussed further later
in the chapter (15.4.5). Under this Act, a person paying compensation to another
as a consequence of injuries or disease su¬ered by that other person for which the
payer is, or is alleged to be, liable13 is required to deduct from the compensation
paid to the injured person the amount of speci¬ed social security bene¬ts received
by that person in respect of the injuries resulting from the tort in the period of
5 years from the date of the injuries, or in the period from the date of the injury
to the date of the compensation payment, if this is less than 5 years. The sum thus
deducted is to be paid to the DWP. These provisions do not apply to payments made
by the CICA under the CICS; to payments made under criminal compensation
orders; to damages or settlements under the Fatal Accidents Act 1976;14 or to
government-funded compensation payments made under the Vaccine Damage
Payments Act 1979, or to haemophiliacs su¬ering from AIDS, or under the 1974
scheme for compensating miners su¬ering from pneumoconiosis,15 or to occupa-
tional sick pay. But they do apply to payments by the MIB. The MIB does not have
subrogation rights because it is not under a legal obligation to claimants to make
the payments it does.16

11 If a person is covered by two liability policies issued by di¬erent insurers and covering the same
liability, neither insurer has subrogation rights against the other, but either or both can claim con-
tribution: Austin v. Zurich Insurance Co. [1945] KB 250.
12 Lister v. Romford Ice and Cold Storage Co. Ltd [1957] AC 555; but see Morris v. Ford Motor Co.
[1973] 1 QB 792 and
13 The Act does not apply to payments under ¬rst-party insurance policies.
14 The Law Commission is of the view that the rules about collateral bene¬ts in fatal accident cases
should be consistent with those in personal injury cases: Law Com. No. 263, Claims for Wrongful
Death (1999), para. 5.39.
15 See 4.10 and 13.6.
16 This also used to be true of the CICB; but now payments under the CICS have a statutory basis
in the Criminal Injuries Compensation Act 1995. In relation to the old scheme, the CICB once
expressed the view that there were very few cases in which it could usefully have sued the crim-
inal even if it had had the power to do so: CICB Eighth Report (Cmnd 5127, 1972), para. 18. But
in 2004 the 1995 Act was amended to give the CICA power to recover directly from criminals com-
pensation paid to victims under the CICS (12.2).
382 Chapter 15

The NHS has statutory rights analogous to subrogation rights under the Road
Tra¬c (NHS Charges) Act 1999. Under the Act, where a person, who has died or
been injured as the result of a road accident and has received treatment in an NHS
hospital, receives compensation, the person who pays the compensation is liable to
pay to the government an amount in respect of that treatment, at a ¬‚at rate (in
2005) of £483 for treatment without admission and at a daily rate (in 2005) of £593
for treatment with admission, up to a maximum in any one case of £35,500. Under
Part 3 of the Health and Social Care (Community Health and Standards) Act 2003
this recoupment scheme has been extended to all cases where a person legally liable
or alleged to be legally liable pays compensation for personal injury, regardless of
cause; and to ambulance services as well as hospital treatment. This extension will
have its greatest impact in relation to work accidents which, after road accidents,
are the main cause of tort claims. Because of recent concern about the state of the
employers™ liability insurance market (7.7), the new scheme will not be introduced
until October 2006. When it comes into e¬ect, it will supersede the 1999 scheme.
Certain compensation payments, including payments made pursuant to a criminal
compensation order or as a result of a fatal accident claim, are not subject to the
2003 recoupment scheme.17 A noteworthy limit on the scheme arises from the fact
that ˜personal injury™ does not include any disease. The explanation given by the
Department of Health for this limitation is that ˜it would be di¬cult to quantify
the cost of diseases . . . to the NHS due to the complexity of the treatment path and
the period of time over which a person with a . . . disease would need treatment™.18
From time to time suggestions have been made for a wider use of subrogation
in the compensation process. For example, the Law Reform Committee once pro-
posed that an employer who pays wages to a person injured as a result of a tort
should be subrogated to the injured person™s rights against the tortfeasor.19 Again,
when the CICS was under consideration, it was suggested that the State should have
subrogation rights against the criminal;20 but it was not until 2004 that such a pro-
posal was adopted (12.2). The Winn Committee on personal injury litigation put
forward for consideration a scheme under which the State would pay some, or all,
of the damages which a road accident victim might be able to recover in subsequent
litigation, and would then recoup itself by taking proceedings against the tortfea-
sor.21 And the Society of Labour Lawyers (in evidence to the Pearson Commission)
once advocated a similar scheme for industrial injuries.

17 See Sch. 10 to the 2003 Act. Also not included in the scheme are payments made under s. 158 of
the Road Tra¬c Act 1988 by users of motor vehicles to medical practitioners who give emergency
18 The Recovery of National Health Service Costs in Cases Involving Personal Injury Compensation:
A Consultation (2002), para. 4.3.
19 Eleventh Report (Loss of Services, etc.) (Cmnd 2017, 1963), para. 5. The Law Commission rejected
the idea (Law Com. No. 56, paras. 146“50) as did the Pearson Commission (Pearson Report, vol. 1,
para. 446). The employer™s own right to sue for loss of an employee™s services was abolished by s.
2 of the Administration of Justice Act 1982.
20 JUSTICE, Report on Compensation for Victims of Crimes of Violence (London, 1962), 20.
21 Winn Committee Report, paras. 107“10 and Appendix 7.
A plethora of systems 383

The main argument in favour of subrogation rights is to ensure that the cost of
compensation ultimately rests on the party legally responsible for the harm which
the compensation redresses.22 However, various arguments can be made against
subrogation. Two types of subrogation claim require consideration: ¬rst, a claim by
the State against a tortfeasor with private liability insurance, or by a private (loss)
insurer against a tortfeasor with private liability insurance; and, secondly, a claim
by the State or an insurer against an individual tortfeasor who has no liability insur-
ance. In the ¬rst type of case, the main objection to subrogation is that it ought not
to be allowed if the party claiming subrogation is the most appropriate party to bear
the loss; and if this is not the case, the law should perhaps be changed so that the
cost is placed on the most appropriate party in the ¬rst instance. However, this
objection may be too simplistic. For example, even if we accept that tortfeasors
ought, prima facie, to bear the costs of their torts, we may still approve of the pro-
visions of the Social Security (Recovery of Bene¬ts) Act 1997 discussed above on
the ground that tort victims may need ¬nancial assistance in the short term, and in
reality the social security system is much more likely to provide assistance quickly
than the tort system. Delays inherent in the tort system may justify the scheme con-
tained in the 1997 Act. On the other hand, we might say that what the 1997 Act
really shows is the need to replace the whole tort system with a more e¬cient per-
sonal injury compensation system, which could deliver compensation as quickly as
the social security system.23
Where subrogation rights are asserted against an individual who is neither
insured nor an e¬ective self-insurer, the principal objection is the same as the
objection against making individuals pay tort damages, namely that it is not prac-
ticable nor, in the majority of cases, is it just or in the public interest. If we accept
that one of the principal aims of any compensation system should be to spread
losses widely so that the cost does not fall too heavily on any individual, it follows
that a loss which has fallen on a party able to spread it (such as an insurer) ought
not to be shifted again to a party unable to spread it e¬ectively. In the case of the
CICA, an argument used against giving it subrogation rights against a criminal who
had already been or was likely to be successfully prosecuted was that in pursuing
the criminal, the CICA might appear to be seeking a second punishment, even
though this would not technically be the case.24 By 2004, a desire to limit the cost
of the CICS was felt to justify giving the CICA power to recover directly from con-
victed criminals amounts paid out under the Scheme.

22 A similar case could be made against rights of contribution. For a fuller discussion in a di¬erent
context see P. Cane, Tort Law and Economic Interests, 2nd edn (Oxford, 1996), 435“43.
23 ˜The recovery of NHS costs is but a symptom of a New Labour desire to make the public services
more e¬cient in terms that a Thatcherite would approve, but which would dishearten many wel-
farists™: R. Lewis, ˜Recovery of NHS Accident Costs: Tort as a Vehicle for Raising Public Funds™
(1999) 62 MLR 903, 911.
24 JUSTICE, Report on Compensation for Victims of Crimes of Violence, 20; Hansard, 5th series, HC,
vol. 694, col. 1159.
384 Chapter 15

Another objection to subrogation is that shifting losses around is costly.25 The
initial allocation of liability is expensive26 and subrogation adds further to the cost
by shifting it again. Even if we thought that subrogation was in principle a good
idea in certain circumstances (e.g. under the Social Security (Recovery of Bene¬ts)
Act 1997), we might want to reconsider the matter if it turned out that the cost of
enforcing rights of subrogation was very high relative to the amounts recovered. It
was probably for this reason that when the scheme for recovery of social security
bene¬ts was ¬rst introduced in 1989, compensation payments of £2,500 or less
were excluded from its scope. The 1997 Act removed this exclusion, but allows regu-
lations to be made in the future to exclude ˜small payments™ from the recoupment
scheme. In fact, the administrative costs of the social security and NHS recoupment
schemes seem to be very modest. In 1997“8 £177 million was recovered under the
social security scheme at a cost of less than £4 million; and in 2002“3 just over £98
million was recovered under the NHS scheme at a cost of around £1 million, to
which must be added the (relatively small) additional costs to insurers of making
the payments.27
There is no doubt that the schemes for recoupment of social security payments
and NHS costs out of tort compensation payments are ¬nancially justi¬able in the
sense that they generate a substantial net bene¬t for the public purse. Moreover,
they do this at the expense of the activities which, in legal terms, caused the harm
that attracted the compensation, thus furthering goals of the tort system, such as
deterrence (17.7) and corrective justice (17.5). On the other hand, given that most
of the costs of tort compensation are widely spread across society as a whole
through insurance and the cost of goods and services, we might seriously question
the value of spending any money to transfer a proportion of such costs (in the form
of social security bene¬ts and NHS services) from the public in the guise of ˜the
taxpayer™ to the public in the guise of consumers of insurance, goods and services.
No doubt, this process of transferring costs from the public to the private ledger is
symbolically and politically important; but whether it makes any economic sense
is more doubtful. However, the main objection to the social security and NHS
recoupment schemes is not their cost, but that they entrench and impliedly endorse
the tort system as a mechanism for compensating disabled people.
Insurance companies, which are potentially the main bene¬ciaries of the
doctrine of subrogation, realize that it may not always be very useful.28 So-called

25 This was the main reason given by the Monckton Committee for rejecting subrogation in favour
of the National Insurance Fund: Final Report on the Committee on Alternative Remedies (Cmnd
6860, 1946).
26 Remember that the Pearson Commission found that it cost 85p to deliver every £1 of tort com-
27 The cost of the payments themselves fall, of course, on premium payers. It has been estimated that
extension of the NHS scheme to all compensation payments in 2006 will add about 5% to pre-
miums for relevant lines of insurance: Department of Health, The Recovery of National Health
Service Costs in Cases Involving Personal Injury Compensation: A Consultation (2002), para. 4.16.
28 R. Lewis, ˜Deducting Collateral Bene¬ts from Damages: Principles and Policy™ (1998) 18 LS 15,
A plethora of systems 385

˜knock-for-knock agreements™ (9.5) represented an abandonment of subrogation
rights “ although only in relation to property damage. In cases where an employee
renders the employer vicariously liable and is in law liable to recoup the employer™s
insurers, the insurance industry has voluntarily abandoned the right of sub-
rogation by agreement even in relation to personal injuries.29 In some countries
subrogation is severely limited by law; for example, in Denmark, a person who has
insured property against accidental damage has no tort action against a person
who damages it, and hence the insurers have no subrogation rights.30 Other
Scandinavian countries permit a tort action, but the insurance proceeds are
deducted from tort damages and once again there are no subrogation rights.31
Insurers say that without subrogation rights, premiums would have to be higher;
but others deny that subrogation rights are of much economic value overall, given
the cost of enforcing them.

15.4 Tort damages and other compensation
15.4.1 General principles
Two questions are addressed in this section. First, to what extent does tort law tol-
erate over-compensation by allowing a claimant to recover tort damages in respect
of any particular loss in addition to compensation from some other source in
respect of that same loss? Conversely, to what extent does tort law pursue a policy
of ˜reduction™ by setting o¬ against tort damages compensation received from some
other source? Secondly, to the extent that tort law does allow over-compensation,
how, if at all, can this be justi¬ed?
We will consider the second question ¬rst by looking at three traditional answers
to it. The ¬rst answer is that compensation from other sources should be ignored
if it is ˜collateral™ or comes from a ˜collateral source™. Taken literally, this might be
thought to mean that bene¬ts deriving other than from the tortfeasor will be
ignored, while bene¬ts coming from the tortfeasor (other than the tort compensa-
tion itself) will be taken into account. However, such a straightforward approach
has never been accepted by the courts: some bene¬ts provided by tortfeasors are set
o¬, but not all; and some bene¬ts received from third parties can be kept, but not
all. As a result, the word ˜collateral™ does not provide an independent criterion for
the decision whether to set o¬ or not to set o¬, but merely expresses a conclusion
based on a value-judgment about whether the injured person should or should not
be allowed to keep the bene¬t. The second answer is that compensation from other
sources should only a¬ect the tort compensation if the injury occasioned by the tort
was the causa causans (the ˜e¬ective™ or ˜proximate™ cause) of the receipt of the non-
tort compensation and not merely a causa sine qua non (or ˜the occasion™) of its

29 P.S. Atiyah, Vicarious Liability in the Law of Torts (London, 1967), 426“7.
30 J. Hellner, ˜Tort Liability and Liability Insurance™ in F. Schmidt ed., Scandinavian Studies in Law
1962 (Stockholm, 1962), 140“1.
31 Ibid.
386 Chapter 15

receipt. But as we saw earlier (5.2), the answer to the question of whether one thing
is merely a necessary condition of another or its ˜real cause™ depends on why the
question is being asked. Whether the tort ˜caused™ the receipt of the bene¬t boils
down to whether the recipient ought to be allowed to keep it or not.
The third answer is based on the assumption that tortfeasors actually pay tort
damages, and that therefore the reduction of tort damages because of compensa-
tion received from another source will ˜bene¬t™ the tortfeasor. Why, it is asked,
should this outside source of compensation ˜bene¬t™ the tortfeasor rather than the
victim? But once it is appreciated that only in a very small proportion of cases in
practice does the tortfeasor personally pay any damages at all, the argument col-
lapses. On the whole, this argument tends to receive short shrift in the courts, not
so much on the ground that tortfeasors do not pay damages but because it is essen-
tially an argument in favour of penal damages. If the claimant is already adequately
compensated, then obliging a defendant to pay money to the claimant for the same
loss on the ground that the defendant should not get the ˜bene¬t™ of the other com-
pensation, is merely to penalize the defendant. And on the whole this is now
The principal considerations relevant to answering the set-o¬ question are
these.32 First, in so far as the tort and the non-tort compensation are paid for by
essentially the same people, over-compensation should be avoided. This is particu-
larly important when the source from which the compensation comes is the State.
Public money should not be wasted by over-compensating some personal injury
victims, particularly when so many other deserving cases (such as victims of
disease) receive much less. In modern conditions, ˜public money™ is not just money
actually collected by the State in the form of taxes or social security contributions.
Tort damages too are, for all practical purposes, paid out of public money, since
they are mostly ¬nanced by road tra¬c and employers™ liability insurance pre-
miums, which are paid (or paid for) by a very large proportion of the public, and
are required by law to be paid.
A second consideration is the purpose of the compensation. If it replaces some-
thing (such as lost income) with a measurable ¬nancial value, then there can be
little justi¬cation for paying compensation more than once from more than one
source. If, on the other hand, the compensation is for something with no measur-
able ¬nancial value such as pain and su¬ering, the argument against paying com-
pensation from more than one source may seem less strong. Because nobody can
say what value we should put on pain and su¬ering, it cannot be said that a person
has been over-compensated by receiving money from more than one source. This
argument also appealed to the dissenting minority on the Monckton Committee,
who thought that full tort damages and social security payments should be pay-
able to accident victims because compensation for disabilities could never be

32 See also R. Lewis, ˜Deducting Collateral Bene¬ts From Damages: Principle and Policy™ (1998) 18
LS 15; Law Com. Consultation Paper 147, Damages for Personal Injury: Collateral Bene¬ts (1997).
A plethora of systems 387

excessive.33 However, this argument is undermined by the fact that compensation
for non-pecuniary loss is calculated and paid by each compensation system on the
assumption that this sum will be the only compensation payable on this account.
A third consideration is whether the people ¬nancing one or other of the sources
of compensation want the recipient to have compensation twice over. This is partic-
ularly relevant to charitable donations: donors sometimes intend the bene¬ciaries to
receive some bene¬t in addition to their legal claims, and this may be a good reason
for refusing to deduct the value of such donations from tort damages. On the other
hand, most charitable donors probably give no thought to the question of whether
the bene¬ciaries have legal compensation rights.
A fourth, and more di¬cult, argument which has played an important role in
some cases is that double compensation is less objectionable if the recipient has in
some sense ˜paid for™ the compensation. In Parry v. Cleaver34 Lord Reid said that the
real justi¬cation for allowing an accident victim to receive tort damages una¬ected
by the amount of any personal accident insurance payments was that the victim
had paid for the accident insurance and should therefore receive the bene¬ts of the
premiums. This is a tricky argument. On the one hand, there is no reason why a
person should not buy several life insurance policies, for example, and recover
under them all. However, the reason for this is not simply that the insured has paid
for all the policies. There are circumstances in which allowing a person to recover
from two sources may provide serious temptation for fraud. For example, a person
is not generally allowed to recover the value of damaged property twice over, even
if it has been insured twice over and the insurance has been paid for with two lots
of premiums. Furthermore, where a person receives payment under a ¬rst-party
insurance policy and is also awarded tort damages, that person will have paid for
the insurance but will not have ˜paid for™ the tort damages in the same sense.
A ¬fth policy consideration concerns the administrative cost of avoiding over-
compensation. If the amount it would cost to avoid over-compensation would
likely be as much as or more than the value of any excess compensation, this would
be a good reason for not seeking to avoid over-compensation. This may be why
statutory sick pay (SSP) is payable to those who receive occupational sick pay, and
why employers may not always deduct the value of SSP from occupational sick pay.
Let us now consider how the law actually deals with cases of potential over-

15.4.2 Tort damages and sick pay
First, if the employer is bound by contract or statute to pay the employee wages
while away from work through sickness, the employee has not ˜lost™ any income and
cannot recover it in damages.35 Secondly, if the employer is not so bound but pays
none the less, it is not clear whether this should be treated as a ˜charitable™ payment
33 Cmnd 6860, 53.
34 [1970] AC 1.
35 Parry v. Cleaver [1970] AC 1.
388 Chapter 15

which is not to be deducted from the damages even if the employer is also the defen-
dant,36 or as deductible because, even though voluntary, it is in the nature of sick
pay.37 Thirdly, if the employer, while not bound to pay the wages, pays them in the
form of a ˜loan™ subject to an undertaking to repay them out of any tort damages
that may be recovered, then again there will be no deduction.

15.4.3 Tort damages and personal insurance
In respect of property damage the owner cannot recover both tort damages and
private insurance. Once the insurer has paid out under the policy it is subrogated
to the owner™s tort claim: the property owner cannot, after collecting the insurance
money, sue in tort for their own bene¬t; any damages recovered by the owner must
be paid over to the insurer. By contrast, a victim of personal injury can recover
under a ¬rst-party insurance policy in respect of the injury and can also sue in tort
for damages on his or her own behalf. The insurer is not subrogated to the insured™s
right of action because personal injury insurance is not ˜indemnity™ insurance; and
so the insured is not required to hand any tort damages recovered over to the
insurer. Furthermore, the proceeds of personal injury insurance policies are not set
o¬ against tort damages.38 The most commonly given reason why the proceeds of
such policies are not set o¬ is that they are bene¬ts which the insured has bought
and paid for personally; but, of course, this is also true of property insurance.
This rule has caused great di¬culty in the case of payments made by employers
to injured workers. If the payments are essentially sick pay, then the tort damages
are reduced by their value even if they are funded by an insurance policy taken out
by the employer.39 Similarly, payments under a group personal accident policy
taken out by an employer for the bene¬t of employees will be set o¬ unless the payee
actually paid or contributed to the premiums.40 If a person injured at work is made
redundant or accepts redundancy, any redundancy payments received will be set o¬
against tort damages in respect of the injury, provided the worker would probably
not have been made redundant if they had not been injured.41 But if the payments
are in the form of a pension payable to an employee who is no longer able to work
because of disabilities, they will not be set o¬ against tort damages.42 This is so
whether the pension scheme is voluntary or compulsory and (apparently) whether

36 This view is indirectly supported by McCamley v. Cammell Laird Shipbuilders Ltd [1990] 1 WLR
963. See 15.4.4 concerning charitable payments.
37 This view is indirectly supported by Parry v. Cleaver [1970] AC 1 in which the non-deductibility
of an occupational pension was una¬ected by whether the pension scheme was voluntary or not.
38 Bradburn v. Great Western Railway (1874) LR 10 Ex 1.
39 Hussain v. New Taplow Paper Mills [1988] AC 514.
40 Gaca v. Pirelli General Plc [2004] 1 WLR 2683.
41 Colledge v. Bass Mitchells & Butlers Ltd [1988] 1 All ER 536.
42 Parry v. Cleaver [1970] AC 1. However, in respect of the period after the date on which the
employee would normally have retired, the proceeds of a disablement pension are set o¬ against
any retirement pension the employee would have received if the employee had worked to normal
retiring age when assessing damages for loss of pension rights: Longden v. British Coal Corporation
[1998] AC 653.
A plethora of systems 389

the employee makes contributions or not;43 and even if the employer is also defen-
dant and provider of the pension (that is, even if the scheme is not funded by an
insurance policy taken out by the employer).44
The confusion here results from the fact that di¬erent criteria are in use: the
basis on which sick pay is set o¬ is its nature “ sick pay is a form of wages, and an
employee who receives sick pay su¬ers, to that extent, no loss of wages. The basis
on which redundancy payments are set o¬ is not that they are a form of wages
(which they are not), but that if the redundancy was a result of the injury, the
redundancy payment is a bene¬t accruing from the injury and must, as a matter of
general principle, be set o¬ against loss resulting from the injury. The basis on
which occupational pensions are not set o¬ is di¬erent again, namely that they are
in the nature of insurance proceeds paid for, either directly or indirectly, by the
employee. The trouble is that these di¬erent criteria may con¬‚ict. For example,
an occupational pension paid to a disabled worker who retires early as a result of
the disability does provide a substitute for wages even if it is also the proceeds of
an insurance policy. Even more problematically, whereas payments from an
employer™s non-contributory pension scheme will not be set o¬ because they are in
the nature of insurance proceeds, payments from an employer™s group personal
accident insurance policy will be set o¬ unless the employee paid or contributed to
the premiums; and an employee will not be treated as having contributed to the
cost of such a policy merely by reason of the fact that the employer bought the
policy for the bene¬t of its employees.45 In fact, it seems that the law is in a transi-
tional state, and its ultimate destination may be a rule that all payments made by
the tortfeasor to the victim will be set o¬.

15.4.4 Tort damages and charitable payments
Payments made to the injury victim by a person who is under no legal obligation
to make them “ i.e. payments in the nature of charitable donations or made out of
˜benevolence™ “ are ignored unless made by the tortfeasor.46 Similarly, no deduc-
tion from damages is made for the value of services gratuitously rendered to an
injured person by relatives or friends (other than the tortfeasor); furthermore, the
damages will include a sum to enable the victim to pay for the services.47 Nor is the
value of free public health care deducted (individuals pay for this through tax-
ation), except that if the injured person saves living expenses by being looked after

43 On the basis that even if the employee does not contribute directly, he or she will have paid for
the pension indirectly in the form e.g. of lower wages. But this is also true of occupational sick
pay. The di¬erence between sick pay and an occupational pension is said to be that the former is,
but the latter is not, a form of wages.
44 Smoker v. London Fire and Civil Defence Authority [1991] 2 AC 502.
45 Gaca v. Pirelli General Plc [2004] 1 WLR 2682.
46 Gaca v. Pirelli General Plc [2004] 1 WLR 2682; Williams v. BOC Gases Ltd [2000] ICR 1181.
47 Hunt v. Severs [1994] 2 AC 350. In Australia, damages under this head can be awarded even if the
carer is the tortfeasor; and the Law Commission has recommended that this should be the law in
England, too: Law Com. No. 262, Damages for Personal Injury: Medical, Nursing and Other
Expenses; Collateral Bene¬ts (1999), 3.67“76.
390 Chapter 15

in a public institution, the value of such savings is deducted from damages for loss
of earnings.48

15.4.5 Tort damages and social security bene¬ts
Under the workers™ compensation system the injured worker was required to elect
between suing the employer in tort at common law and claiming workers™ com-
pensation. Beveridge apparently favoured the abolition of the employee™s tort
action: one of the justi¬cations he gave for treating victims of industrial injuries
preferentially in the social security system was the argument that only if this were
done would it be possible to restrict liability at common law in industrial injury
cases. However, Beveridge did not consider this question fully but recommended
that a special committee should be set up to inquire into the question of ˜alternative
A committee (consisting principally of lawyers) was set up under the chair-
manship of Sir Walter Monckton.49 The committee recommended that the tort
action should be retained and that the rule of election should be abolished. It also
decided that entitlement to National Insurance bene¬ts should be una¬ected by
the possibility that tort damages would be recovered in the future, if only because
it is necessary to pay such bene¬ts at once and not to wait and see whether any
damages may be recovered. The committee was then faced with the question
whether the amount awarded in tort damages should be a¬ected by the receipt of
National Insurance bene¬ts. The committee thought that there was a funda-
mental di¬erence in principle between voluntary insurance and compulsory
national insurance, which is more like a tax.50 The committee recommended that
social security bene¬ts should be deducted in full from tort damages on the
ground (unjusti¬ed, as things have turned out) that this would discourage tort
actions. The strongest argument against duplication is that tort damages and
social security bene¬ts (whether contributory or not) are paid for by much the
same group of people (that is, a signi¬cant section of the public), and there is no
justi¬cation for paying double compensation for the same loss at expense of the
same group.
The trades unions opposed the o¬-setting of bene¬ts, and the political compro-
mise ¬nally enacted in s. 2 of the Law Reform (Personal Injuries) Act 1948 was far
more favourable to injury victims (and especially industrial injury victims) than

48 Administration of Justice Act 1982, s. 5.
49 The committee issued three reports: Cmnd 6580, 1944; Cmnd 6642, 1945; and Cmnd 6860, 1946.
The ¬rst two were concerned with the Law Reform (Contributory Negligence) Act 1945 which
was then under consideration. On the last see P. Bartrip, Workmen™s Compensation in Twentieth
Century Britain (Aldershot, 1987), ch. 10. It is not clear what e¬ect abolition of the rule of elec-
tion and the introduction of apportionment for contributory negligence had on the incidence of
tort actions against employers.
50 Cmnd 6860, para. 32. In fact, the distinction between voluntary and compulsory insurance is not
very relevant: a property owner cannot keep both tort damages and property insurance proceeds
even though such insurance is not compulsory. Again, occupational pensions are not set o¬
whether the pension scheme is voluntary or compulsory.
A plethora of systems 391

the committee had recommended. Instead of providing for full deduction of the
value of social security bene¬ts, the 1948 Act provided for the deduction from
damages for loss of earnings resulting from personal injuries of half the value of
certain (but not all) bene¬ts received or likely to be received during a period of
5 years from the date of the injury. The justi¬cation51 for deducting only half the
value of bene¬ts was that in respect of industrial bene¬ts the employee was paying
nearly half the cost (¬ve-twelfths) and so was entitled to receive half the bene¬t
without a¬ecting a tort claim. In fact, only in the case of industrial injuries was ¬ve-
twelfths of the cost paid by employees, yet the half-deduction rule was applied to
all bene¬ts. Moreover, that workers paid for half of their industrial bene¬ts is false
because contribution rates were not risk-related, which meant that workers in some
occupations were heavily subsidized by workers in others.52 There also appears to
have been no justi¬cation for the 5-year limitation. Although few tort victims
received social security bene¬ts for more than 5 years, it simply meant that where
common law damages were at their highest, the amount of duplication was the
greatest. Finally, the provision only applied to court awards and not to settlements
which, of course, greatly limited its impact.
The 1948 legislation only covered speci¬ed bene¬ts. The rules about bene¬ts
other than those speci¬ed in the Act were not uniform. State retirement pension
was not deducted at all;53 but attendance and mobility allowance,54 family income
supplement,55 income support,56 statutory sick pay57 and unemployment bene¬t,58
at least in respect of payments received before trial, were deducted in full.59
In 1989 a new system for recovery of social security bene¬ts was set up under
the Social Security Act 1989. The provisions of the 1948 Act continued to apply to
cases in which the value of any compensation awarded was £2,500 or less. In the
case of compensation payments over £2,500 (whether made in pursuance of a
court order or by way of a settlement out of court) the relevant rules were contained
in the 1989 Act (subsequently re-enacted as Part IV of the Social Security Admin-
istration Act 1992). The Act provided, in e¬ect, that the party paying the compen-
sation had to deduct from it the full value of speci¬ed social security bene¬ts
received by the payee in respect of the injury in a period of 5 years after the date of
the injury60 (or in the period from the date of the injury to the date of the

51 Which was speci¬cally rejected by the Monckton Committee.
52 13.1.2.
53 Hewson v. Downs [1970] 1 QB 93.
54 Hodgson v. Trapp [1989] AC 807. These bene¬ts have now been replaced by disability living
55 Gaskill v. Preston [1981] 3 All ER 427. The current equivalent of this bene¬t is working tax credit.
56 Formerly supplementary bene¬t: Lincoln v. Hayman [1982] 1 WLR 488.
57 Palfrey v. Greater London Council [1988] ICR 437.
58 Nabi v. British Leyland (UK) Ltd [1980] 1 WLR 529. The current equivalent of this bene¬t is job-
seeker™s allowance.
59 No deduction was made in respect of bene¬ts for which the claimant could have applied but had
not: Eley v. Bedford [1972] 1 QB 155.
Or, in the case of diseases, the date on which the claimant ¬rst claimed a relevant social security
bene¬t as a consequence of the disease.
392 Chapter 15

compensation payment, if this was less than 5 years),61 and had to pay the amount
deducted to the DWP. The recoupment scheme was (and is) administered by the
Compensation Recovery Unit (CRU) within the DWP, decisions of which are
subject to appeal to a social security appeals tribunal and thence, on a point of law,
to a Social Security Commissioner.
In e¬ect, the 1989 Act achieved what the Monckton Committee recommended.
But the mechanism by which it was achieved was quite di¬erent from that used in
the 1948 Act. Under the 1948 Act, the value of the social security bene¬ts was
deducted from the tort damages on the ground that the injured person should not
be compensated twice for the same need. However, the bene¬ciary of this deduc-
tion was not the source of the deducted payments but rather the tortfeasor. By con-
trast, under the 1989 Act it was not the compensation payer but the taxpayer who
bene¬ted. From one point of view, this di¬erence is not of great importance
because, at the end of the day, compensation for personal injuries and social secur-
ity bene¬ts are both paid for by a large section of members of the public. But the
distinction between public expenditure and private expenditure is of political
importance, and the main motivation for the 1989 provisions was to reduce public
spending. From the government™s point of view, the 1989 scheme also had the
advantage of casting the administrative cost of recoupment on to defendants and
their insurers.
The 1989 scheme was criticized on a number of grounds. First, it was said that
since National Insurance bene¬ts are paid for wholly by employers and employees,
these bene¬ts, at least, should not be deducted; but this is a weak argument because
National Insurance contributions are, in e¬ect, a tax. Secondly, it was argued that
in practice, many tort claimants do not receive full compensation, and that this
problem would be made worse by deduction of social security bene¬ts; but there is
no good reason why tort defendants should be subsidized by the taxpayer. Thirdly,
it was initially suggested that the total costs of recoupment (administrative costs,
increased delays in the settlement process and so on) might be greater than the
resulting bene¬t to the public purse; but this has proved not to be the case. In the
year to March 1998, the CRU recovered amounts equivalent to bene¬ts of
some £177 million at an administrative cost of less than £4 million. Fourthly,
certain details of the scheme and its operation caused much consternation. One
was the provision that social security bene¬ts were to be deducted in full from com-
pensation payments even if the compensation was reduced on account of contrib-
utory negligence, and even though contributory negligence is ignored in assessing
social security bene¬ts. Another was the provision that social security bene¬ts
were to be deducted from the total compensation and not just from that part
which ˜performed the same function™ as the relevant bene¬t. For instance, income

61 Because the recoupment period ends when the claim is settled, it is unnecessary to take account
of social security bene¬ts in calculating compensation for future loss. Receipt of compensation for
future loss may, however, a¬ect entitlement to social security bene¬ts after the date of settlement.
As in the case of the 1948 Act, there was no principled justi¬cation for the 5-year limitation.
A plethora of systems 393

replacement bene¬ts could be deducted not only from compensation for loss of
income but also from compensation for non-pecuniary loss. People were also out-
raged by the fact that after deduction of bene¬ts, a claimant might end up with little
or no compensation. Moreover, it was alleged that the level of settlements was being
depressed as insurance companies put pressure on injury victims to accept £2,500
or less in order to avoid the operation of the recoupment provisions. If we view the
1989 scheme as designed to prevent over-compensation and not merely to reduce
public spending, the concerns about the failure to set like o¬ against like and the
alleged behaviour of insurers certainly have some force.
The recoupment scheme was re-enacted and amended in 1997.62 The £2,500
threshold was removed,63 and social security bene¬ts are now deducted only from
compensation64 which performs a like function.65 There is no deduction from
damages for non-pecuniary loss. Nevertheless, the 1997 Act makes it clear that the
relevant social security payments are to be deducted in full even if the result is that
the injured person ends up with no compensation (because, for instance, the com-
pensation has been reduced for contributory negligence).66 Leaving these amend-
ments aside, the 1997 scheme is the same as the 1989 scheme, which it replaces.
One notable feature of both the 1948 and the 1997 provisions is that they (did)
do not apply to damages or settlements awarded under the Fatal Accidents Act
1976. Indeed, under this Act no bene¬ts accruing from the death of the deceased
are deducted from the damages awarded. Such generosity is hard to justify.
Criminal injuries compensation is exempt from recoupment because awards under
the CICS are reduced to take account of the receipt of bene¬ts.67 Payments under
criminal compensation orders are exempt presumably because they are typically
made by individuals, not insurers.
Apart from social security bene¬ts, we saw in chapter 14 that there are many
other social welfare services of which injured persons might be able to take advan-
tage. Few of these are taken account of by the tort system.68 Following the recom-
mendations of the Monckton Committee,69 the availability of treatment under the
NHS is ignored in assessment of tort damages, so that a person may have private

62 Social Security (Recovery of Bene¬ts) Act 1997 and Social Security (Recovery of Bene¬ts)
Regulations 1997.
63 But the Minister has power to make regulations to exempt ˜small payments™ from the operation
of the scheme.
64 And interest: Gri¬ths v. British Coal Corporation [2001] 1 WLR 1493. This case also held that com-
pensation for the value of gratuitous care constitutes ˜compensation for cost of care incurred™ for
the purposes of o¬-setting care-related social security bene¬ts.
65 But note that disablement bene¬t ( is deducted from compensation for loss of income.
Statutory sick pay is not set o¬, but will be taken into account in calculating loss of earnings.
66 The defendant must repay all the bene¬ts even if greater than the compensation.
67 The 1997 Act expressly provides (in s. 17) that bene¬ts covered by the recoupment scheme are to
be disregarded in assessing damages.
68 But see Administration of Justice Act 1982, s. 5 ( n. 101). See also Cunningham v. Harrison
[1973] QB 942, 954, 957; Taylor v. Bristol Omnibus Co. [1975] 1 WLR 1054, 1058, 1063.
69 Cmnd 6860, para. 56. Although, it seems, contrary to Beveridge™s inclinations: Beveridge Report,
para. 262.
394 Chapter 15

medical treatment and recover tort damages to cover the cost.70 Private medical
insurance is also ignored in the assessment of tort damages, so that a person can
have private treatment and recover the cost of this twice over.71

15.5 Criminal injuries compensation
The value of any present or future entitlement to social security bene¬ts is deducted
in full from compensation under the CICS (other than the standard amount of
compensation),72 and the CICA may refuse to make an award until reasonable steps
to claim social security bene¬ts have been made. With certain exceptions “ notably,
private health insurance “ the Board ignores the proceeds of insurance policies
taken out and paid for by the victim.73 Unlike the position at common law, the value
of an occupational pension received as a result of the injuries is taken into account
in assessing compensation for loss of earnings or dependency. If the pension is not
taxable, its whole value is deducted; if it is taxable, half of its value is deducted.74
This latter rule is rather generous, but may to some extent compensate for the fact
that the pension may have been contributory. Personal pensions are ignored.

70 The Pearson Commission recommended that the cost of private medical treatment should be
recoverable only if it was medically reasonable to incur it: Pearson Report, vol. 1, para. 342. The
Chief Medical O¬cer has recently recommended that the cost of private medical treatment
should not be recoverable in claims against the NHS: 6.2.3 n. 73.
71 BUPA tries to avoid this by requiring members to inform it of any negligence claim made and to
include medical expenses in any such claim. Since in practice the expenses will usually need to be
paid long before any damages are received, the member is expected to reimburse BUPA when
damages are recovered. It seems probable, nevertheless, that many members secure double recov-
72 The rules are complex. See D. Miers, State Compensation for Criminal Injuries (London, 1997),
73 Ibid., 235“7; but the cost of private medical treatment will not be awarded unless the treatment
and its cost were both reasonable (2001 Scheme, para. 35(c)).
74 Ibid., 237“9.

The cost of compensation and who
pays it

In this chapter we examine what the various compensation systems we have con-
sidered cost, and how those costs are paid. The Pearson Commission provided a lot
of information about the costs of the tort system in particular. There are two
di¬erent types of costs, private costs and social costs. The main function of com-
pensation systems is to transfer money from some people to others; the sums so
transferred are a (private) cost to those who have to pay them, but they are not a
social cost. They do not reduce society™s resources as a whole. In economic terms,
they are transfer payments. So far as transfer payments are concerned, the questions
of interest concern the total value of the payments and the way the burden is dis-
tributed. In the context of compensation for personal injuries, social costs are, in
essence, the administrative costs of making transfer payments. Administrative costs
are social costs because they are the measure of the administrative resources con-
sumed in making the transfer payments. The question of how much ought to be
paid out in compensation is answered by reference to the goals of the compensa-
tion system in question; but the question of how much ought to be spent in making
the compensation payments is purely a matter of e¬ciency. The lower the admin-
istrative costs of a system as a proportion of compensation paid out under the
system, the more e¬cient the system is as a compensation system, whatever its
goals.1 As we shall see, some compensation systems are much more expensive to
administer than others. Higher relative costs can be justi¬ed only if the system in
question delivers desired bene¬ts additional to the amount of compensation paid
out,2 and if the higher administrative costs are referable to those additional

16.1 The cost of tort compensation
The Pearson Commission estimated that in the years 1971“6 total tort compensa-
tion payments (at 1977 currency values) averaged some £202 million per annum,
of which £69 million (about 34%) was paid out in industrial injury cases and
1 Unless, perhaps, one of the goals of the system is to keep a certain number of people in employ-
ment as administrators.
2 Or, in other words, only if the system has some goal other than and additional to compensation.

396 Chapter 16

£118 million in road accident cases.3 The total costs of the system are nearly double
the amounts paid out in compensation because the tort liability insurance system
is so staggeringly expensive to operate. The Pearson Commission estimated that the
administrative cost of making the annual payments of £202 million averaged some
£175 million during this same period, 1971“6 (again in terms of 1977 currency
values). To translate these ¬gures into 2005 currency values, they would need to be
multiplied by about four, producing a ¬gure of about £1.5 billion. However, we
have also seen that the annual number of successful tort claims has approximately
trebled since the 1970s (8.1.4), suggesting a total cost of the tort system in the region
of £4.5 billion. This is very much less than estimates in the region of £10 billion per
annum made, for example, by the Institute of Actuaries in 2002.4 However, simply
scaling up the Pearson ¬gures in line with in¬‚ation and to take account of the
increase in claims would produce an underestimate because general levels of com-
pensation payments may (for various reasons) have increased at a faster rate than
general in¬‚ation.5 Advances in medical technology enable increasing numbers of
seriously disabled people (including children born disabled) to be kept alive for
long periods at great ¬nancial cost;6 several new heads of damages have been intro-
duced since 1977; the discount rate (6.4) has been halved, and levels of damages for
non-pecuniary loss (6.5) have been signi¬cantly increased, in recent years; the
schemes for recovery of social security bene¬ts paid to claimants and NHS hospi-
tal costs (15.3) has increased the total cost of tort compensation. There is also some
evidence to suggest that the Woolf reforms may have signi¬cantly increased legal
costs by what is called ˜front-loading™ “ i.e. by making it necessary to spend at an
early stage in the settlement process amounts the expenditure of which would for-
merly have been deferred until later. The earlier costs have to be incurred, the larger
the proportion of cases in which they have to be incurred, and so the greater the
legal costs of the system overall.7
But even if the cost of the tort system has doubled in real terms (taking
account of in¬‚ation and increased claiming), it is still much less than the most
recent guesstimates, which go as high as £14 billion or more. According to a
3 Pearson Report, vol. 1, tables 158, 159.
4 In The Cost of the Compensation Culture.
5 For example, the EU Working Time Directive, which limits working hours to 48 per week, is said to
have increased the cost of care: Department of Health, Making Amends: A Report by the Chief
Medical O¬cer (2003), 67. According to a large-scale survey of motor accident claims, the cost of
personal injury claims (compensation plus costs) increased by nearly 10% per annum in the decade
up to 2003 as a result of increasing numbers of claims (3% per annum) and increasing average cost
of claims (6.7% per annum). In¬‚ationary pressures are apparently greatest in relation to high-value
claims. Legal costs as a proportion of total cost have remained steady at around 30%: International
Underwriting Association of London, Third UK Bodily Injury Awards Study (London, 2003).
6 According to the Third UK Bodily Injury Awards Study increasing use of prosthetics (such as
arti¬cial limbs) has added signi¬cantly to medical costs in motor accident cases.
7 P. Fenn and N. Rickman, Cost of Low-Value Employers™ Liability Claims, 1997“2002 (DCA, 2003); J.
Peysner and M. Seneviratne, The Management of Civil Cases: The Courts and the Post-Woolf Landscape
(DCA Research Series 9/05, November 2005), p. 71. Note that the abolition of legal aid for most per-
sonal injury claims transferred costs from the public to the private sector but did not itself increase
costs. The uplifts associated with CFAs and the cost of ATE insurance represent new real costs.
The cost of compensation and who pays it 397

government report published in 2004, the cost of the tort system in the UK in 2000
was about 0.6% of GDP, which gives a ¬gure in the region of £6 billion;8 and a more
recent o¬cial document puts the cost in 2004 at £7.2 billion.9 We might then, with
some con¬dence, estimate the total annual cost of the tort system at between £6 and
£8 billion. This is obviously a very large ¬gure; but it is important to maintain a
sense of proportion. For instance, estimated total current (as opposed to capital)
expenditure on the NHS in 2002“3 was more than £64 billion, whereas the amount
paid out in tort compensation and costs by the National Health Service Litigation
Authority (NHSLA) in 2004“5 was about £528 million “ i.e. about 0.8% of total
NHS current account expenditure.
The aggregate Pearson ¬gure (£377 million in 1977 currency values) included
legal costs paid by insurers and self-insurers to claimants in settlements (as well as
those awarded in cases tried); legal costs paid by insurers and self-insurers to their
own legal advisers; other disbursements paid by insurers and self-insurers to
claimants (e.g. to cover the cost of medical reports); and the general administrative
costs of the insurers,10 or in the case of self-insurers, of their claims departments.
But this ¬gure does not include anything for the costs of running the courts, which
hear the very small proportion of cases that go to trial or appeal. Thus according to
the Pearson ¬gures, the administrative expenses of the system as a whole amount
to about 85% of the value of the sums paid out, or about 45% of the total costs of
the system.11 So (if we ignore insurers™ investment income) we can say that about
55 pence of the insurance premium pound is paid out to injured victims, and
45 pence is swallowed up in administration.
As we shall see later, no other compensation system is anything like as expensive
to operate as the tort system. The social security system, for instance, runs at a cost
of about 12% of the total amounts paid out;12 although it should be noted that this
¬gure does not include the cost of raising the revenue out of which social security
payments are made. The great expense of the tort system seems to be largely due to
two factors: ¬rst, most insurance companies pay large sums of money to brokers
and advertising agencies to sell their insurance policies; and, secondly, the tort
system requires a detailed examination of every claim for personal injury damages.
This processing of claims is very expensive not because of the high cost of liti-
gation (only a very small minority of claims are litigated), but for several other
reasons. First, it is necessary (in theory, at least) to ascertain in every case who, if

8 Better Regulation Task Force, Better Routes to Redress (2004), 15. This ¬gure is low by international
9 Compensation Bill Final Regulatory Impact Assessment (2005), para. 2.10. The predicted cost in
2008“9 is £10.2 billion.
10 Including the cost of selling insurance, buying reinsurance and meeting claims.
11 Ibid., para. 261.
12 Pearson Report, vol. 2, para. 158. Some bene¬ts are more expensive to administer than others
depending e.g. on the di¬culty of establishing the conditions of eligibility. The administrative
costs of the New Zealand accident compensation scheme are about 7% of the amounts paid out:
G. Palmer, ˜New Zealand™s Accident Compensation Scheme: Twenty Years On™ (1994) 44 U. of
Toronto LJ 223, 227.
398 Chapter 16

anyone, was at fault and whether that fault caused the claimant™s injuries. These are
very di¬cult questions to answer in many cases, and the process of investigation
may be time-consuming (and time is money); it requires expertise (which costs
money); and it may require the interviewing of witnesses, the taking of statements,
the commissioning of technical or medical experts to analyse the evidence and so
on (all of which can be very expensive). Secondly, the compensation payable in a
tort case depends on the medical condition and other circumstances of the particu-
lar claimant; there are no rigid tari¬s or formulae for assessing tort compensation “
every case is di¬erent. Thirdly, because of the adversarial nature of the tort process,
both parties will employ advisers to make the same inquiries about the causes and
consequences of the accident. This duplication itself accounts for a considerable
proportion of the total cost.
The Pearson Commission estimated that fees (i.e. legal costs) paid by insurers
amounted to about 20% of compensation payments.13 As compensation payments
totalled £202 million, this means that fees paid to claimants must have totalled
about £40 million.14 If we assume that insurers had similar costs of their own in
dealing with the claim itself, this would account for about £80 million of the esti-
mated total of £175 million. The Civil Justice Review estimated that average (legal)
costs incurred in cases tried in the High Court are between 50% and 75% of the
amount recovered (depending on the basis of calculation); and in the county court
costs are (staggeringly) between 125% and 175% of the compensation awarded
(again, depending on the basis of calculation). These ¬gures also show that in rela-
tive terms, the smaller the claim the higher the costs: in the Civil Justice Review
survey the amount awarded in 89% of the county court cases was less than £3,000,
while only 41% of the High Court cases were disposed of for less than £3,000. Of
course, these ¬gures represent a maximum because they relate to cases that went to
trial. But the bulk of the costs of a personal injury action will be expended before
trial and will, moreover, need to be incurred whether or not proceedings are com-
menced. Clearly then, a major part of the total cost of the system consists of legal
costs “ 30% according to a recent survey.15 And it is the nature of the system that
renders these legal costs necessary, because it is almost indispensable for a person
claiming damages to have the assistance of a solicitor. By contrast, a person claim-
ing for damage to a car under a comprehensive policy would very rarely need to
consult a solicitor. Yet many such claims are more substantial than many personal
injury damages claims.
Who pays for all this? So far, we have contented ourselves with the general answer
that the ˜public™ pays. But it is now necessary to delve a little further into this ques-

13 Pearson Report, vol. para. 526.
14 In fact fees as such were paid in only 75% of cases, but presumably in other cases the claimant
received a su¬cient sum in settlement to pay their own fees. In relation to claims closed in 2004“5
by the NHSLA, defence costs were about 14% of damages paid and claimant costs were about
15 International Underwriting Association of London, Third UK Bodily Injury Awards Study.
The cost of compensation and who pays it 399

tion to see how the burden of payment is distributed. In a recent report, the Law
Commission suggested that the government is probably the largest single payer of
personal injury compensation in England and Wales; and in this sense, the public,
in the form of the taxpayer, bears a signi¬cant proportion of the costs of the tort
system.16 According to recent ¬gures,17 more than half of all personal injury com-
pensation claims arise out of road accidents, around one-quarter out of workplace
accidents and illness, about one-¬fth out of accidents in public places (˜public lia-
bility™) and a very small proportion out of clinical negligence and all other types of
liability, including product liability. In the case of workplace claims the cost is, in the
¬rst instance, paid for by employers, mostly in the form of premiums for employers™
liability insurance. These premiums are business expenses and are therefore tax
deductible. In this way part of the cost is borne by the taxpayer. The rest of the
burden is distributed by employers in the same way in which they distribute all their
costs, that is by passing them on to employees (in the form of lower wages), cus-
tomers (in the form of higher prices) and shareholders (in the form of lower divi-
dends). But how any particular cost is distributed between these groups is impossible
to say. One e¬ect of liability insurance is that the cost of injuries in one business will
be paid for partly out of premiums paid by other businesses. As between di¬erent
classes of businesses, insurance rates are principally adjusted according to the risk
they present (this is called ˜feature rating™ or ˜classi¬ed rating™). Businesses in an
industry with a higher injury rate will pay higher premiums than businesses in an
industry with a lower injury rate. Hence consumers, employees and shareholders of
more dangerous industries will pay rather more; though the di¬erences are in most
cases likely to be so small when spread among these groups that they will not be per-
ceptible at all. As between di¬erent enterprises within a single industry, the scope for
variation according to risk is limited by a number of technical considerations.
Although it might be thought that an employer with a bad injury record would have
to pay a higher premium than an employer with a good record, the use of such ˜expe-
rience rating™ is in fact quite restricted because it is actuarially unsound to base pre-
miums on the past record of a business unless it is quite large.18
In the case of road accidents, the cost, though still largely met by means of lia-
bility insurance, is spread in di¬erent ways. That part of the cost attributable to
commercial vehicles is distributed in much the same way as the cost of industrial
injuries, while the cost of accidents caused by publicly owned vehicles (of which
there are about 1 million in Britain) falls on the taxpayer. About four-¬fths of vehi-
cles on the road are cars and motorcycles, and most of them are privately owned;
so insurance costs in this area cannot be distributed beyond the vehicle owner. The
car owner pays for liability insurance with no tax relief, and cannot spread the cost
16 Law Com. No. 287, Pre-Judgment Interest on Debts and Damages (2004), para. 7.30.
17 The Recovery of National Health Service Costs in Cases Involving Personal Injury Compensation:
A Consultation (2002), para. 2.7.
18 P.S. Atiyah, ˜Accident Prevention and Variable Premium Rates for Work-Connected Accidents™
(1975) 4 Industrial LJ 1 and 89; but for a rather di¬erent view see J. Phillips,˜Economic Deterrence
and the Prevention of Industrial Accidents™ (1976) 5 Industrial LJ 148. See also
400 Chapter 16

any further. In the workplace context, the cost of liability is paid by a large class con-
sisting of employees, shareholders and consumers for the bene¬t, generally speak-
ing, of a subset of that class, namely employees. By contrast, in the road tra¬c
context the cost of liability is largely borne by a subset of the class of bene¬ciaries,
namely private motorists, for the bene¬t, generally speaking, of all road users,
including pedestrians and cyclists as well as motorists themselves.
As between motorists, the burden is distributed basically according to accident-
causing potential. In practice, insurers classify motorists into quite broad groups
according to certain factors which have been shown to be statistically signi¬cant in
accident involvement; in particular age, claims record, place of residence and type
of vehicle. A person who is a member of a high-risk group pays a larger premium
than a person who belongs to a low-risk group, despite the fact that the former may
in fact be a more responsible and careful driver than the latter and might have had
fewer accidents. In this way, some members of a risk group with a claims record
better than the average for that group may be subsidizing other members with a
worse-than-average claims record. The main reason why insurers classify motorists
into risk groups rather than according to the risk presented by each individual
motorist is that increased sub-classi¬cation would be expensive and complex to
administer.19 On the other hand, it may not be thought fair that a person should
have to pay a higher premium simply by virtue of being statistically more likely to
be involved in an accident despite the fact that they have a very good accident
record. However, the force of this argument is somewhat reduced by the fact that
the typical motor accident insurance policy incorporates a no-claims bonus based
entirely on the claims record of the policy holder. Also, insurers cannot a¬ord to
allow the risk-variation within an insurance group to become too large because, if
it does, it may cause the less risky members to terminate their insurance and seek
lower-cost cover from another insurer. In other words, competition between insur-
ers encourages a certain amount of risk-di¬erentiation.
Just as subsidisation may occur within risk groups, so too an insurer may decide
to let one risk group be subsidised by another. For instance, if an insurer thought
that the premiums chargeable to young drivers on strictly statistical principles were
so high that few young drivers would be able to a¬ord them, it might be tempted
to undercharge young drivers and to compensate for this by overcharging older
drivers. Once again, there are limits to which this is possible in a competitive insur-
ance market, although it is a common feature of non-competitive (or ˜mutual™)
insurance, such as that organized for the bene¬t of professional groups including
doctors and lawyers: for professional reasons, premiums for such insurance are
often much lower for newly quali¬ed professionals than is statistically justi¬ed on
the basis of risk.
We must also consider the costs of road accidents caused by uninsured
motorists. As we have seen, the MIB meets the cost of liability in cases of this kind

19 Moreover, as sub-categories are multiplied, they become smaller and less reliable statistically.
The cost of compensation and who pays it 401

as well as in other similar cases, such as where a person is injured by a ˜hit-and-run™
driver. The MIB collects its income from its constituent members, who are them-
selves insurers engaged in road tra¬c business in the UK, and all such insurers are
now required to be members of the MIB. This has the curious result that dis-
quali¬ed motorists who still drive but cannot get insurance and motorists who
simply fail to insure, are having their insurance paid for them by other motorists
who are insured with members of the MIB.
So far as public liability claims are concerned, the bulk of these are made against
local authorities, and so their cost is borne ultimately by taxpayers. Concerning the
cost of medical negligence liability, general practitioners insure with one of the
three medical defence societies, the best known of which is the Medical Defence
Union. These societies are mutual insurance organizations funded by subscriptions
from members. Such subscriptions are in e¬ect paid by the Department of Health
because they are taken into account in ¬xing remuneration levels. The societies also
o¬er members a range of claims-related services. Most medical negligence claims
are made against NHS hospital doctors for whose negligence, in most cases, a health
authority is vicariously liable. Until relatively recently, hospital doctors paid sub-
scriptions to one of the defence societies to cover the risk of liability, and the cost
of liability was shared between the doctor (that is, the defence society) and the hos-
pital authority (that is, the taxpayer). But in the 1980s subscriptions for hospital
doctors increased very rapidly; and there were plans to change the traditional prac-
tice of ¬xing subscriptions according to the number of years each doctor had been
in practice and to introduce risk classi¬cations according to the nature of the work
done. As a result, doctors in high-risk specialties, such as obstetrics and neuro-
surgery, would have paid much higher subscriptions than doctors in lower-risk spe-
cialties. The ¬rst response to rising subscriptions was that the government agreed
to pay doctors™ medical-defence subscriptions, whereas previously doctors had paid
them personally. Later, in response to fears that high, risk-related subscriptions
would have an adverse e¬ect on patterns of medical practice,20 the government
decided to introduce what is called ˜Crown (or NHS) indemnity™. Under this
arrangement, hospital doctors no longer paid for liability insurance, and the whole
cost of liability was to be borne by hospital authorities out of their own resources,
and hence, by the taxpayer. Now there is a central fund, the Clinical Negligence
Scheme for Trusts (CNST), administered by the NHSLA, into which NHS hospital
trusts make risk-related annual payments in the nature of insurance premiums to
cover compensation awards against them.
Finally, it must be remembered that in the case of injuries not caused by anyone™s
fault, or injuries for which the victim is partly to blame (and so for which they
receive only partial compensation), the cost of the injuries may be borne partly by
the victim personally. But part of it, at least, is likely to be met in other ways, such
as by the NHS, the social security system, social services and also by friends and

20 For a further discussion see
402 Chapter 16

family. The same is often true even in the case of fault-caused injuries, because
many victims of such injuries receive no tort compensation or an amount inade-
quate to meet their needs.

16.2 Costs not paid through the tort system
16.2.1 The cost of social services
The cost of social services is paid predominantly by payers of central and local gov-
ernment taxes. For example, the cost of medical treatment for injuries and diseases
falls mainly on the NHS, which is largely paid for out of central government taxes.
This is so regardless of whether the injury or disease was due to anyone™s fault, and
regardless of whether the victim recovers any tort compensation. A proportion of
the costs to the NHS of negligently caused road accidents is recovered from payers
of tort compensation under the Road Tra¬c (NHS Charges) Act 1999 (15.3).
In 2000, it was estimated that the total cost to the NHS of treating road accident
victims was £500 million;21 but this ¬gure relates to all road accidents, not just to
those in relation to which compensation is paid. In 2002“3 about £100 million was
recovered under the NHS recoupment scheme, but the total cost to the NHS of
treating compensated victims of road accidents is unknown. From October 2006
the cost-recovery scheme extends to all personal injuries (but not diseases) for
which tort compensation is paid, regardless of cause, and to ambulance as well as
hospital costs. 22
Legal aid is paid for out of general taxes. However, since 2000 only a few cate-
gories of personal injury claims qualify for legal aid; and the proportion of the legal
aid budget spent on such claims is very small. For present purposes, the most
important category is medical negligence claims. The great majority of legally aided
medical negligence claims end in payment of compensation to the claimant, and
this means that the claimant™s costs will be paid by the defendant rather than by the
Legal Services Commission. Any costs not recovered by the Commission from the
defendant will be recovered from the claimant out of the damages received.
However, since the defendant in the typical medical negligence case will be the
NHS, the taxpayer will bear the legal costs of most such successful claims whether
they are legally aided or not. Other relevant social services paid for by taxpayers or
ratepayers include the police, rehabilitation units, local authority welfare services
and so on. Part of the costs of these services is certainly attributable to road and
industrial injuries, but premium payers make no direct contribution to meeting it.
The Pearson Commission made a rough estimate that the annual total cost
of ˜public services™ provided to tort victims was around £525 million (in 1977

21 Royal College of Surgeons of England, Better Care for the Severely Disabled (2000), para. 4.3.
22 In 2000 it was estimated the ambulance costs associated with road accidents were about £20
million. In 2002 it was estimated that the annual cost to the NHS of treating ˜accidental injury and
poisoning™ was £2.2 billion (not including expenditure on rehabilitation): Department of Health,
Preventing Accidental Injury “ Priorities for Action (2002), 6.
The cost of compensation and who pays it 403

currency values, say £2.1 billion in 2005).23 A signi¬cant proportion of this large
total is externalized and not paid directly by the tortfeasors responsible for the acci-
dents or diseases, or by their insurers. Prior to the availability of the Pearson statis-
tics, it was estimated that in 1970, 42% of the cost of road accidents and 76% of the
cost of industrial accidents was borne by ˜external™ sources, such as the social secur-
ity system, the NHS, private insurance and sick pay arrangements.24 Now, of course,
as a result of the schemes for recouping the cost of social security bene¬ts and NHS
care, a smaller proportion of the costs of tort-caused injury and disability is exter-
However, paying liability insurance premiums is not the only way in which
vehicle owners contribute to the social costs of road transport, including road acci-
dents. They also make an indirect contribution by paying excise taxes on fuel and
vehicles, which raise many billions of pounds a year. Similarly, even if industry does
not directly pay the full cost of industrial injuries, it no doubt makes a substantial
indirect contribution in the form of taxation.

16.2.2 The cost of the social security system
The Pearson Commission estimated the total annual cost of social security pay-
ments to victims of tortiously caused personal injuries at £421 million (at an
administrative cost of £47 million “ both in 1977 currency values); but there have
been so many changes in the social security system since then that it is very di¬cult
to assess the contemporary validity of such ¬gures. It is probably safe to say,
however, that the equivalent amount in 2005 would not have been less than the
current value of the Pearson ¬gure “ say about £1.7 billion. The total cost of indus-
trial injuries bene¬ts in the late 1970s was around £259 million per annum (1977
currency values); the administrative cost of the IIS was some £28 million.25 The cost
of the IIS (and of other contributory bene¬ts) is borne by employee and employer
National Insurance contributions, which are e¬ectively a form of taxation. The cost
of non-contributory bene¬ts is met by the general taxation.
Although the industrial injuries system is still often treated as a species of
national insurance, the insurance element in the system is minimal. In the ¬rst
place, there are no ˜contribution conditions™; every employed person who earns
more than the contributions threshold is entitled to the same bene¬ts irrespec-
tive of contributions paid. Secondly, National Insurance contributions are not
related to the risk that the employee contributor will make a claim or that the con-
tributing employer™s business will precipitate a claim.26 This means, in e¬ect, that
the ˜good risks™ subsidize the ˜bad risks™. There are obviously sound reasons for not

23 Pearson Report, vol. 2, table 158.
24 D. Lees and N. Doherty, ˜Compensation for Personal Injury™ (April 1973) Lloyds Bank Review
18, 20.
25 These ¬gures are no longer identi¬able in the statistics; those given are based on the Pearson
Report, vol. 2, table 158, and Social Security Statistics 1978. Note that only a proportion of recipi-
ents of industrial injury bene¬ts would be victims of tortiously caused injury and illness.
26 See T.H. Marshall, Social Policy in the Twentieth Century, 2nd edn (London, 1967), ch. 4.
404 Chapter 16

adjusting employee premiums according to the risk of illness and incapacity.
Quite apart from the administrative complexity of doing so, the result would be
that those in the most serious need “ the chronically sick “ would either be
required to pay una¬ordably high premiums or would be given uselessly small
bene¬ts. Apart from the subsidy inherent in contributions unrelated to risk, the
other main source of subsidy in the National Insurance system for those most in
need of assistance derives from the fact that National Insurance contributions are
Beveridge gave some thought to whether, in the IIS, contribution rates could be
varied (as were rates for worker™s compensation insurance, which industrial
injuries was replacing) with the extent of the risk or the accident rate of the indus-
try in question.27 The trades unions (particularly the Mineworkers™ Federation)
were strongly opposed to variable premiums. They thought that miners™ wages were
depressed by the cost of mining accidents, and they argued that nearly all indus-
tries were dependent on coal and so ought to pay their share of the cost of coal
mining accidents through ¬‚at-rate premiums.28 Beveridge accepted these argu-
ments, though he wished to impose a special levy on industries with a particularly
high accident rate, primarily to encourage accident prevention. The government
was unimpressed by the suggestion that risk-related premiums deterred accidents,
and therefore rejected the idea for a special levy and adopted a uniform rate for con-
One has only to compare the rate of reported accidents in di¬erent industries to
get some indication of the extent of the subsidy involved in the uniform-rate prin-
ciple. In 2002“3 in the construction industry, for every 100,000 employees there
were 4 reported fatal accidents, 1,166 reported accidents causing injury leading to
more than 3 days o¬ work (including ˜major injuries™); whereas in the service sector
the corresponding ¬gures were 0.3 and 487 respectively.30 Another element of
subsidy inherent in the uniform-rate principle is that employees injured at work
receive higher bene¬ts (under the IIS) than other disabled employees. The Robens
Committee called for an urgent re-examination of the uniform-rate principle,31
but a majority of the Pearson Commission rea¬rmed it, mainly on the ground that
the cost of risk-related contributions would outweigh the bene¬ts.32
A third respect in which the industrial injuries system does not work on insur-
ance principles is that while National Insurance contributions are earnings-related,
with only very minor exceptions industrial injury bene¬ts are not.

27 Beveridge Report, paras. 86“9.
28 But the price of coal itself would have included some amount representing the cost of coal mining
accidents, and so industries dependent on coal were already contributing in this way to the cost
of coal mining accidents.
29 Cmnd 6551, para. 30.
30 Health and Safety Statistics Highlights 2002/3, supplementary tables 10 and 12. These ¬gures relate
to reported injuries (1.4.1 n. 22 for an explanation).
31 Robens Committee Report, para. 447.
32 See further
The cost of compensation and who pays it 405

Perhaps the major respect in which the social security system is based on insur-
ance principles is that in respect of a few bene¬ts, certain conditions relating to the
amount of contributions paid have to be satis¬ed before the bene¬t is payable. But
even here, the contribution conditions are not usually actuarially determined.
On the whole, then, the social security system is, in very many respects, based
not on insurance principles, but on welfare principles. This is most obviously true
of bene¬ts entitlement to which is not subject to contribution conditions and
which are not funded by contributions but entirely by general taxation.

16.2.3 Other sources of compensation
Other sources of compensation and provision for tort victims include contractual
sick pay, occupational pensions, personal accident and private health insurance and
medical services provided by employers. The Pearson Commission estimated that
tort victims annually received ˜other payments™ (omitting life insurance payments,
but including criminal injuries compensation: see 16.3) of some £204 million (in
1977 currency values); and ˜private services™ worth some £50 million (1977 values).
There is really no way of knowing how meaningful these estimates are; but no other
more recent estimates are available.

16.2.4 Costs in perspective
The Pearson Commission estimated the annual combined total value of all com-
pensation payments made to, and all services provided for, tort victims (and rela-
tives in fatal cases) at some £1.45 billion “ around £6 billion in 2005 currency.
Although there is no way of reliably testing this ¬gure, we can say with some
con¬dence that the actual total in 2005 was probably no less than this, and that it
may have been considerably more. Indeed, taking account of the fact that the there
are about three times as many tort claims per annum now as in the 1970s, and that
the cost of compensation and other bene¬ts to tort victims has probably increased
faster than the general rate of in¬‚ation, the actual total might be as high as
£20 billion.
In attempting to make some sense of such a ¬gure, several points in particular
should, perhaps, be borne in mind. First, according to the Pearson Commission,
only about a quarter of compensation payments to tort victims are made by tort-
feasors and liability insurers; and compensation payments represent only about
60% of the combined total. In other words, even taking account of the social secur-
ity and NHS recoupment schemes,33 it is probably the case that signi¬cantly less
than 20% of the combined total cost of compensation payments and services pro-
vided to tort victims is internalized to (i.e. paid by) tortious activities. The rest is
spread, by various mechanisms, such as taxation and the cost of goods and services,
widely across society. It is in this sense that it can be said (and is frequently said in

33 Which together bring in nearly £300 million a year and will yield perhaps £100 million more when
the NHS scheme is extended in 2006.
406 Chapter 16

this book) that no matter how or by whom the various bene¬ts are provided in the
¬rst instance, ultimately they are paid for by the public or, in other words, by society
as a whole.
Secondly, the bene¬ts provided to tort victims (and their relatives) represent
only a proportion of the total social costs of the death, illness and injury they su¬er.
For example, in calculating the ˜value™ of preventing road accidents, the Depart-
ment for Transport adopts a willingness to pay methodology34 that takes account
not only of ˜loss of output™ and medical expenses (which may be the subject of tort
compensation) but also costs which the tort system ignores, such as the cost of
policing and insurance administration, and ˜human costs™ representing ˜pain, grief
and su¬ering™ not only to the victim but also to relatives and friends; and in rela-
tion to fatal casualties, ˜the intrinsic loss of enjoyment of life over and above the
consumption of goods and services™.35 Application of this method produces an
average value of preventing a fatal road accident of almost £1.5 million (in 2002
currency values), and an average value of preventing a personal-injury accident
(whether serious or slight) of about £73,000. On the basis of such ¬gures, it is calcu-
lated that the total cost of road accidents in 2002 was more than £17.7 billion.
Thirdly, the combined total of the cost of bene¬ts provided to tort victims rep-
resents only a relatively small proportion of all the bene¬ts in money and kind pro-
vided in respect of death, illness and injury: most such bene¬ts are provided to
people who receive nothing from the tort system (including many tort victims). On
the other hand, fourthly, the group of bene¬ciaries of tort compensation no doubt
receives a disproportionately large share of the combined total of bene¬ts in respect
of death, illness and injury.

16.3 The cost of criminal injuries compensation
Compensation payments under the old CICS amounted to some £190 million in
1996“7. The number of claims and the amounts paid out increased greatly after the
Scheme was ¬rst introduced.36 The administrative cost of the Scheme in 1996“7
was 8.8% of total expenditure. In 2002“3, more than £160 million was paid out in
compensation under the new CICS. One of the main justi¬cations for introducing
the new enhanced-tari¬ scheme in 1996 was to reduce the cost of criminal injuries
compensation, including administrative costs. The former aim has been achieved
but, surprisingly, the cost of administering the new, supposedly simpler, Scheme is
signi¬cantly greater than that of administering the old Scheme: 13.5% of total
expenditure in 2002“3. Although the CICS requires (like the tort system) an analy-
sis of the circumstances of each case to ascertain the cause and gravity of the

34 6.5.1.
35 Highway Economics Note No. 1 (2002).
36 £400,000 was paid out in the ¬rst full year of operation. The total compensation paid out between
the date the Scheme started (on 1 August 1964) and 31 March 1997 was about £1.6 billion. Almost
a third of this sum was paid out in the years 1994“7.
The cost of compensation and who pays it 407

injuries, it is very much cheaper than the tort system to operate. In the ¬rst place,
the cost of raising the money to pay for the compensation is hidden in the cost of
the taxation system from which it is funded. The cost of raising the money to pay
compensation is a very substantial item in the tort system because of the high costs
of insurance brokerage and advertising. Secondly, the criminal injuries system is
cheaper than the tort system because it eliminates the adversary process and there-
fore involves only one set of costs rather than two. Legal aid is not available for
applications to the CICS, nor for appeals to the CICAP; and the costs of claimants
who are represented are not paid by the CICA.
Almost the whole cost of the CICS falls on the taxpayer, though it is in some
respects a mistake to look at the Scheme in isolation from other compensation
systems. It is by no means true that the whole cost of criminal injuries is met by the
Scheme. As we have seen, social security bene¬ts are deducted in full from awards
under the CICS. In 2002“3 the CICA recovered about £730,000 from the proceeds
of compensation orders and damages awards in favour of recipients of CICS
awards; and in an attempt to increase this ¬gure, the CICA now has power to
recover directly from convicted criminals. The main reason why the scheme is
¬nanced by taxation and not by any form of insurance is simply that the risk of
being injured by criminal attack is so small that the cost of administration would
be out of all proportion to the size of the premium which it would be necessary
to levy.37
One feature of the scheme, which may be thought to bear on the question of
its ¬nancing, is that a proportion of claimants receive injuries from assaults or
attacks which are in some sense related to their occupation. For example, police
o¬cers, security guards, bartenders, wages clerks, bus conductors and post o¬ce
workers are all particularly vulnerable to criminal attack. It could be argued that
employers ought to be made responsible for compensation awards when the
o¬ence is related to the claimant™s occupation on the basis that assaults of this kind
should be a charge on the business in question and not on the taxpayer. In practice,
however, this would probably create more problems than it would solve. In any
event, a substantial proportion of those whose occupations are relevant to the
o¬ence are public o¬cials of one sort or another “ police o¬cers, customs o¬cials
and so on.


. 14
( 18)