Private Law Liability of Public Authorities for Negligent
Inspection and Regulation
M. Kevin Woodall*
The growth of the public sector in this century has
greatly expanded the number of situations in which
persons have come to rely upon or are affected by gov-
ernment action and inaction. One result of this expan-
sion is that courts, especially in the last two decades,
have been forced to determine the extent to which
principles of private law liability apply to public
authorities. In the United States, the policy-operational
distinction combined with principles of specific detri-
mental reliance have been used by courts to determine
the private law liability of public authorities in cases
of negligent inspection and regulation. In the United
Kingdom, Australia, and New Zealand the policy-
operational distinction, however, has been rejected in
favour of an approach which relies exclusively on spe-
cific detrimental reliance. In contrast, the Supreme
Court of Canada has recently applied the policy-
operational distinction in conjunction with an Anns
foreseeability approach to duty of care when evaluat-
ing liability in comparable cases of negligent inspec-
tion and regulation.
In this article, the author criticizes the approach
currently used by the Supreme Court in assessing the
private law liability of public authorities for negligent
inspection and regulation. The author suggests that the
Court’s use of both the policy-operational distinction
and the Anns approach to duty of care as touchstones
of liability is misguided and should be replaced by a
specific detrimental reliance analysis. In the first part,
he criticizes the policy-operational distinction. He
argues that it is fundamentally ambiguous and fails
adequately to distinguish between distinctively gov-
ernmental and distinctively non-governmental activity.
In the second part, he criticizes the Anns prima facie
foreseeability approach adopted by the courts in deter-
mining the duties owed by governments to individuals.
He argues that the use of foreseeability obscures the
unique nature of a government’s relationship with its
citizens and the special duties which arise out of this
relationship. Following the move away from both the
policy-operational distinction and the primafacie duty
of care in the United States and Commonwealth
courts, the author suggests a theory of liability for pub-
lie authorities grounded in the specific relationship
which arises when a government induces a citizen to
rely on it for inspection and regulation, and the citizen
reasonably relies.
Au cours de ce sidcle, la croissance du secteur
public a multipli6 les situations ott les individus sont
touchds par l’action ou l’inaction des gouvernements.
Aussi, les tribunaux ont dO determiner dans quelle
mesure les principes de responsabilit6 civile s’appli-
quent aux organismes publics et ce, particulirement
au cours des deux deri~res ddcennies. Aux ttats-
Unis, deux principes sont employts par les tribunaux
pour delimiter la responsabilit6 civile ddlictuelle des
pouvoirs publics dans les cas d’inspections ou de
r~glements ndgligents. Le premier principe distingue
une dtcision de politique gouvernementale de sa mise
en oeuvre : seule cette demire peut engager la respon-
sabilit6 du gouvernement. Selon le second principe,
les tribunaux retiennent la responsabilit6 du gouveme-
ment lorsque la victime du dommage pouvait raison-
nablement compter sur l’action ou l’inaction du gou-
vernement ou d’un de ses organes (specific
detrimental reliance). Au Royaume-Uni, en Australie
et en Nouvelle-ZWlande, la distinction entre politique
et mise en oeuvre a 6t6 rejet e en faveur d’une appro-
che centrde uniquement sur les principes du specific
detrimental reliance. Rdcemment, la Cour supreme du
Canada a appliqu6 la distinction entre politique et mise
en oeuvre en combinaison avec les principes de prdvi-
sibilit6 61abords dans l’arret Anns de la Chambre des
Lords britannique.
Dans cet article,
‘auteur critique l’approche de la
Cour suprEme. I1 propose qu’on substitue A cette ana-
lyse ‘approche du specific detrimental reliance. Selon
lui, la dualit6 politique/mise en oeuvre est fondamen-
talement ambigue et ne permet pas de distinguer ad6-
quatement les activitds gouvemementales des activitds
non-gouvernementales. Puis, il critique l’approche
basde sur Ia prdvisibilit6 de l’arrdt Ans, utilise par
les tribunaux pour ddterminer l’tendue des obliga-
tions des gouvemements envers les individus. I1
soutient que Ie crit~re de Ia pr6visibilitd ne rend pas
justice A la relation unique d’un gouvernement avec
ses citoyens et aux devoirs particuliers qui naissent de
cette relation. Comme les tribunaux des l~tats-Unis et
‘auteur 6carterait ces thories. I1
du Commonwealth,
propose une thdorie centre davantage sur la spdcifi-
cit6 de Ia relation qui s’6tablit lorsqu’un gouvemement
incite un citoyen h reposer sur lui pour l’inspection et
la rdglementation et qu’en retour l’individu lui accorde
une confiance raisonnable.
* B.A., LL.B., B.C.L. Of the Bar of British Columbia. I would like to thank Mr. Justice W.J. Wal-
lace of the British Columbia Court of Appeal, who suggested the topic, made important comments
on several drafts, and generally encouraged me to persevere. I would like to thank Simon Coval
for his perceptive observations. I am especially grateful for the assistance of Susan Coristine,
whose suggestions, both editorial and substantive, are found throughout.
McGill Law Journal 1992
Revue de droit de McGill
To be cited as: (1992) 37 McGill L.J. 83
Mode de citation: (1992) 37 R.D. McGill 83
McGILL LAW JOURNAL
[Vol. 37
Synopsis
Introduction
1.
The Policy-Operational Distinction
A. Historical and Judicial Origins of the Policy-Operational
Distinction
1. Roots in American Law
2.
Anns and the Policy-Operational Distinction in the
Commonwealth
B. Critique of the Policy-Operational Distinction
1. Discretion and Policy
2.
3.
Liability for Failing to Make a Policy Decision
“Immunity” for Policy Decisions
C. The Policy-Operational Distinction Since Anns
1. England
2. Australia
3.
New Zealand
D. The Supreme Court of Canada Trilogy and the Policy-Operational
Distinction
1. Laurentide Motels v. Beauport
Immunity for Policy Decisions
a. The Policy-Operational Distinction
b. Firefighters’ Lack of Diligence
Just v. B.C.
a.
b. Applying the Policy-Operational Distinction as a Test
c. The Dissent of Justice Sopinka
d. Conclusion
Rothfield v. Manolakos
2.
3.
E. Policy-Operational Distinction: Conclusion
II.
Private Law Duties of Care
A. Duties of Care
1. The Anns Foreseeability Test
2.
Foreseeability, Duty of Care and Omissions
B. Specific Detrimental Reliance –
A Relationship of Proximity,
Good Samaritan Rules and the Duty to Warn
1. United States
2. Australia
19921
LIABILITY OF PUBLIC AUTHORITIES
3.
Specific Detrimental Reliance: Summary
C. The Supreme Court Trilogy
1. Rothfield v. Manolakos
a. Justice La Forest
b. The Dissent of Justice Cory
c.
Laurentide Motels v. Beauport
Just v. B.C.
a. Standard of Care
b. Specific Reliance
c.
Summary of Just
2.
3.
Summary of Rothfield v. Manolakos
Conclusion
Of course, it is not a tort for the Government to govern!
Introduction
When should a government agency be held liable in damages to a private
citizen for negligently faling to protect the citizen’s person or property? In prin-
ciple, governments should bear the same liability in tort and delict as similarly
situated persons. The question is: when is government activity sufficiently like
the activity of persons that the law regulating the civil relations between persons
can apply to the government?’ In three recent decisions, the Supreme Court of
Canada turned to the so-called “policy-operational distinction” to answer this
question.3 Once it has been determined that a government agency should bear
‘Dalehite v. U.S., 346 U.S. 15 at 57 (1953) [hereinafter Dalehite].
2There is some debate about whether the liability of governments to persons ought to be gov-
erned by a separate body of law altogether. The fact remains, however, that Canadian law does not
have, nor could it easily create, a separate body of droit administratif as in France. Analogizing
government activity to private activity so as to bring it under private law rules therefore remains
necessary. On the debate about whether a separate droit administratif should be recognized in the
common law, see B. Schwartz, “Public Tort Liability in France” (1954) 29 N.Y.U. L. Rev. 1432;
J.B.D. Mitchell, “The Causes and Effects of the Absence of a System of Public Law in United
Kingdom” [1965] P.L. 95; C. Harlow, “‘Public’ and ‘Private’ Law: Definition without Distinction”
(1980) 43 Mod. L. Rev. 241; G. Samuel, “Public Law and Private Law: A Private Lawyer’s
Response” (1983) 46 Mod. L. Rev. 558; C. Harlow, “Fault Liability in French and English Public
Law” (1976) 39 Mod. L. Rev. 516; E. Taborsky, “A Case for Administrative Tribunals” (1943-4)
7 Mod. L. Rev. 209.
3Laurentide Motels v. Beauport, [1989] 1 S.C.R. 705, 94 N.R. 1, 23 Q.A.C. 1 [hereinafter Lau-
rentide Motels cited to S.C.R.]; Just v. B.C., [1989] 2 S.C.R. 1228, [1990] 1 W.W.R. 385 [here-
REVUE DE DROIT DE McGILL
[Vol. 37
liability as if it were a person, the question becomes: would a person have a duty
to prevent injury to another in that situation? In seeking to determine whether
such a duty existed in the common law cases, the Supreme Court has turned to
the foreseeability test proposed by Lord Wilberforce in Anns v. Merton London
Borough Council.4
In the first part of this paper I look at the roots of the policy-operational
distinction in American law, and then trace the reception and ultimate rejection
of the distinction in the law of England, Australia and New Zealand. I conclude
that the concept of discretion, which lies at the centre of the policy-operational
distinction, is too ambiguous to provide a clear basis for distinguishing catego-
ries of governmental activity. Moreover, I argue that there is nothing in the con-
cept of discretion that is unique to government. Determining that a decision is
in the nature of policy or that a function is operational does not answer the ques-
tion of whether the defendant agency was or was not acting like a person.
In Part II, I trace the history of the foreseeability test since Anns. I show
that the Anns foreseeability test has been roundly rejected as a broad approach
to duties of care in negligence. I suggest instead that the Canadian courts should
follow the lead of the Australian and English courts, taking into account the par-
ticularities of the relationship between the parties, analysing categories of duties
on the basis of categories of relationships, and developing new duties by anal-
ogy to existing duties. I argue that in the category of cases represented by the
Supreme Court trilogy, where in each case it was alleged that a government
agency failed to prevent injury caused by a third person or by an act of nature,
it may be appropriate to recognize new duties based on an analogy to the duty
to rescue. A person may claim compensation for faulty regulation or inspection
where a government has intentionally acted in such a way: (1) that it would be
reasonable for the plaintiff to rely on the government agency to ensure his or her
safety or interests and (2) that the agency has actually induced the plaintiff to
forego other remedies or precautions against the risk.
It may be that even this approach is too restrictive to embrace the duties
which we expect a government to shoulder in an increasingly communitarian
society. However, if it is indeed desirable to create duties for which there is no
private law analogy, I believe that the justification for those duties must be fully
and openly articulated.
I. The Policy-Operational Distinction
The policy-operational distinction lies at the centre of the Supreme Court
of Canada’s attempt to distinguish those areas of governmental activity which
are sufficiently similar to personal activity to warrant the imposition of tortious
inafter Just cited to S.C.R.]; Rothfield v. Manolakos, [1989] 2 S.C.R. 1259, [1990] 1 W.W.R. 408
[hereinafter Rothfield cited to S.C.R.].
4[1978] A.C. 728, [1977] 2 All E.R. 492 (H.L.) [hereinafter Anns cited to A.C.].
1992]
LIABILITY OF PUBLIC AUTHORITIES
liability from those areas of governmental activity which are outside the scope
of the law of tort. If an impugned decision is said to be in the nature of “policy”
it is generally beyond review by the courts. In contrast, “operational” decisions
made during the implementation of a policy decision can give rise to private
rights of action, especially in negligence.
In this part I look at the history and judicial origins of the policy-
operational distinction. Through an examination of the treatment of the distinc-
tion by the courts in England, Australia, New Zealand and finally Canada, I will
demonstrate that the distinction is at best unhelpful and at worst confusing.
A. Historical and Judicial Origins of the Policy-Operational Distinction
1.
Roots in American Law
The discretionary (or policy)-operational distinction has its roots in the
American Federal Tort Claims Act.’ The Federal Tort Claims Act was enacted
to make the government liable in tort as if it were a person, so far as that is pos-
sible. It simplified the procedures for bringing actions against the federal gov-
ernment and for the first time abrogated the federal government’s sovereign
immunity from actions in tort.’
The following year the British Crown Proceedings Act was enacted with
the same objectives. “Sovereign immunity” in the U.S. and “Crown immunity”
in English law are rough equivalents. The enactment of the American and Brit-
ish acts gave “fresh impetus to the demand for similar … legislation in Canada.”‘
While the Federal Tort Claims Act (and later the state acts9) had the effect
of waiving immunity from tort actions generally, it did so in an indirect way. A
“tort claims act may be described as inverting the customary rule [that applies
5Federal Tort Claims Act, 28 U.S.C. 2671ff (1946).
6L.L. Jaffe, “Suits against Governments and Officers: Sovereign Immunity (Part 1) and Damage
Actions (Part I)” (1963-64) 77 Harv. L. Rev. 1, 209. Sovereign immunity is a quasi-constitutional
doctrine of questionable origin that holds that a sovereign power is immune from suit in its own
courts without its consent. See W. Holdsworth, A History of English Law, vol. 9, 3d ed. (London:
Methuen, 1944) at 8. But see Kawananakoa v. Polyblank, 205 U.S. 349 at 353 (1907) (per Holmes
J.); L.H. Tribe, American Constitutional Law, 2d ed. (Mineoloa, N.Y.: Foundation Press, 1988) at
174 n. 5.
71947 (U.K.), 10-11 Geo. VI, c. 44.
8D.P. Jamieson, “Proceedings by and against the Crown in Canada” (1948) 26 Can. Bar Rev. 373
at 373.
9At the time of the enactment of the Federal Tort Claims Act only some states had begun to limit
their own sovereign immunity. See E. Borchard, “Government Liability in Tort” (1948) 26 Can.
Bar Rev. 399 at 409ff. Most state constitutions bar actions against the state in state courts except
by consent. See Restatement of the Law (Second) of Torts, 2d ed. (Washington: American Law
Institute, 1965) [hereinafter Restatement] at Reporter’s Note to 895B; see ibid. at 256ff. for a list
of states, their constitutional provisions relative to tort immunity, and the present law.
McGILL LAW JOURNAL
[Vol. 37
to individuals] by creating immunity and establishing liability only by excep-
tion.”” The Federal Tort Claims Act provides that:
The United States shall be liable … in the same manner and to the same extent as
a private individual under like circumstances, but shall not be liable for interest
prior to judgment or for punitive damages …
The provisions of this chapter … do not apply to … any act or omission … based
upon the exercise or performance or the failure to exercise or perform a discretion-
ary function or duty … whether or not the discretion involved be abused.”
The “discretionary” half of the discretionary-operational dichotomy there-
fore comes from the act itself. As Reed J. observed in Dalehite, the first U.S.
Supreme Court case to consider the Federal Tort Claims Act, “[t]he statute is
unique in Anglo-American jurisprudence in its explicit exception for discre-
tion.”‘2 The original purpose of the exception was political: Congress feared that
opponents of activist New Deal projects would oppose them with vexatious
actions in negligence, trespass and nuisance. 3 The phrase “operational level”
was coined by courts to describe activities where the “discretionary function”
exception would not apply.” Thus the original dichotomy was between “discre-
tionary” decisions on one side, and “operational” decisions on the other. Discre-
tionary decisions were, by virtue of the Act, absolutely unreviewable. Opera-
tional functions, on the other hand, did not benefit from the qualified sovereign
immunity, and could give rise to the same private law duties as would be
imposed on persons.
It must be emphasized that the discretionary-operational distinction has
always been a rule of immunity from suit, not a source of private law duties.
Finding that a function is operational in nature does not create a presumption
in favour of finding private law liability; it simply means that the doors to that
1057 Am. Jur. 2d 27. The Federal Tort Claims Act is a series of nested exceptions. It holds that
the government is generally not liable, save as provided in the Act. The sections mentioned below
hold that it is liable as if it were a person, except where a list of express exceptions apply. The
discretionary function exception is among those exceptions.
“28 U.S.C. 2674 and 2680(a), cited in Indian Towing v. U.S., 350 U.S. 61 at 63 (1955) [here-
inafter Indian Towing].
12Supra, note 1 at 82 n. 27.
13Reed J. in Dalehite, ibid. at 30:
[The discretionary function exception is] a highly important exception, intended to pre-
clude any possibility that the bill might be construed to authorize suit for damages
against the government arising out of an authorized activity, such as a flood control or
irrigation project, where no negligence on the part of any Government agent is shown.
… The bill is not intended to authorize a suit for damages to test the validity of or pro-
vide a remedy on account of such discretionary acts even though negligently performed
and involving an abuse of discretion (emphasis added).
‘4lbid at 42. A similar dichotomy was developed by state courts considering state tort claims
acts. See Johnson v. California, 447 P.2d 352 (Cal. 1968) [hereinafter Johnson].
1992]
LIABILITY OF PUBLIC AUTHORITIES
possibility are not closed. Whether the parties are in a relationship of proximity
must still be determined by the ordinary private law rules. 5
2.
Anns and the Policy-Operational Distinction in the Commonwealth
The House of Lords decision in Anns is important because it introduced the
American “discretionary-operational” distinction (renamed the “policy-
operational” distinction) into the Commonwealth. 6 Anns also raised foreseea-
bility to a position of unwarranted preeminence among the many criteria for
determining the existence of a duty of care. Our present interest is in the policy-
operational distinction. I will discuss the duty of care issue below.’7
Prior to Anns the leading English case on the liability of public authorities
in tort was East Suffolk Rivers Catchment Board v. Kent. 8 Lord Romer held that
an authority could be liable when it created fresh risk and inflicted injury, but
that it could never be liable for failing to exercise a “discretionary power.” For
example, an authority that had the power to light city streets could not be liable
for negligence in carrying out that power. 9 This line of analysis focused on
whether the authority had a duty or a mere power to do what the plaintiff alleged
it had failed to do. Moreover, the House affmned the more general principle that
the duties of public agencies are governed by statute law (public law) not the
common law (private law).
In Anns, Lord Wilberforce recognized that the scope of the duties owed by
public agencies to persons is primarily a matter of public law, but added that it
is not exclusively so. In his view, the development of the concept of proximity
in the modem law of torts required a reconsideration of the civil responsibility
of public authorities to persons who are closely affected by their activities. “The
problem …” he wrote, “is to define the circumstances in which the law should
impose, over and above, or perhaps alongside, these public law powers and
15I will discuss the significance of this point in more detail in part H.
161t has been suggested that the policy-operational distinction was the basis of the decision of
Laskin J. (as he then was) in Welbridge Holdings Ltd v. Winnipeg (City), [1971] S.C.R. 957, 22
D.L.R. (3d) 470, [1972] 3 W.W.R. 433 [hereinafter Welbridge Holdings cited to S.C.R.]. See P P
Craig, “Negligence in the Exercise of a Statutory Power” (1978) 94 L.Q. Rev. 428 at 438 n. 55;
PW. Hogg, Liability of the Crown, 2d ed. (Toronto: Carswell, 1989) at 125. This view involves
a misreading of the case and a misunderstanding of the policy-operational distinction. Welbridge
Holdings held that a municipal council could not be held liable in negligence to building devel-
opers who lost money when the failure to observe legislative formalities invalidated a by-law. The
case is concerned with the immunity of municipalities in their “legislative and quasi-judicial”
capacities contrasted with their corporate capacity; it was not concerned with distinguishing among
levels of discretion within the executive branch of government.
17See below, subsection II.A.1.
11[1941] A.C. 74, [1940] 4 All E.R. 527 (H.L.) [hereinafter Kent cited to A.C.].
19Discussing Sheppard v. Corporation of Glossop, [1921] 3 K.B. 132, [1921] All E.R. 61 [here-
inafter Sheppard], cited in Kent, ibid. at 99ff. Lord Atkin dissented.
REVUE DE DROIT DE McGILL
[Vol. 37
duties, a duty in private law towards individuals such that they may sue for
damages in a civil court.”2 In other words, public agencies should in principle
be subject to common law duties of care as well as to statutory duties. Lord Wil-
berforce seemed to be saying that the Kent analysis, which turned on whether
the authority was acting under a power or a specific duty, did not address the
question of proximity, which relates to the actual interaction between the par-
ties. In other words, it did not address whether the agency’s actions brought it
into a relationship of proximity with the plaintiff.
In distinguishing Kent, Lord Wilberforce turned to the discretionary-
operational distinction. Kent was correct so far as it went, he wrote, but it failed
to recognize the discretionary-operational distinction. The rule in Kent’s case
must now “be understood and applied with the recognition that, quite apart from
such consequences as may flow from an examination of the duties laid down by
a particular statute, there may be room, once one is outside the area of legiti-
mate discretion orpolicy, for a duty of care at common law.”‘” Since operational
functions are generally outside the scope of legitimate discretion or policy
“there is room” for a duty of care at common law. The idea seems to be,
although it is nowhere stated directly, that it is more likely that a relationship
of proximity would arise from operational functions than from policy decisions.
Whether the agency was acting under a “power” or a “duty” is irrelevant.
B. Critique of the Policy-Operational Distinction
It appears that Lord Wilberforce considered the policy-operational distinc-
tion to be part of the test for determining when a government, a political abstrac-
tion, is in a relationship of proximity with a person in a manner analogous to
the relationships of proximity between persons. That, however, is not the pur-
pose for which the discretionary function exception was created by the drafters
of the Tort Claims Act.
Our basic theme is that the policy-operational distinction was intended to
define those situations where governments act sufficiently like persons so that
the rules governing the relations between persons can apply to the government
as well. There are, however, at least two ways in which governments are not like
persons. First, government is political and legislative: it has the power and the
duty to act in ways that persons may not and it redistributes costs, benefits and
risks among the persons it governs. This is the aspect of government with which
the American drafters of the discretionary function exception were dealing. Sec-
ond, governments are abstract: we cannot have a relationship of proximity with
a government, in either a physical or moral sense, in the same way we have rela-
tionships of proximity with persons. When we speak of a person being in a rela-
2 0Supra, note 4 at 754.
2’Ibid. at 758.
1992]
LIABILITY OF PUBLIC AUTHORITIES
tionship of proximity with the government, we really mean that the person has
come to be affected by a program or human agent of the government much more
directly than other persons. This is the aspect of government Lord Wilberforce
seems to have had foremost in his mind, although he was clearly concerned with
the political and legislative aspect as well. Discretion, which lies at the heart of
the policy-operational distinction, does not adequately address either of these
issues.
1.
Discretion and Policy
In American law, agencies exercising “discretion” are immune from suit in
tort. Agencies performing operational functions, by definition non-
discretionary, are subject to normal private law rules.22 In Anns, the
“discretionary-operational” distinction, transformed into the policy-operational
distinction, continues to rely on discretion as the decisive criterion for determin-
ing whether the private law rules of tort will be applied. American commenta-
tors find the concept of discretion unhelpful due to its multiplicity of meanings.
Lord Wilberforce, in Anns, imported this ambiguity into the Commonwealth. In
addition, he introduced the concept of discretion into the operational level, fur-
ther muddying the distinction.
The problem is that the concept of discretion is ambiguous. In different
contexts it can mean very different things.’ To say a decision-maker has non-
reviewable discretion might be an allusion to the constitutional separation
between the judicial and executive branches: the judiciary is cautious about
commenting on executive decisions, and vice versa. The weakness with this
interpretation is that the force of executive privilege declines as one descends
through the bureaucracy. The implication of the policy-operational distinction is
that the decision-maker’s status might be relevant, with discretionary decisions
being made at “high levels.” However, the policy-operational distinction does
not provide a test for distinguishing between high and low level decision-
makers, nor does it address the principal question of when government action
is so similar to private action that it should be subject to private law. Discretion
might also mean that a decision-maker has a broad range of choice from which
to choose a course of conduct. If one saw that a decision-maker had only two
or three options one might conclude that his or her choice involved very little
discretion. But to conclude that “bad” decisions from among a narrow range of
choices should give rise to private law liability is a non sequitur. There is noth-
ing in the idea of range of choice that is either uniquely governmental or
uniquely akin to personal conduct.
22See text accompanying note 11.
23See, for example, R. Dworkin, Taking Rights Seriously (Cambridge: Harvard U. Press, 1978)
at 31, 69.
McGILL LAW JOURNAL
[Vol. 37
Finally, “discretion” might mean that the decision-maker must take into
account a variety of complicated and perhaps technical considerations, and
render a decision that is the product of judgment on matters about which rea-
sonable persons would differ. An example is “the striking of a just balance
between the rival claims of efficiency and thrift” mentioned by duParcq L.J. in
the Court of Appeal judgment in Kent.24 American authorities generally cite
similar reasons for refusing to permit executive decisions to be second-guessed.
In some cases courts focus on the specialized expertise of government agencies.
It is said that the “vigorous decision-making” of the executive would be
impaired if every functionary were required to “exercise his judgment and at the
same time [be] held responsible according to the judgment of others who may
have no experience in the area and may be much less qualified than he to pass
judgment in a discerning fashion or who may now be acting largely on the basis
of hindsight.”‘ Unfortunately, the discretionary-operational distinction is of no
use in segregating “simple” decisions from those which require special exper-
tise. Discretionary (or policy) decisions might involve weighing competing
interests which are easily understood by lay-persons. On the other hand, day to
day operations may require judgment which is beyond the reach of the average
person.26 Courts are properly reluctant to second-guess decisions of this type.
“Discretion,” then, has a variety of meanings which courts seldom articu-
late when they rely upon it as the basis of an authority’s liability. Even if a court
is careful to identify the type of discretion it has in mind, the discretionary-
operational distinction would not tell us whether a “bad” decision ought to incur
private law liability. Lord Wilberforce sometimes equates “legitimate discre-
tion” with policy,27 implying, in line with the American cases, that that which
is not policy, i.e. “operational” activity, is not discretionary. At other times he
24[1940] I K.B. 319 at 338, [1939] 2 All E.R. 207 (C.A.), cited with approval by Viscount Simon
L.C. in the House of Lords decision, supra, note 18 at 86.
2’Restatement, supra, note 9 at 895D(b), commentary. See also O.M. Reynolds, “The Discre-
tionary Function Exception of the Federal Tort Claims Act” (1968) 57 Georgetown L.J. 81 at 121.
261n Weiss v. Fote and the City of Buffalo, 167 N.E.2d 63 (N.Y. 1960) for example, the New York
Court of Appeals held that the statute abolishing state immunity did not permit actions in negli-
gence based on allegedly negligent placement of street lights. The Court held that recognizing such
suits would require turning over to lay-persons (the jury) decisions about traffic control that could
only be decided by experts. Is the placement of traffic lights in the nature of policy, or is it oper-
ational? One would expect this answer to be the same as the answer to the question of whether
the decision about when to inspect for unstable rocks is by nature policy or operational: in prin-
ciple, the decisions seem similar. But the court in Weiss evidently thought the traffic light decision
was discretionary (unreviewable), while the Supreme Court of Canada thought the rock inspection
decision in Just was operational (reviewable). The point is not that one court was wrong and the
other right. The point is that the discretionary (or policy)-operational distinction does not tell us
whether the decisions ought to be reviewable or not.
27Anns, supra, note 4 at 758: “[Tihere may be room, once one is outside the area of legitimate
discretion or policy, for a duty of care at common law.”
19921
LIABILITY OF PUBLIC AUTHORITIES
says that “operational” decisions may have discretionary elements.’ Opera-
tional decisions therefore seem to fall into two, or perhaps more, categories:
first, those that are merely mechanical, where the functionary has no discretion,
and which are therefore fully reviewable and second, those that involve the
implementation of a higher level policy decision, but nevertheless permit wide
(non-reviewable) discretion. Thus, while “policy” decisions (whatever they may
be) are clearly discretionary, “operational” decisions may or may not be. In
other words, describing a decision as being in the nature of “policy” seems to
dispose of the issue, but finding that it is operational does not.
For these and other reasons the American literature has criticized the dis-
tinction.29 As Louis Jaffe, a commentator on American public liability, has
stated, “the dichotomy … is at least unclear, and one may suspect that it is a way
of stating rather than arriving at a result.”3 The Americans are stuck with the
dichotomy because it is embodied in their statute. In Canada we should exorcise
it.
2.
Liability for Failing to Make a Policy Decision
As we have seen, operational functions may or may not be discretionary
in nature, so they may or may not attract private law liability. It gets worse.
Under American law, discretionary (policy) powers are simply incapable of
attracting private law duties. In contrast, according to Lord Wilberforce, when
authorities exercise discretionary powers, “[t]heir immunity from attack, though
great, is not absolute.”31 The immunity is “not absolute” because the authority
is “under a duty to give proper consideration to the question whether they
should [exercise the discretionary power] or not.”32 The duty to give proper con-
sideration is a common law duty, the breach of which gives rise to a claim in
damages by persons. This “somewhat revolutionary and illogical proposition” ’33
merits close consideration because it is the basis of the Supreme Court decisions
25Ibid, at 755.
29See, for example, Note, “Discretionary Planning-Operational Dichotomy Revisited” (1966) 41
Wash. L. Rev. 340; L. Green, “Fort Law: Public Law in Disguise” (1959-60) 38 Texas L. Rev. 1,
257 (reprinted in L. Green, The Litigation Process in Tort Law, 2d ed. (Indianapolis: Bobbs-Merrill,
1977)) at 115; H.R. Hink & D.C. Shutter, “Some Thoughts on American Law of Government Tort
Liability” (1965-66) 29 Rutgers L.J. 710; L.S. Jayson, “Application of the Discretionary Function
Exception” (1964) 24 Fed. Bar J. 153; C.I Peck, “Laird v. Nelms: A Call for Review and Revision
of the Federal Tort Claims Act” (1972-73) 48 Wash. L. Rev. 391 [hereinafter Review and Revision];
C.J. Peck, ‘The Federal Tort Claims Act: A Proposed Construction of the Discretionary Function
Exception” (1956) 31 Wash. L. Rev. 207; W.E. Steffen, “A Federal Tort Claims Act Tort” (1967)
34 Tenn. L. Rev. 367.
30Jaffe, supra, note 6 at 218.
31Supra, note 4 at 755.
321bid.
33Craig, supra, note 16 at 448.
REVUE DE DROIT DE McGILL
[Vol. 37
in City of Kamloops v. Neilson,34 Laurentide Motels, 35 and Just,36 as I will show
later.37
To begin with, it seems that Lord Wilberforce intended that the term
“proper consideration” be given the narrowest of meanings. An authority would
fail to give a question “proper consideration” only if it made a decision in bad
faith, or refused to make a decision in circumstances that amounted to bad faith.
If “improper” were to include “unwise” or “imprudent'” or “negligent” the pas-
sage quoted immediately above would be incompatible with the rest of his
speech. However, as will be seen, that is the interpretation of the Supreme Court
of Canada. 8
It is at best doubtful whether negligence is the appropriate doctrine for
dealing with gross inaction or bad faith of public authorities. Negligence implies
inattention or inadvertence where attention and care are required; bad faith
implies deliberate action for unacceptable reasons. Such a policy decision
would be ultra vires. One commentator suggests that a plaintiff who was injured
by an ultra vires decision (or absence of decision) should be entitled to compen-
sation. A cause of action claiming such compensation would, however, be sui
generis: it would be an action with no private law counterpart.39 Insisting that
such a cause of action is not in negligence is more than a quibble about forms
of action. By emphasizing the special nature of the cause of action the court nat-
urally turns its mind to the proper tests and the different bases upon which lia-
bility can be imposed.4″
31[1984] 2 S.C.R. 2, 10 D.L.R. (4th) 641 [hereinafter Kamloops].
35Supra, note 3.
361bid.
37See discussion below, subsections II.C.2 and 3.
38See text accompanying note 58.
39p. p. Craig, “Compensation in Public Law” (1980) 96 L.Q. Rev. 413 at 438-40.
40Gibbs C.J. and Mason J. in Sutherland Shire Council v. Heyman (1985), 60 A.L.R. 1, 59
A.L.J.R. 564 (H.C.) [hereinafter Sutherland Shire cited to A.L.R.] at 15-17 explicitly reject the pos-
sibility of private law liability for failing to make a policy decision. Mason J. wrote:
Although Anns has since been applied in the House of Lords … and by the Supreme
Court of Canada in Kamloops … it is evident from what I have written that I am unable
to accept all that Lord Wilberforce stated in that speech. Moreover, although a public
authority may be under a public duty, enforceable by mandamus, to give proper con-
sideration to the question whether it should exercise a power, this duty cannot be
regarded as afoundation for imposing a duty of care on the public authority in relation
to that power (ibid. at 31, emphasis added).
Essentially the same position was taken by Sopinka J. in dissent in Just, supra, note 3 at 1251.
It is notable that only three common law decisions have found a public authority liable in neg-
ligence when the court found the decision in question to be one of policy: Kamloops, supra, note
34, Laurentide Motels, supra, note 3, and, in the New Zealand courts Takaro Properties v. Rowling
[1986] 1 N.Z.L.R. 22 (H.C.), aff’d in part [1986] 1 N.Z.L.R. 51 (C.A.), rev’d [1987] 2 N.Z.L.R.
700 (P.C.) [hereinafter Takaro Properties]. In both Kamloops and Takaro Properties there were
serious allegations that the decision-maker had acted in bad faith. Those cases might, therefore, be
1992]
LIABILITY OF PUBLIC AUTHORITIES
3.
“Immunity” for Policy Decisions
The foregoing illustrates why the policy-operational distinction does not
answer the question of when government action is sufficiently like private
action so as to incur private law liability. The idea that the distinction is a test
for immunity further confuses the issue.’ “Immunity” suggests a special rule
that protects one from liability that would have been incurred but for the rule.
If the idea of a “policy decision” has any meaning, it must be that the impugned
decision is uniquely governmental in nature. If that is so, there can be no private
law duty of care: the government agency making a policy decision is acting like
a government and not like a person, so the rules that govern the civil relations
between persons do not apply as a matter of law, and cannot apply as a matter
of logic. It follows that a government rendering a policy decision is not
“immune” from private suit; rather, private law is simply irrelevant.
This is not a mere doctrinal quibble. If one thinks of policy decisions as
immune from private law suits, and operational decisions as not immune, it is
easy to make the mistake that operational decisions (whatever they are) fall
automatically into the realm of private law. This fallacious reasoning conceals
the basic question the distinction purports to answer: when are government
agencies sufficiently like persons that they should incur private law liability?
C. The Policy-Operational Distinction Since Anns
1.
England
As we have seen, the policy-operational test in Anns only confuses the
issue of when a public agency owes a duty of care arising out of public func-
tions. While in some cases it may be possible to state confidently that a partic-
ular decision is in the nature of policy or, conversely, that it is operational in
nature, in between there remains a vast grey area. The policy-operational dis-
tinction itself does not provide a criterion for dealing with cases in this grey
area. Furthermore, the consequences of making such a determination are un-
seen as cases granting compensation for ultra vires decisions rather than cases recognizing liability
for imprudence or negligence. In Review and Revision, supra, note 29 at 417, the author cites John-
son, supra, note 14 for the proposition that an authority cannot raise the defence that it was engag-
ing in a discretionary function if in fact no discretionary decision were made. On its face this seems
to support Lord Wilberforce’s position. The Court’s position must, however, be limited to the con-
text of that case. In Johnson, an adoption agency failed to notify the adoptive parents of the hom-
icidal tendencies of the sixteen year old child they were adopting. The agency’s fault, therefore,
was not simply a failure to make a policy decision: its failure was in acting (placing the child with
the adoptive parents) in a way that increased the risk to the parents. Once an agency acts in such
a way normal negligence principles can be applied.
41The idea that policy decisions are “immune” from tort liability is found in the judgments of
Lord Wilberforce in Anns, supra, note 4 and Mason J. in Sutherland Shire, ibid. at 34-35.
McGILL LAW JOURNAL
[Vol. 37
clear: some policy decisions may attract private law liability, and some opera-
tional functions may be discretionary and therefore immune from private law
evaluation.
It will not come as a surprise, then, that this aspect of Anns has virtually
disappeared from English jurisprudence. When English courts refer to the
“Ans test” they mean the test for the existence of a private law duty of care
based on foreseeability. The leading decisions of the House of Lords on the pri-
vate law liability of public authorities since Anns, Peabody Donation Fund v.
Lindsay Parkinson & Co. Ltd,42 Yuen Kun-yeu v. A-G Hong Kong43 and Hill v.
Chief Constable of West Yorkshire,” do not even mention the policy-operational
distinction.45
2.
Australia
Of the four judges who wrote decisions in Sutherland Shire,” the leading
Australian case in governmental liability in tort, only two mentioned the distinc-
tion. Gibbs C.J. found the distinction as employed by Lord Wilberforce in Anns
to be useful, but nothing in his decision turned on it. Mason J., whose decision
was cited with approval by Cory J. in Just,7 discussed the discretionary-
operational distinction quite fully. Like Lord Wilberforce, he proposed the
policy-operational distinction as an immunity rule alternative to the rule in Kent:
a decision not to act is immune if it is in the nature of policy, but not immune
if itis operational in nature. Whether liability will actually be imposed depends
on private law tort rules. According to Mason J., however, the policy-
operational distinction is not a general test for government liability in tort. It is
relevant only to the very narrow range of cases where it is alleged that the gov-
ernment agency made a conscious decision not to take certain actions. 4
1 More-
over, the policy-operational distinction does not provide a reason for imposing
liability in those limited cases. It is invoked only to rebut the suggestion that an
agency could never be liable for an affirmative decision not to act. As we will
see, government liability, like private law liability generally, arises from rela-
tionships of proximity, an issue wholly unconnected to the policy-operational
distinction.49
42(1985] A.C. 210, [1984] 3 All E.R. 529. (H.L.) [hereinafter Peabody cited to A.C.].
41(1988] A.C. 175, [1987] 2 All E.R. 705, [1987] 3 W.L.R. 776 (P.C. (H.K.)) [hereinafter Yeun
Kun-yeu cited to All E.R.].
44[1989] A.C. 53, [1988] 2 All E.R. 238 (H.L.).
45The Privy Council in Takaro Properties, supra, note 40 discussed the distinction only to reject
it. See below, text accompanying note 53.
46Supra, note 40.
471bid. at 34-35 cited in Just, supra, note 3 at 1241-42.
48Mason J. explicitly rejected the notion that an agency could be liable in private law for failing
to make a policy decision. Supra, note 40.
49See below, part H.
1992]
LIABILITY OF PUBLIC AUTHORITIES
The several decisions in Sutherland Shire have been very influential in
both England and Australia for the development of negligence law in general,
and governmental liability in tort in particular, but the policy-operational dis-
tinction has played almost no part in the jurisprudence of either country.
3.
New Zealand
In England the policy-operational distinction simply faded away after
Anns. In Australia it never really took hold. In New Zealand, however, the lower
courts applied the distinction with the same baffling results as the Canadian
courts.5 The appeals courts were therefore obliged to consider the merits of the
distinction carefully. They concluded that the distinction was a difficult one and
that the lower courts were misapplying it.
According to the Court of Appeal in Brown v. Heathcote County Council,
“the presence of a policy element is certainly relevant both to whether a duty
of care exists and to the question of breach, but only as one factor to be con-
sidered.”‘” In Takaro Properties Woodhouse P. stated:
All kinds of decisions which are made in the execution of policy will still involve
a considerable element of discretion. In the same way it may not be possible on
purely logical grounds to distinguish aspects of the practical development of pol-
icy from its broad formulation. So in my opinion it would be wrong to conclude
that these distinctions can have final significance in a case such as this. Nor do I
think there will be automatic exclusion of a duty of care simply because accent is
on policy rather than operational aspects of statutory decisions. 52
Because the matter had been raised in the lower courts, the Privy Council in
Takaro Properties thought it necessary to state its views regarding the role of the
policy-operational distinction. Lord Keith agreed with the Court of Appeal,
stating:
[The conclusion of Quilliam J. of the New Zealand High Court] was expressed as
follows:
The distinction between the policy and the operational areas can be
both fine and confusing. Various labels have been used instead of oper-
ational….
50See, for example, the High Court decision of Quilliam J. in Takaro Properties, supra, note 40
at [1986] 1 N.Z.L.R. 34-35. Quilliam J. stated that when a Minister of Finance makes a decision,
on the recommendations of a Cabinet Committee, based on irrelevant criteria, the “discretionary”
decision becomes “operational,” and therefore subject to private law duties. The gap in this logic
is obvious.
51Brown v. Heathcote County Council, [1982] N.Z.L.R. 584 at 618 (H.C.), aff’d for different rea-
sons [1986] 1 N.Z.L.R. 76 (C.A.), aff’d for different reasons [1987] 1 N.Z.L.R. 720 (P.C.). The
citation quoted here is from Cooke P.’s judgment at the Court of Appeal at 81 in. 34. At trial Hardie
Boys J. of the High Court declined to find that a drainage authority had a general duty to warn
home builders of potential flood damage because to do so would intrude on the authority’s policy-
making powers.
52Supra, note 40 at [1986] 1 N.Z.L.R. 59.
REVUE DE DROIT DE McGILL
[Vol. 37
Their Lordships feel considerable sympathy with Quilliam J.’s difficulty in
solving the problem by simple reference to this distinction. They are well aware
of the references in the literature to this distinction (which appears to have orig-
inated in the United States of America), and of the critical analysis to which it has
been subjected. They incline to the opinion, expressed in the literature, that this
distinction does not provide a touchstone of liability, but rather is expressive of the
need to exclude altogether those cases in which the decision under attack is …
unsuitable for judicial resolution, of which notable examples are discretionary
decisions on the allocation of scarce resources or the distribution of risks.5 3
In other words, whether a decision is by nature policy may assist in determining
whether the courts should decline to comment altogether, but it is of no assist-
ance in determining whether private law liability should be imposed.
D. The Supreme Court of Canada Trilogy and The Policy-Operational
Distinction
The reader will perhaps be surprised to find the Supreme Court of Canada
firmly wedded to the policy-operational distinction given the theoretical prob-
lems with and the movement in the Commonwealth away from the distinction.
This section will consider the Supreme Court’s recent application of the distinc-
tion and demonstrate the heights of confusion which the distinction has
produced.
1.
Laurentide Motels v. Beauport (City)
In Laurentide Motels5’ a fire department responded to a fire caused by a
guest of the plaintiff motel. The fire department arrived at the scene promptly,
but was unable to find the fire hydrants, which had become buried in the snow.
After the fire fighters found the hydrants they were unable to obtain water. The
motel sued on two grounds: (1) that the city was liable for the negligence of its
fire fighters at the scene; and (2) that the city was liable for failing to maintain
the fire hydrant system. The Supreme Court agreed on both counts. On the first
ground, the Court held that firefighter functions were operational in nature, and
therefore subject to private law duties. In so doing, the Court overruled the bulk
of authority.55 On the second ground, the Court inferred from the municipality’s
53Takaro Properties, ibid. at [1987] 2 N.Z.L.R. 709.
54Supra, note 3.
55The bulk of the authorities have held that municipalities owe no private law duties for negli-
gence in the operation of a fire service.
Quebec: Fournier v. Victoriaville (1918), 28 B.R. 216; Peletier v. Gatineau Pointe (1937), 76
C.S. 180; but see Point- Wau v. GauthierMfg, [1979] C.A. 77 (where a fire service is actually cre-
ated the chief is obliged to act according to basic fire-fighting principles).
Common Law Canada: Alexander v. London (1919), 15 O.W.N. 320; Sequin v. Hawkesbury,
[1955] O.R. 956, 5 D.L.R. 809 (C.A.); Soderberg v. Medstead, [1921] 1 W.W.R. 1130, 14 Sask.
L.R. 201, 59 D.L.R. 702 (C.A.); Stevens-Wilson v. Chatham, [1933] O.R. 305 at 326, (1933] 2
D.L.R. 407 (C.A.); Wing v. Moncton, [1940] 2 D.L.R. 740, 14 M.P.R. 515 (N.B.C.A.); but see
19921
LIABILITY OF PUBLIC AUTHORITIES
power to operate a waterworks a duty to maintain the system. It found further
that the city’s ad hoc inspections were inadequate and in breach of the inferred
statutory duty. This amounted to fault for the purposes of art. 1053 of the
C.C.L.C.
56
The Court was faced with three questions. First, are the boundaries of
municipal delictual liability in Quebec defined by public law or private law? In
other words, does one look to public law for the demarcation between the pri-
vate and public character of municipal corporations, or is the demarcation found
in private law?57 The second question, following from the first, was: what spe-
cific public law rule defines the limits of municipal delictual responsibility? The
third question was: once we know which municipal functions are subject to pri-
vate law, what are the precise civil law rules that govern municipal delictual lia-
bility for those functions?
a. The Policy-Operational Distinction
Beetz J. decided the first question by holding that the determination of the
dividing line between the public and private aspects of municipal liability is a
matter of public law. Accordingly, he then turned his attention to finding which
particular public law rules apply. He wrote:
Where the legislator confers a power upon a public authority, the conferral of
power is usually couched in terms of a discretion. The discretionary power is nec-
essary to allow the public authority latitude in which to make decisions that can
be categorized as policy decisions: decisions of a political nature for which the
authority should be accountable not before the courts but before the electorate or
the legislature. Anns, supra, and Kamloops, supra, indicate that the form such pol-
icy decisions may take varies, ranging from by-laws and resolutions to internal
directives, administrative decisions and even a discretion in the execution of activ-
Hesketh v. Toronto (1898), 25 O.A.R. 449 (no duty to fight fires, but once one takes on the duty
it must be followed through).
United States: Dalehite, supra, note 1 (referring to the assertion that liability could lie for neg-
ligence in fighting fires: “The Federal Torts Claims Act did not create a cause of action where none
existed before. … That cities, by maintaining fire fighting organizations, assume no liability for
their lapses, is much more securely entrenched.” (ibiaL at 43-44)); Moytka v. City of Amsterdam,
204 N.E.2d 635 (N.Y. 1965); but see Rayonier Inc. v. United States, 352 U.S. 315 (1957) (where
state law imposes liability on persons for failing to prevent the spread of fire, that liability will also
be imposed on federal fire fighters for failing to contain fires on federal lands).
Australia: Bennett & Wood Ltd v. Orange City Council (1967), 67 S.R. (N.S.W.) 426 (duties of
a fire authority “described” in a statute are “descriptive only”; cannot be the basis of a duty of
care).
56Art. 1053 states, “Every person capable of discerning right from wrong is responsible for the
damage caused by his fault to another, whether by positive act, imprudence, neglect or want of
skill.”57For discussion of the dual nature of municipal corporations, see Bowles v. Winnipeg (1919),
29 Man. R. 480,24 C.C.C. 327,45 D.L.R. 94 at 99 (K.B.); Pon Yin v. Edmonton (1918), 24 C.C.C.
327 at 333, 8 W.W.R. 809 (Alta K.B.); Welbridge Holdings, supra, note 16 at 968ff.
McGILL LAW JOURNAL
[Vol. 37
ities within the operational sphere. The only duty incumbent upon the authority in
the policy sphere is to make its decisions responsibly and in accordance with the
object of the Act which conferred the power However, the discretion conferred by
the legislator is not so broad as to exclude all liability for the authority’s actions.
Once the authority moves into the operational sphere of its power, i.e., the prac-
tical execution of its policy decision, the authority will be liable for damage
caused to an individual by its negligence.
The question now arises as to whether this common law rule is a rule of pub-
lic law or of private law. If it is a rule of private law, it cannot apply in Quebec….
In fact, I do not believe that the rule can be categorized as other than a rule of
public law.58
This passage is interesting because it emphasizes that the policy-operational dis-
tinction is a matter of public law, not private law. That being so, it cannot by
itself be a test for the existence of private law liability. Second, this enunciation
of the policy-operational test is consistent with the American sources: if a deci-
sion is in the nature of policy, it can be reviewed only according to public law.
On the other hand, merely operational decisions may be subject to the private
law, and in particular to arts 1053 et seq. Unfortunately, the test he then applied
did not conform to the test he enunciated. Beetz J. wrote:
The testimony of Armand Grenier, a superintendent of the city of Beauport
whose duties included ensuring that the fire hydrants were inspected, maintained
and cleared, shows that no real policy decision was ever taken with respect to the
inspection and repair of the fire hydrants. As a matter of practice, all hydrants
would be inspected during the summer to ensure that they were functional, but no
records were kept of those inspections. In winter, again as a matter of practice, all
fire hydrants would be cleared after each snowfall, but again, no records were kept
of these activities. There is no doubt that the municipality could have, as a matter
of policy, established some scheme of inspection and repair. However, failure to
make such a policy decision does not allow the municipality to escape the appli-
cation of the private law. Wilson J. wrote for the majority in Kamloops, supra,
that:
In my view, inaction for no reason or inaction for an improper reason
cannot be a policy decision taken in the bona fide exercise of
discretion.
Therefore, in absence of a policy decision to which the omission alleged to
have caused damage can be attributed, the inspection and repair of the fire
hydrants must be taken to be in the operational sphere, since they are the practical
execution of the municipality’s policy decision to establish the water system and
to allocate personnel and money to the maintenance of the system. Private law
standards therefore apply to the municipality’s conduct. 59
5SLaurentide Motels, supra, note 3 at 722-23 (emphasis added). This conclusion follows from
art. 356 C.C.LC.
591bid., note 3 at 726.
19921
LIABILITY OF PUBLIC AUTHORITIES
Observe that the municipality did have a practice of inspecting the fire
hydrants. The municipality’s alleged fault lay only in not enunciating an explicit
policy. One would have thought that according to Beetz J., if an authority does
not explicitly say it has decided not to perform its functions as a matter of policy,
whatever procedures it adopts will be deemed operational. By Beetz J.’s trans-
mutation, an unwise policy decision turns into an operational function, which is
then judged according to private law standards. According to the first passage
cited above, “[t]he only duty incumbent upon the authority in the policy sphere
is to make its decisions responsibly and in accordance with the object of the Act
which conferred the power.”‘ On these terms, the authority could be faulted
only if it was irresponsible. What, then, does “irresponsible” mean? If the dis-
tinction between the policy and the operational spheres has any meaning, an
“irresponsible” decision cannot simply be one that appears “unwise.” Yet Beetz
J.’s second formulation has done exactly that.
There is certainly no support for this proposition in the American, Austra-
lian or English decisions, nor does it follow from the nature of policy or discre-
tionary decisions. In fact, as it was originally understood, discretionary (or pol-
icy) and operational decisions were mutually exclusive categories. Perhaps
Beetz J. was thinking of Lord Wilberforce’s tentative suggestion that in appro-
priate circumstances a public authority might be held liable if it fails to consider
whether to exercise a statutory power.6′ Lord Wilberforce, however, did not say
that the failure to make an explicit policy decision rendered all that follows
operational in nature.
b. Firefighters’ Lack of Diligence
In addition to the question of whether the municipality had been negligent
by failing to inspect fire hydrants, there was an allegation that the firefighters
had lacked diligence in the techniques used once they arrived at the scene of the
fire. On that question Beetz J. wrote:
The characterization of the acts and omissions of the city of Beauport’s fire-
fighters poses no difficulties: they are clearly operational in nature. Doubtless
there exists a discretion in the execution of the firefighters’ operational activities.
But this discretion exists in the operational rather than the policy sphere, and the
exercise of this discretion is itself an operational decision. Here too, private law
standards apply.61A
Beetz J. asserted that the fighters’ activities were more operational than policy
in nature (despite the element of discretion), and that liability could be imposed
6Ibid. at 722.
61See above, text accompanying notes 31, 32. Lord Wilberforce suggested that when an authority
is negligent in considering whether to exercise a power it might be held liable to private individ-
uals. The basis for this proposition has never been explained.
61ASupra, note 3 at 727.
REVUE DE DROIT DE McGILL
[Vol. 37
as a result. If “policy” means planning while “operational” means doing, it does
not seem extravagant to assert that fighting fires is more operational in nature
than policy in nature. But the policy-operational distinction adds nothing to this
intuition. The fact that the assertion seems plausible does not make it more than
an assertion.
Neither Beetz I. nor L’Heureux-Dub6 J. ever considered whether the
municipality or the firefighters conducted themselves in the manner of a prudent
person similarly situated as required by art. 1053 C.C.L.C.62 Beetz J. adopted
L’Heureux-Dub6 J.’s reasoning on this question, and the latter inferred a duty
to fight fires prudently from the fact that the municipality had allocated funds
for the creation and maintenance of a waterworks system. As a result, the
policy-operational distinction became the source of the duty to compensate
rather than a preliminary determination of the appropriateness of applying pri-
vate law standards to public conduct. This decision is impossible to justify on
its own logic or to reconcile with the jurisprudence.
2.
Just v. British Columbia
In Just,63 the plaintiff’s car was struck by a boulder that had been dislodged
by natural forces from the slopes above the highway on which he was travel-
ling.’ The plaintiff was severely injured and his daughter was killed.
McLachlin J. (as she then was) found for the province on the ground that
the decisions on how to inspect the road-side slopes, and on what to do about
potential hazards, fell “within the area of policy, and [could not] be reviewed by
this court.”‘6 She did not go on to consider what standards would be reasonable,
or whether the province had met those standards because she considered the
province’s decisions on road maintenance to be in the nature of policy, and
hence incapable of giving rise to a private law duty of care. A unanimous Brit-
ish Columbia Court of Appeal agreed.’
The Supreme Court of Canada reversed.67 Cory J. found first that the prov-
ince was under a duty of care to maintain the highways reasonably.” Second,
he found that these duties were operational (not policy) in nature and therefore
that the court was not barred from reviewing the adequacy of the maintenance
62See above, subsection II.C.2.
63Supra, note 3.
64″In all probability the fall was caused by the action of tree roots working into the natural dis-
continuities of the rock. The resultant cracks were expanded by successive freeze-thaw cycles.
Heavy wet snow settling on the tree, perhaps aided by high winds, caused it to act as a lever on
the rock” ((1985), 64 B.C.L.R. 349 at 350, [1985] 5 W.W.R. 570 (S.C.) [cited to B.C.L.R.]).
651bid. at 355.
6610 B.C.L.R. (2d) 223, [1987] 2 W.W.R. 231 [cited to B.C.L.R.I.
67Per Cory J., Sopinka J. dissenting.
68Supra, note 3 at 1236.
1992]
LIABILITY OF PUBLIC AUTHORITIES
program. The case was returned for a new trial on that issue as the lower courts
had not made any findings as to the adequacy of the province’s maintenance
program.
a.
Immunity for Policy Decisions
Cory J. began his analysis by setting out his views on how the tort liability
of government agencies relates to tort liability generally. He wrote:
It may be convenient at this stage to summarize what I consider to be the
principles applicable and the manner of proceeding in cases of this kind. As a gen-
eral rule, the traditional tort law duty of care will apply to a government agency
in the same way that it will apply to an individual. In determining whether a duty
of care exists, the first question to be resolved is whether the parties are in a rela-
tionship of sufficient proximity to warrant the imposition of such a duty. In the
case of a government agency, exemption from this imposition of duty may occur
as a result of an explicit statutory exemption. Alternatively, the exemption may
arise as a result of the nature of the decision made by the government agency. That
is, a government agency will be exempt from the imposition of a duty of care in
situations which arise from its pure policy decisions.69
It is evident that Cory J. regarded the policy-operational distinction as a test for
immunity from liability that would otherwise be imposed, in the way discussed
and criticized above. The first three sentences are uncontroversial. If a govern-
ment agency is in a proximate relationship with a specific person a duty of care
towards that person may arise. However, he goes on to say that if the plaintiff
and the agency were in a relationship of proximity, the agency may be exempted
from liability if the decision in question is a “pure policy decision.”
With respect, this view misconceives the very question the policy-
operational distinction is supposed to answer. We say that governments ought
to bear the same civil responsibility as persons, but the fact is that governments
are different from persons in important respects.70 The American discretionary
function exception was intended to identify those government decisions which
are political and legislative and therefore uniquely governmental. The civil
duties which arise from relationships between persons are simply inapplicable
to a relationship between the government and the governed. Similarly, Lord
Wilberforce saw the policy-operational distinction as a way to identify when the
government ceases to act at large towards the general public, and begins to act
in a specific way towards a specific person so that it could be said that a rela-
tionship of proximity had arisen. If a decision is “policy” and not “operational,”
69 bi&L at 1244-45, per Cory J. This schema is borne out in the structure of his judgment. After
laying out the facts he decided that the province was under a private law duty of care in respect
of the highway inspections; then he turned to the policy-operational distinction to see whether the
province should be granted immunity.
70See above, text accompanying section I.B.
McGILL LAW JOURNAL
[Vol. 37
then by extension the government cannot be in a relationship of proximity with
the plaintiff. According to both the American legislation and Lord Wilberforce
in Anns, a policy decision is generally outside the scope of private law.
Cory J., in contrast, assumed that government activities are prima facie
subject to private law rules. He decided that all government functions are capa-
ble of attracting private law liability, although the government will be immune
from the liability that would otherwise be enforced if the operative decision was
in the nature of policy. This is the converse of the view of Lord Wilberforce,
who held in Anns that government functions were presumptively governed by
public law, although there might be room for private law liability “over and
above, or perhaps alongside” the public law duties. In the long run, it would
probably not be very important where the presumption lay, so long as there was
a test which would unambiguously decide when the presumption should hold
and when it should be set aside. The policy-operational distinction is not such
a test. As applied by the Supreme Court of Canada it simply confirms the pre-
sumption that government functions are subject to private law liability.
b. Applying the Policy-Operational Distinction as a Test
After approving the use of the policy-operational test generally,7 Cory J.
applied it to the facts in Just. The difficulties he encountered illustrate the ambi-
guity of the test.
After reviewing the jurisprudence, Cory J. summarized the policy-
operational distinction as follows:
In determining what constitutes such a policy decision it should be borne in
mind that such decisions are generally made by persons of a high level of authority
in the agency, but may also properly be made by persons of a lower level of
authority. The characterization of such a decision rests on the nature of the deci-
sion and not on the identity of the actors. As a general rule, decisions concerning
budgetary allotments for departments or government agencies will be classified as
policy decisions. Further, it must be recalled that a policy decision is open to chal-
lenge if it is not made in the bonafide exercise of discretion. If after due consid-
eration it is found that a duty of care is owed by the government agency and no
exemption by way of statute or policy decision-making is found, a traditional tort
analysis ensues and the issue of standard of care required of the government must
next be considered.72
71But supra, note 3 at 1239 he cited Blessing v. U.S., 447 F.Supp. 1160 (E.D. Penn 1978) [here-
inafter Blessing] which cautions against the temptation to subject broad administrative programs
to the crucible of tort law; at 1243 he cited Indian Towing, supra, note 11, where it was made clear
that the discretionary function exception does not say anything about the private law duties of care;
and at 1243 he cited U.S. v. Varig Airlines, 467 U.S. 797 (1984) [hereinafter Varig Airlines] which
made it clear that inspection systems are matters of discretion, and cannot be subject to a “reason-
able person” test.
72Supra, note 3 at 1245.
1992]
LIABILITY OF PUBLIC AUTHORITIES
Thus, policy decisions may be made at high or low levels. The test turns on the
“nature of the decision” but there is no discussion about the “nature” of a policy
decision.73 Generally, budgetary decisions are policy in nature, but then again,
even those may be challenged if they are not made in the bonafide exercise of
discretion. What are the boundaries of a bona fide decision? None are set out,
but two examples are given: where a municipality inspects its roads only every
five years74 and where a lighthouse service inspects its lighthouses only every
two years.75 Yet the implication of Just is that it is open to a court to find that
an elaborate inspection and maintenance program could be attacked as being
beyond the bonafide exercise of discretion.76 In short, policy is what the court
says it is. Once again, the policy-operational distinction is “a way of stating,
rather than arriving at, a result.”‘
After considering various examples of policy and operational decisions
Cory J. wrote:
Thus, a true policy decision may be made at a lower level provided that the
government agency establishes that it was a reasonable decision in light of the sur-
rounding circumstances.
78
First, Cory J. had earlier stated that no liability could attach to actions following
from “pure policy decisions.” Now he seems to be saying that the agency will
be exempt only so long as the “true” policy decision is reasonable. Second,
there is no suggestion in any of the previous jurisprudence that the onus is on
the government to show that even its operational decisions were reasonable. The
onus, as in private law generally, is on the plaintiff to show that the decision
taken was unreasonable.
He then turned to the question of whether the decisions concerning the
inspection system in the case at hand were in the nature of policy or were oper-
ational. He wrote:79
The manner and quality of an inspection system is clearly part of the oper-
ational aspect of a government activity and falls to be assessed in consideration of
the standard of care issue.8 0
73To be fair, Cory J. does identify criteria, ibid. at 1245. However, as my discussion of standard
of care demonstrates above, subsection II.C.3.a., these criteria break down at the level of
application.
741bid. at 1238.
75lbiL at 1243. Cory J. here states that in Indian Towing the Coast Guard was obliged to conduct
the inspections carefully because it had decided to inspect. That is far from the ratio of that case.
See text accompanying note 139ff., infra.
76The role of budgetary allotments in the liability of government agencies is explored more fully
below, subsection II.C.3.a.
“Jaffe, supra, note 6.
78Just, supra, note 3 at 1243 (emphasis added).
791bid. at 1245.
801t might be argued that he meant only to say that the inspections themselves were operational
and subject to assessment. However, from the aspects of the operation listed in the subsequent pas-
REVUE DE DROIT DE McGILL
[Vol. 37
Why this is “clearly” so is not explained. In fact, one source he had cited earlier,
Varig Airlines,$’ held unequivocally that the design of an inspection system is
unreviewable policy. Barratt v. North Vancouver,82 a decision of Martland J. of
the Supreme Court of Canada, stated the same thing. In Barratt, a cyclist sued
the city for failing to detect and repair potholes. A unanimous bench found that:
[A municipality’s] method of exercising its power [of maintaining municipal
roads] was a matter of policy to be determined by the Municipality itself. If, in the
implementation of policy its servants acted negligently, causing damage, liability
could arise, but the Municipality cannot be held negligent because it formulated
one policy of operation rather than another.s2A
Cory J. did not distinguish the case. He simply stated that Martland J.’s view
was obiter, and that it went “too far.” (Sopinka J. disagreed, saying it was on
“all fours”).” Cory J. then sought to show that all road inspection decisions are
in principle assessable by private law standards by offering the example of a
municipality which decided to inspect its roads only once every five years.’
This is a non sequitur: the possibility that policy-makers will make unrea-
sonable discretionary decisions does not show that those decisions are really
operational in nature. Moreover, the example, which amounts to a total abdica-
tion of responsibility, is irrelevant to the Just case. There was no suggestion that
the province had failed entirely to conduct inspections; rather, an elaborate pro-
gram was in place.
c. The Dissent of Justice Sopinka
Sopinka J.’s dissent was based on three propositions: first, that the “explo-
sion” of tort liability generally ought to be defused; second, that where an
agency is acting under a power, rather than a duty it cannot be held liable to per-
sons; and third, that the decision in Barratt was correct insofar as it held that
decisions concerning the maintenance of public roads are in the nature of policy.
The second and third propositions bear no logical relation to each other or to
Sopinka J.’s major thesis, that the expansion of tort liability ought to be
reversed.
The second proposition is contained in the following passages:
The starting point for the application of the Anns principle is that there is no
statutory duty in favour of the plaintiff to do the thing the lack of which is alleged
to have caused the injury to the plaintiff.8 5
sages it seems that Cory J. intends to permit a challenge to the system as a whole. I must admit
that that is not absolutely clear. That in itself is a problem with the decision.
81Supra, note 71.
82[1980] 2 S.C.R. 418, 114 D.L.R. (3d) 577, 27 B.C.L.R. 182 [hereinafter Barratt].
SZAIbid. at 428.
83Just, supra, note 3 at 1238, Cory J. and at 1257-58, Sopinka J.
84Ibid. at 1238.
85Ibid. at 1250.
LIABILITY OF PUBLIC AUTHORITIES
1992]
And later:
If, as here, the statute creates no duty to inspect at all, but simply confers a
power to do so, it follows logically that a decision to inspect and the extent and
manner thereof are all discretionary powers of the authority.8 6
This reasoning appears to resurrect the rule in Kent’ that omissions can never
give rise to civil liability unless there is a statutory duty to act. With respect, the
power-duty dichotomy is no more helpful than the policy-operational distinc-
tion. It has no bearing whatsoever on the question of when a government is suf-
ficiently like a person that private law rules can apply.
The core of Sopinka J.’s third proposition is found in the following
passage:
My colleague’s [Cory J.’s] reasons are based essentially on an attack on the
policy of the respondent with respect to the extent and the manner of the inspec-
tion program. In my opinion, absent evidence that a policy was adopted for some
ulterior motive and not for a municipal purpose, it is not open to a litigant to attack
it, nor is it appropriate for a court to pass upon it. 8
Citing Barratt,”9 he concluded that decisions concerning the form and extent of
road inspection programs are policy in nature and therefore exempt from
review. This conclusion seems correct. However, his reasoning in this case will
not be of much assistance in future cases because it relies on the policy-
operational distinction. As I have suggested, finding that a decision is by nature
policy or operational is invariably a simple assertion.
This leaves only the first proposition that courts ought to be wary of admit-
ting new forms of liability because tort liability has gotten out of hand. Citing
a number of studies and academic articles, Sopinka J. observed that insurance-
conscious judges have amended the classical function of tort law, shifting the
loss suffered by the plaintiff to the defendant who caused the loss, by adding a
loss-spreading function, which compensates those who suffer harm irrespective
of the fault of the plaintiff or the defendant. Sopinka J. hinted that this devel-
opment is detrimental, and that recognizing liability in a case like Just would
exacerbate this undesirable trend.
I cannot say that loss-spreading has no place. Indeed, there are powerful
arguments that a progressive society should alleviate distress where it is capable
of doing so, and loss-spreading in tort law is a reflection of that value. My
objection is that by employing the empty rhetorical device of the policy-
operational distinction, the Court prevents principled discussion of the merits of
86Ibid. at 1253.
87Supra, note 18.
88Supra, note 3 at 1251.
89Supra, note 82.
McGILL LAW JOURNAL
[Vol. 37
the loss-spreading approach in theory and the limits to loss-spreading in partic-
ular cases.’
d. Conclusion
In older cases like Kent, the possibility of government liability in tort was
dismissed virtually out of hand because it was thought that the fundamental dif-
ferences between public agencies and persons was so great that private law lia-
bility was simply inapplicable. Sopinka J. adopted this view, although in a more
moderate way.
In Cory J.’s judgment the presumption was reversed. Cory J. presumed that
government activity is like private activity unless the converse is proven. The
converse might be proven if the government, not the plaintiff, can show that the
operative decision was in the nature of policy. However, the ambiguities of the
policy-operational distinction make it virtually impossible to prove that a deci-
sion was a “pure” policy decision. Even if the government convinces the court
that the decision was in the nature of “pure policy” there remains the possibility
that it will have to show that the decision was reasonable. In short, the presump-
tion that public activity can generate private law liability is virtually
irrebuttable.
Neither extreme view is helpful. A more balanced position is that of Lord
W’ilberforce in Anns, who began with the assumption that government activity
is essentially public in nature, but then recognized that in certain circumstances
a government agent or agency could be in a relationship of proximity with a per-
son such that a duty of care could arise. This view focuses attention on the par-
ticularities of the relationship between the government agent or agency and the
plaintiff, always keeping in mind the essential public nature of the agency.
3.
Rothfield v. Manolakos
In Rothfield the plaintiffs wished to build up and level their sloping back-
yard. They hired a contractor to build a retaining wall, and on their own initia-
tive discovered that they required a building permit. Owner-builders were nor-
mally obliged to submit engineer’s diagrams with their application for a
building permit, but the by-laws also provided the chief inspector with a discre-
tion to dispense with this requirement for less ambitious projects. The city did
so in this case, on the understanding that the owner or contractor would notify
the city at each stage of construction so that the work could be inspected visu-
ally. This understanding merely duplicated the requirements of the by-law.
90See in regard to possible loss spreading approaches: D. Cohen & J.C. Smith, “Entitlement and
the Body Politic: Rethinking Negligence in Public Law” (1986) 64 Can. Bar Rev. 1; R.A. Mac-
donald, “Jurisdiction, Illegality and Fault: An Unholy Trinity” (1985) 16 R.G.D. 69.
1992]
LIABILITY OF PUBLIC AUTHORITIES
Despite these twin obligations the owners failed to summon the city during the
important first stages of construction. When the inspector was finally sum-
moned he was unable to examine the foundations properly because they had
already been covered over. He permitted work to continue, but again on the
understanding that he would be summoned at each stage. Work progressed
almost to completion until a crack appeared in the wall. The inspector returned,
but he was unable to evaluate the seriousness of the problem. He ordered work
halted for a period during which the wall would be watched to see if further evi-
dence of structural weakness manifested itself. The wall did not deteriorate, so
he granted permission to complete the backfilling. The wall collapsed a short
time later.
Professional engineers hired by the owner-builders in contemplation of lit-
igation determined that the wall had been designed and constructed inade-
quately. The owner-builders sued the contractors, the city, and its inspectors,
alleging that they had been negligent in not ensuring that the construction con-
formed to the by-laws. The trial court found the contractors 60% liable (for
faulty construction and design) and the city 40% liable (for negligently issuing
the building permit and failing to conduct adequate inspections).
The Supreme Court of Canada found that the city, by making a policy deci-
sion to inspect houses, came under a duty of care to owner-builders to ensure
that their construction projects conformed to the by-laws. The Court also found
that the owner-builders had been contributorily negligent in failing to notify the
city, so their damages were reduced accordingly. The portion of liability the trial
judge had allocated to the city was re-apportioned, 70% to the city and 30% to
the owner-builders.
The decisions in Rothfield did not rely on the policy-operational distinction
to the same extent as the other cases did. However, the plurality judgment of
La Forest J. did mention it.9 His decision reflects the danger of relying on the
policy-operational distinction, and the analytical poverty of the foreseeability
test for establishing duties of care.
La Forest J.’s basic position is set out in the following passage:
By application of the test formulated by Lord Wilberforce in Anns, and adopted
by this court in Kamloops, the city, once it made the policy decision to inspect
building plans and construction, owed a duty of care to all who it is reasonable to
conclude might be injured by the negligent exercise of those powers. This duty is,
of course, subject to such limitations as may arise from statutes bearing on the
powers of the building inspector.92
9’Only two of the remaining six judges agreed with him, but his judgment, with three signato-
ries, became the basis of the order because the other two had only two judges each.
92Supra, note 3 at 1266-67.
REVUE DE DROIT DE McGILL
[Vol. 37
La Forest J. implied that once one made a policy decision, all subsequent
decisions following from the first decision are necessarily non-policy (opera-
tional). in nature. But there is nothing in the nature of either policy or discretion
that would support this conclusion.93 For example, it is conceivable that a policy
decision (passing a by-law) could establish a review process that itself grants
broad, non-reviewable discretion to a functionary. In fact, the by-law in this case
gave the city’s chief inspector considerable discretion in the making of a
number of decisions. Thus, the fact that a “policy” decision was taken to inspect
buildings tells us nothing whatsoever about the discretionary nature of subse-
quent implementation decisions, or whether courts ought to refrain from
second-guessing those decisions.94
To reiterate, the policy-operational distinction is supposed to distinguish
the government acting as a government from the government acting as a person.
Of the three areas of government activity that gave rise to the trilogy of cases,
the house inspections in Rothfield appear at first blush to bear the greatest
resemblance to an activity that could be performed by a person. Once one looks
at the matter more closely, as we will below, it becomes clear that there are fun-
damental differences between the relationship between a municipal inspector
and a home-builder on the one hand, and the relationship between a private
industry design professional and his or her client on the other. The policy-
operational distinction, especially as misapplied by La Forest J., is incapable of
bringing these differences into relief.
E. Policy-Operational Distinction: Conclusion
The objective of the law regulating the private law responsibility of public
agencies is to hold the government civilly responsible as if it were a person,
insofar as that is possible. Under American law, which initially focused on polit-
ical or legislative aspects of government, “discretion” was nominated as the
touchstone of truly governmental activity. The discretionary function exception,
however, does not address the principal question. First, “discretion” is a hope-
lessly vague concept. Second, there is nothing in the concept of discretion that
is uniquely governmental. Third, and most important, civil responsibility is
about relationships among persons in society. The concept of discretion takes
account only of the agency’s actions and ignores the plaintiff’s relationship with
the government. Discretion cannot tell us when the government is acting like a
person because, as it is aimed at the government only, it does not take into
account how persons interact.
The discretionary-operational distinction was mentioned by Lord Wilber-
force in Anns. That was a mistake. It was not necessary for Lord Wilberforce’s
93See the discussion of discretion, above subsection I.B.1.
94This seems to be Sopinka J.’s criticism of the majority in Just, supra, note 3 at 1251.
19921
LIABILITY OF PUBLIC AUTHORITIES
ill1
argument that the statutory character of public agencies does not automatically
preclude private law responsibility. Moreover, while discussing the distinction,
Lord Wilberforce confused it further. Fortunately, English, Australian and New
Zealand courts have ignored the policy-operational distinction or deliberately
set it aside.
The three Supreme Court decisions rely on the policy-operational distinc-
tion to varying degrees. In all three cases the analysis is circular and unhelpful.
The judges simply assert that the government should compensate the victims,
and their analysis confirms that assertion.
As we will see, this hazy reasoning is perpetuated by the foreseeability the-
ory of the existence of duties of care. The Anns-based foreseeability theory does
not and cannot relate the hypothetical foresight of a government agency to the
hypothetical foresight of an individual. But when one uses an alternative theory
for the existence of private law duties, focusing on the nature of the relationship
between the parties, as suggested by recent decisions of the House of Lords and
the Australian High Court, one’s mind turns naturally to the question of whether
the government acted towards the plaintiff in a manner analogous to relations
between persons. The fundamental question that the policy-operational question
seeks, but fails, to answer will therefore be addressed.
II. Private Law Duties of Care
The policy-operational distinction is, at best, a test for deciding whether
government activities are sufficiently like the activities of a person to render the
government liable “in the same manner and to the same extent as a private indi-
vidual under like circumstances.”’95 This does not answer the further question of
whether a person would owe a duty of care in the circumstances alleged by the
plaintiff. In this section I leave the policy-operational distinction aside, and turn
to consider the private law rules which apply to cases like the three Supreme
Court cases reviewed in this article.
A. Duties of Care
1.
The Anns Foreseeability Test
What then is the test for the existence of a common law (private law) duty
of care? One of the most controversial passages in the modem law of torts is
found in Lord Wfilberforce’s speech in Anns:
Through the trilogy of cases in this House – Donoghue v. Stevenson, Hedley
Byrne & Co. Ltd. v. Heller & Partners Ltd., and Dorset Yacht Co. Ltd. v. Home
Office, the position has now been reached that in order to establish that a duty of
95From the U.S. Tort Claims Act, supra, text accompanying note 11.
McGILL LAW JOURNAL
[Vol. 37
care arises in a particular situation, it is not necessary to bring the facts of that sit-
uation within those of previous situations in which a duty of care has been held
to exist. Rather the question has to be approached in two stages. First one has to
ask whether, as between the alleged wrongdoer and the person who has suffered
damage there is a sufficient relationship of proximity or neighbourhood such that,
in the reasonable contemplation of the former, carelessness on his part may be
likely to cause damage to the latter –
in which case a prima facie duty of care
arises. Secondly, if the first question is answered affirmatively, it is necessary to
consider whether there are any considerations which ought to negative, or to
reduce or limit the scope of the duty or the class of person to whom it is owed or
the damages to which a breach of it may give rise.96
This passage has been much followed and much doubted. The fault is not in the
passage itself but in the way it has been interpreted. Lord Wilberforce set out
a complex, perhaps inevitably complex formula, but other judges have col-
lapsed it into a single rule: the so-called foreseeability test.
Lord Wilberforce’s formula has two stages. The first stage focuses on the
specific parties and the circumstances that brought them to court. The second
stage introduces broader considerations not limited to the circumstances of the
specific parties, sometimes called “policy” considerations.’
The first stage has been the most seriously misunderstood. The test for a
primafacie duty of care in Anns involves three considerations: whether there is
a “relationship of proximity or neighbourhood” between the parties; whether
injury was “within the reasonable contemplation” of the defendant; and whether
“carelessness on [the part of the defendant] may be likely to cause damage” to
the plaintiff. Lord Wilberforce did not set out in any detail how these elements,
proximity, foreseeability, and causation, relate to one another conceptually, or,
on a more practical level, how a judge would apply them to real cases.98 In the
hands of other judges, however, “reasonable contemplation,” re-named “fore-
seeability,” emerged as the pre-eminent consideration. The concept of proximity
was swallowed up by foreseeability, so that foreseeability of injury automati-
cally created proximity, which in turn created a duty of care. This was perhaps
inevitable, given Lord Atkin’s now-sacred speech in Donoghue v. Stevenson.99
But, as Lord Wilberforce wrote soon after Anns in McLoughlin v. O’Brian:
96Supra, note 4 at 751-52.
97This “policy” is judicial policy. It is not to be confused with executive discretion or policy in
the context of the policy-operational distinction.
98Recall that Anns was an appeal on a preliminary point of law rather than on a final determi-
nation of the 1is between the parties.
99Donoghue v. Stevenson, [1932] A.C. 562, [1932] All E.R. I at 19 (H.L.). Lord Atkin’s judg-
ment is criticized in Sutherland Shire, supra, note 40 at 41. Brennan J. commented on Lord Atkin’s
biblical allusion and the neighbourhood principle:
If foreseeability of injury were the exhaustive criterion of a duty to act to prevent injury
occurring, the “neighbour” of the law would include not only the biblical samaritan but
1992]
LIABILITY OF PUBLIC AUTHORITIES
Foreseeability, which involves a hypothetical person, looking with hindsight at an
event which has occurred, is a formula adopted by English law, not merely for
defining, but also for limiting the persons to whom a duty may be owed…. It is
not merely an issue of fact to be left to be found as such. When it is said to result
in a duty of care being owed to a person or a class, the statement that there is a
“duty of care” denotes a conclusion into the forming of which considerations of
policy have entered. That foreseeability alone does not of itself, and automatically,
lead to a duty of care is, I think, clear…. [As Lord Reid stated,] A defender is not
liable for a consequence of a kind which is not foreseeable. But it does not follow
that he is liable for every consequence which a reasonable man could foresee.10
The Supreme Court of Canada is the last adherent to the foreseeability the-
sis. In recent years Commonwealth courts have considered anew the concept of
duty of care. In Peabody Lord Keith of Kinkel observed that “[tihere has been
a tendency in recent years to treat [the passages from Anns reproduced above]
as being themselves of a definitive character. This is a temptation that should
be resisted.’ ‘.. He suggested that, rather than becoming preoccupied with fore-
seeability, judges should take into account “all the circumstances of the case.”
Turning too quickly to the question of foreseeability upsets the logical order in
which the various considerations ought to be taken into account. There must be
a relationship of proximity, in Lord Atkin’s sense, before any duty of care can
arise.”r2
The problem with Lord Keith’s formulation is that saying we must take all
the circumstances into account does not help a great deal. Moreover, proximity
“in Lord Atkin’s sense” may, it seems, be created by foreseeability alone. He
therefore added that “in determining whether or not a duty of care of particular
scope was incumbent upon a defendant it is material to take into consideration
also the priest and the levite who passed by the injured man. Windeyer J. commented
on their relationship… :
He obviously was a person whom they had in contemplation and who was
closely and directly affected by their action. Yet the common law does not
require a man to act as the Samaritan did… . The law casts no duty upon
a man to go to the aid of another who is in peril or distress, not caused by
him.
If foreseeability of injury were the exhaustive criterion of a duty to act
to prevent the occurrence of that injury, legal duty would be coterminous
with moral obligation….
See also J.C. Smith & P. Bums, “Donoghue v. Stevenson – The Not So Golden Anniversary”
(1983) 46 Mod. L. Rev. 147; P.P. Craig, “Negligent Misstatements, Negligent Acts and Economic
Loss” (1976) 92 L.Q. Rev. 213; I.N. Duncan Wallace, “From Babylon to Babel, or a New Path for
Negligence? (1977) 93 L.Q. Rev. 16 at 17, 20; S. Todd, “The Negligence Liability of Public
Authorities: Divergence in the Common Law” (1986) 102 L.Q. Rev. 370 at 376; R.F.V. Heuston,
“Donoghue v. Stevenson in Retrospect” (1957) 20 Mod. L. Rev. 1.
100[1983] A.C. 410, [1982] 2 All E.R. 298 at 303 (H.L.) [hereinafter McLoughlin] (emphasis
added).
10’Supra, note 42 at 240.
“121bid.
REVUE DE DROIT DE McGILL
[Vol. 37
whether it is just and reasonable that it should be so.’ ‘ “03 This apparently bound-
less discretion was clearly unsatisfactory, and Lord Keith did not repeat the test
in such broad terms when he had the opportunity soon afterward.”4 Like the
“Anns test,” however, the “just and reasonable test” soon took on a life of its
own, far beyond what its author had intended.”5
In Australia, the strong foreseeability thesis has been authoritatively repu-
diated at least since the decision in Jaensh v. Coffey.”0 6 In Sutherland Shire,”0′
the High Court unanimously rejected the proposition that foreseeability is the
paramount consideration in determining the existence of duties of care. The sev-
eral judgments advocated various alternative tests. Brennan J., whose judgment
found favour with the House of Lords in Caparo Industries,’ suggested a
frankly contextual approach. He stated:
It is preferable, in my view, that the law should develop novel categories of neg-
ligence incrementally and by analogy with established categories, rather than by
a massive extension of a prima facie duty of care restrained only by indefinable
“considerations which ought to negative, or to reduce or limit the scope of the duty
or the class of person to whom it is owed.”” 9
As with Lord Keith in Peabody, Brennan J. would oblige judges to look closely
at the particular facts of each case. However, rather than permitting judges the
discretion to decide cases based on what is “just and reasonable in the circum-
stances,” he would have them justify new duties of care by explaining how the
new situation is similar to ones where duties of care have already been accepted.
This may appear to be a return to the piecemeal approach that Lord Atkin
sought to avoid in Donoghue v. Stevenson, but that is in fact how trial courts
now work. At the appellate level (and with trial judges who have the inclination
and the time to consider cases more thoroughly) judges will be obliged to con-
sider explicitly the nature of the relationship between the parties, whetl~er that
type of relationship makes one party civilly responsible for the welfare of the
other, and the extent and limits of that responsibility. An approach that forces
1031bid. at 241 (emphasis added).
104n Yuen Kun-yeu, supra, note 43 at 712ff.
105See, for example, Norwich City Council v. Harvey, [1989] A.C. 828 (C.A.); Pacific Associates
v. Baxter, [1989] 2 All E.R. 159 (C.A.) where the English Court of Appeal applied Peabody as if
it introduced a length-of-the-Chancellor’s foot test for the existence of a duty of care.
06(1984), 54 A.L.R. 417 at 443 (H.C.): “It is not and never has been the common law that the
reasonable foreseeability of risk of injury to another automatically means that there is a duty to
take care with regard to that risk of injury…. Although Jaensh v. Coffey and Sutherland Shire both
reject the foreseeability thesis of civil responsibility, they differ greatly on how proximity is actu-
ally established. See Note, “The Concepts of ‘Proximity’ and ‘Reasonable Foreseeability”‘ (1986)
60 A.L.J. 4.
’07Supra, note 40 at 41.
10s[1990] 1 All E.R. 568 (H.L.), [1990] 2 A.C. 605 [hereinafter Caparo Industries cited to All
E.R.].
l0gSupra, note 40 at 43.
1992]
LIABILITY OF PUBLIC AUTHORITIES
explicit analysis of the particularities of human relationships cannot help but
come closer to the objectives of civil responsibility than one based on the ret-
roactive prophesies of a “hypothetical person looking with hindsight.”
As stated, Brennan J.’s judgment was adopted by several Law Lords in
Caparo Industries. Lord Bridge’s speech may be regarded as the lead judgment.
After canvassing the line of jurisprudence surrounding Peabody he concluded,
as Lord Keith did in the latter case, that proximity (or neighbourhood) and what
is “just and reasonable in the circumstances” are factors which ought to be given
at least equal weight with foreseeability as foundations for duties of care.’
However, he went somewhat further than Lord Keith. He found that concepts
like foreseeability, proximity, and what is “just and reasonable” are little more
than labels which do not elucidate the real issues, and may in fact provide con-
cealment for the lazy or the intellectually dishonest. Like Brennan J., he stated
that the law should turn back “in the direction of attaching greater significance
to the more traditional categorisation of distinct and recognisable situations as
guides to the existence, the scope, and the limits of the varied duties that the law
imposes. ‘ “OA A great deal of excitement was created by the recent decision of
the House of Lords in Murphy v. Brentwood District Council,”‘ but in my view
Caparo Industries is a far more significant case. Murphy overruled Anns” inso-
far as Anns implied that pure economic loss is recoverable in tort even absent
a special relationship like that in Hedley Byrne”3 or Ross v. Caunters.”‘ Caparo
Industries reformulated the basis of tortious civil responsibility generally.
Brennan J. and Lord Bridge did not offer a specific test so much as a course
of judicial inquiry. What is clear, however, is the following: first, one must con-
cede that there is not, nor could there be, a single overarching principle for
establishing duties of care in negligence; rather, it will be necessary to recognize
various categories of duties. It might be objected that this returns us to the “bad
old days” before Donoghue v. Stevenson when the plaintiff had to prove that his
or her injury fit within one of a limited number of recognized classes of victim.
This should not arouse concern. What may have been objectionable before
Donoghue v. Stevenson was not that negligence duties were categorized, but that
the court was reluctant to admit new tort duties that did not fit into one of the
recognized categories. Caparo Industries follows the spirit of Lord MacMillan’s
dictum in Donoghue v. Stevenson that, “[t]he categories of negligence are never
11 Supra, note 108 at 574.
IlOAlbid.
111[1990] 2 All E.R. 908 (H.L.), [1990] 3 W.L.R. 414.
112It is arguable that just as Lord Wilberforce’s statements on foreseeability were applied more
broadly than he had intended, so too have his views on economic loss been exaggerated. See the
comments by Taylor J. on economic loss in Kamahap Enterprises Ltd v. Chu’s Central Market Ltd,
[1990] 1 W.W.R. 632, 40 B.C.L.R. (2d) 288 at 290-94.
‘”Hedley Byrne & Co. v. Heller & Partners Ltd, [1964] A.C. 465, [1963] 2 All E.R. 575 (H.L.).
14[1980] Ch. 297, [1979] 3 All E.R. 580.
McGILL LAW JOURNAL
[Vol. 37
closed.””1 5 The stress is on development of new categories of duties as opposed
to the rigid adherence to old categories. 6 In short, recognizing that we must
develop categories of duties is not so much a choice, but an imperative rising
out of the variability of human relations. Second, in establishing categories of
duties we should work by analogy from recognized duties. Third, these analo-
gies should focus on, among other things, the particularities of the relationship
between the parties and the type of injury actually suffered. This “neo-
traditional” approach has already been employed by appellate courts in Can-
ada.”‘ One hopes that it will be adopted by the Supreme Court of Canada as
well.
2.
Foreseeability, Duty of Care and Omissions
The move in the Commonwealth away from the all-inclusive foreseeability
approach in Anns suggests a need to consider whether foreseeability alone is a
suitable approach to the category of liability of public authorities. In the vast
majority of negligence cases involving persons, foreseeability is still a useful
and often decisive criterion for the determination of proximity and duty. At the
risk of oversimplification, when defendants have deliberately set in motion a
chain of events that resulted in physical injury to the person or property of the
plaintiff, it is proper that they should be held accountable unless it is shown that
they could not have foreseen the result. This, of course, was the case in Dono-
ghue v. Stevenson. According to the facts as alleged,”‘ Stevenson initiated the
system of manufacturing and distributing that caused the bottle of ginger beer
to be served to Mrs. Donoghue. Causation was not controversial. The issue was
whether the factory owner was in a relationship of proximity with the consumer.
Lord Atkin concluded that there was a relationship of proximity because Ste-
venson could foresee that Mrs. Donoghue, or persons similarly situated, would
come to harm because of a chain of events he put in motion. When factual cau-
sation is not controversial, foreseeability may be a shortcut to determine
“‘Supra, note 99 at 30.
“l6As Smith & Burns observe, supra, note 99 at 149:
One might wonder why it would be necessary to continue the separate categories, once
the general principle has been recognised and formulated…. The reason is that there
are many important policy considerations involving different interests in the context of
different kinds of conduct that could not all be accommodated and taken into account
in the context of a general principle. For example, a mere affront to dignity can be
actionable in assault and battery, while the most damaging and intentional attack on
reputation and character are not actionable in defamation, providing that the allegation
is true.
(1991), 53 B.C.L.R. (2d) 180 (C.A.).
117See the judgment of Lambert J.A. in Edgeworth Construction Ltd v. N.D. Lea & Assoc.
“8Recall that Donoghue v. Stevenson was heard on a preliminary motion so there were no expli-
cit findings of fact.
1992]
LIABILITY OF PUBLIC AUTHORITIES
whether the defendant should compensate the plaintiff. Even so, the source of
the duty to compensate”9 is not the fact that the defendant should have foreseen
the injury. As Smith and Bums stated in an article cited with approval by the
House of Lords and the Australian High Court, “[c]ausation is the principal
foundation of liability in the law of torts.”‘”0 Defendants are liable to compen-
sate plaintiffs because they caused the plaintiffs’ injuries.”‘ This statement must
be regarded cautiously, but it provides a first working hypothesis. In this context
the source of the duty to compensate is the act of the defendant which caused
the injury to the plaintiff. Foreseeability is not a source of a duty to compensate,
but rather a limiting principle: defendants are obliged to compensate only those
who they could have foreseen would be injured by their deliberate actions.,”
However, in the category of cases we are considering it is not obvious that the
defendants caused the accidents at all. In Just, the rock that fell on the car was
not set on the hill by the province, nor was it dislodged by the province. In Lau-
rentide Motels, the fire was started by a careless hotel guest, not by the fire ser-
vice. Finally, in Rothfield, the wall was constructed inadequately by the contrac-
tor, not by the city.
There is considerable debate about whether an omission can ever be the
cause of an accident. For example, Gibbs C.J. stated in Sutherland Shire:
I am disposed to agree that there is a basic difference between causing something
and failing to prevent it from happening… . Thus it has been held that damage
resulting from tick infestation was not caused by a negligent failure to eradicate
ticks.’2 In Kamloops it was held that a council’s negligence caused the plaintiff’s
damage, but with the greatest respect I am not sure that the distinction between
causing damage and failing to avert it was fully examined. If a building inspector
negligently fails to carry out an inspection which, if properly made, would have
revealed defects, and the building is completed, it does not seem right to say that
the inspector’s negligence caused the defects, although had the inspector not been
negligent he might have prevented the building from being completed in a defec-
tive state.’ 24
1191 prefer to speak of a duty to compensate rather than of a duty of care.
12Smith & Bums, supra, note 99 at 148.
121Causation is essential to the Aristotelian conception of commutative (or “corrective,” or “rec-
tificatory”) justice which has influenced thinking in tort law. See Ethica Nicomachea in W.D. Ross
ed., The Works of Aristotle, vol. IX (Oxford: Oxford U. Press, 1925) at 1132a 4: “MIThe law looks
only to the distinctive character of the injury and treats the parties as equal, if one is in the wrong
and the other is being wronged, and if one has inflicted injury and the other received it.” Below,
I look very briefly at a justification for compensation in which causation and fault are irrelevant.
122See the statement of Lord Wilberforce in McLouglin, supra, text accompanying note 100.
123Administration of Papua and New Guinea v. Leahy (1961), 105 C.L.R. 6 at 10, 12, 21 (H.C.
Aus.).
124Supra, note 40 at 15, 17. See also Deane J. in Sutherland Shire, ibid. at 63. He had earlier
described Kamloops as a “decision of the Supreme Court of Canada which provides a striking
example of liability for inaction.”
REVUE DE DROIT DE McGILL
[Vol. 37
This position, or something like it, is shared by Weinrib,'” Smith and Bums, 26
and others. Some of those who hold this view of causation might nevertheless
impose moral and legal responsibility for failing to take positive steps to avert
danger in appropriate circumstances. Here, however, the first step is to articulate
whether there is such a duty in the circumstances. Such a duty could not be
derived from foreseeability of harm alone, but it might be based on other moral
values. I suggested earlier that the statement that “[c]ausation is the principal
foundation of liability in torts” must be regarded cautiously. As Professor
Stephen Perry argues, causation can never provide a complete theory of respon-
sibility even where the defendant’s actions were among the principal causes of
the plaintiff’s injuries.”2 One might go further and say that the fact that causa-
tion is the foundation of responsibility in one class of cases does not mean that
it is the only foundation, or even a necessary element. A ground of responsibil-
ity may be the relationship between the parties. That is, whether they are in a
personal relationship such as to give rise to a positive duty on one party in
favour of the other even where the one did not create risk that endangered the
other.’ If there is such a duty the responsibility of the defendant is engaged. 29
In short, the duty of care arises not from foreseeability alone, but from a pre-
existing relationship of proximity of which foreseeability may form a part.
An alternative position, identified with Hart and Honor6, 30 is that omis-
sions (of which a failure to avert danger may be one) may in fact cause an
injury. If a result occurs because one fails to do what is normally done, then the
failure can be seen as a true cause of the result. The link between what would
otherwise be a non-culpable omission and causation is the notion of “what is
normally done.”‘ I suggest that when we are speaking about a duty of positive
2E.J. Weinrib, “The Case for a Duty to Rescue” (1980) 90 Yale L.J. 247 at 256: “Participation
by the defendant in the creation of the risk, even if such participation is innocent, is thus the crucial
factor in distinguishing misfeasance and nonfeasance.”
126Supra, note 99 at 154: “In [Anns] the inspectors did not cause the damage to the structures.
The damage was caused by the builders who negligently built the foundations. While it is true that
the inspectors may have carelessly approved the inadequate foundations, they did not cause the
damage to happen through their actions. Rather, they allowed it to happen, or failed to prevent it
from happening.”
127S.R. Perry, “The Impossibility of General Strict Liability” (1988) 1 Can. J. L. & Juris. 147
at 169.
128in Weinrib’s case, supra, note 125, the relationship between the parties is not an actual rela-
tionship between the particular parties, but a hypothetical, idealized, universalized relationship
derived from Kantian deontological arguments.
129There is some controversy about whether responsibility which is divorced from causality
should carry with it a further obligation to compensate, or whether penal sanctions would be more
appropriate. This question is beyond the scope of this article.
cussed in T.M. Benditt, “Liability for Failing to Rescue” (1982) 1 Law & Phil. 391.
’30H.L.A. Hart & A.M. Honor6, Causation in the Law (Oxford: Clarendon Press, 1959), dis-
’31Benditt, ibid., is of the opinion that this is not really a theory of causation, but rather a theory
of responsibility.
19921
LIABILITY OF PUBLIC AUTHORITIES
action by one person in favour of another, “what is normally done,” i.e. whether
there is such a duty to act, will depend on the relationship of proximity between
the parties.
Thus, even if one accepts the view that an omission can be the true cause
of an accident, foreseeability is not an appropriate shortcut for finding a duty of
care. The omission is a “cause” of the accident because we believe the parties
were already in a relationship such that it would be “normal” for one to look
out for the interests of the other. Foreseeability is relevant only after that rela-
tionship is established..
The Caparo Industries approach to duties of care, which concentrates on
categories of relationships, has the merit of obliging us to shift our focus among
the elements of a duty of care, proximity, foreseeability, and causation, in dif-
ferent categories of cases. In the category of cases where the defendant’s actions
are a true cause of the injuries suffered by the plaintiff, proximity, or the rela-
tionship between the parties, is relatively unimportant, and we can focus on
whether the defendant should have foreseen the consequences of her or his
actions. ’32 But in the category of cases where someone or something other than
the defendant is the factual cause of the plaintiff’s injuries, we must focus more
directly on the relationship between the parties to see whether the defendant had
a duty to prevent the harm. Foreseeability is always an issue, but it becomes an
issue in this category of cases only after it has been established that the relation-
ship between the parties raised a duty in the defendant to prevent others, or
nature, from injuring the plaintiff. The three cases we are examining fall into the
second category. As we will see, the Supreme Court does not distinguish the two
categories of cases, and focuses immediately on foreseeability in both.
In summary, where the injury in question was not the direct result of a
deliberate act of the defendant we cannot go straight to the question of foresee-
ability. We must first see whether the parties were in a relationship which gives
rise to a recognized common law duty. If the cause of the plaintiff’s injury was
an act of a third party or an act of nature, we must first see whether there was
a common law duty to take positive steps to protect the plaintiff from that dan-
ger. Such positive steps might include issuing warnings, or actually removing
the danger itself.
B. Specific Detrimental Reliance – A Relationship of Proximity, Good
Samaritan Rules, and Duties to* Warn
If the foreseeability thesis is so unhelpful in certain categories of negli-
gence involving liability of public authorities, what should replace it? Caparo
132Altematively, one might say that a relationship of proximity flows automatically from the fact
that a reasonable defendant would have foreseen the consequences of his or her actions.
McGILL LAW JOURNAL
[Vol. 37
Industries suggests that we should not try to find a single basis for negligence
liability. Rather, we ought to develop categories of duties based on, among other
things, the nature of the relationship between the parties and the type of injury
suffered. The first obstacle, when considering a case of negligent inspections,
as in Rothfield and Just, or negligent maintenance of fire hydrants, as in Lau-
rentide Motels, is to see whether there is a private law duty to do that which it
is alleged the government has failed to do. Persons do not normally owe one
another the duty to seek out and eliminate potential sources of danger, nor do
they even owe a duty to warn about clear and imminent peril or effect an easy
rescue.’33 But this does not close the matter. In Caparo Industries and Suther-
land Shire it was suggested that courts may develop new categories of negli-
gence duties by analogy from recognized duties. Courts have found an analogy
to the duty to warn and the duty to rescue fruitful.
In recent years some of the more “sickening examples of callous refusal to
help, followed by immunity from tort liability”‘” have been mitigated,’ but
few jurisdictions have imposed an affirmative duty to rescue one who is in
peril.’36 Nevertheless, many American jurisdictions have begun to accept what
is known as a “Good Samaritan” rule based on the Restatement.3 7 According to
133Osterlind v. Hill, 160 N.E. 301 (Mass. 1928) (drowning person); Gautret v. Egerton (1867),
2 C.P. 371 at 375, cited by Gibbs C.J. in Sutherland Shire, supra, note 40 (no general duty to warn
one who is headed into danger); Buch v. Amory Manufacturing, 44 A. 809 (1897); Dorset Yacht
Co. v. Home Office, [1970] A.C. 1004 at 1027, [1970] 2 All E.R. 294 (H.L.) (no general duty to
assist a person in distress).
But a duty to rescue might arise if the parties are otherwise in a special relationship of proximity:
Horsley v. MacLaren, [1972] 2 S.C.R. 441, 22 D.L.R. (3d) 545 (special duty from proximity rela-
tionship of passenger –
boat owner, derived from statute); Anderson v. Atcheson, 333 U.S. 821
(1948) (employer-employee); Millette v. Cdt , [1971] 2 O.R. 155, 17 D.L.R. (3d) 247, rev’d in part
[1972] 3 O.R. 224, 27 D.L.R. (3d) 676 (C.A.), aff’d in part [1976] 1 S.C.R. 595, 51 D.L.R. (3d)
244 (police officer and driving public); Schacht v. R., [1973] 1 O.R. 221, 30 D.L.R. (3d) 641, aff’d
[1976] 1 S.C.R. 53, 55 D.L.R. (3d) 96, 3 N.R. 453 [hereinafter Schacht] (police and driving pub-
lic); Suarez v. Dosky, 407 A.2d 1237 (N.J. Super. Ct. App. Div. 1979) (police officers assisting chil-
dren on cross-walk).
I34A.M. Linden, “Rescuers and Good Samaritans” (1971) 34 Mod. L. Rev. 241 at 241.
’35Largely by providing rescuers who injure themselves damages against the person who made
136But see the Quebec Charter of Human Rights and Freedoms, R.S.Q. 1975, c. C-12, art. 2:
the rescue necessary. See ibid. at 252ff.
Every human being whose life is in peril has a right to assistance.
Every person must come to the aid of anyone whose life is in peril, either personally
or by calling for aid, by giving him the necessary and immediate physical assistance,
unless it involves danger to himself or a third person, or he has another valid reason.
A Vermont Good Samaritan statute states that a duty is owed not only to those whose lives are
at risk, but also those “exposed to grave physical harm.” However, the penalty for wilfully violat-
ing the duty is a fine of only $100. See Benditt, supra, note 130 at 392.
137Restatement, supra, note 9.
323 Negligent Performance of Undertaking to Render Services
One who undertakes, gratuitously or for consideration, to render services to another
which he should recognize as necessary for the protection of the other’s person or
19921
LIABILITY OF PUBLIC AUTHORITIES
the Restatement there is no general duty to avert danger to others, but a duty to
warn about or avert danger is created where: (1) the defendant acts in such a
way (2) in relation to a particular plaintiff’s safety (3) that it would be reason-
able for the plaintiff to rely on the defendant, and (4) the defendant has actually
“induced [the plaintiff] to forego other remedies or precautions against the
risk.”‘ 3 The Good Samaritan rule has been pleaded in cases of allegedly neg-
ligent inspections by government agencies. This reliance-based duty is a useful
analogy, and it has gained cautious approval in Australia. Its roots in the ungen-
erous common law rule are, however, noticeable, and the elements which the
plaintiff must prove are quite strict. I will now consider the application of a det-
rimental reliance analysis in the United States and Australia.
1.
United States
In the United States, a cause of action based on the duty to warn was rec-
ognized in the 1955 case Indian Towing.’39 In Indian Towing, a tug and barge
went aground on an island in the Mississippi. The island had been marked by
a lighthouse, and the location of the lighthouse was indicated on navigational
charts, but the light in the lighthouse went out following a careless inspection
by a coast guard employee. The Supreme Court decided that lighthouse inspec-
tions were among the “operational” duties of the coast guard. The Court recog-
nized, however, that finding that the function was operational only opened the
door to private law liability; the plaintiff still had to prove the relationship
between the government and himself was similar to a relationship that could
things, is subject to liability to the other for physical harm resulting from his failure
to exercise reasonable care to perform his undertaking, if
(a) his failure to exercise such care increases the risk of such harm, or
(b) the harm is suffered because of the other’s reliance upon the
undertaking.
324A Liability to Third Person for Negligent Performance of an Undertaking
One who undertakes, gratuitously or for consideration, to render services to another
which he should recognize as necessary for the protection of a third person or his
things, is subject to liability to the third person for physical harm resulting from his
failure to exercise reasonable care to protect his undertaking, if
(a) his failure to exercise reasonable care increases the risk of such harm,
or
(b) he has undertaken to perform a duty owed by the other to the third per-
son, or
(c) the harm is suffered because of reliance of the third or the third person
upon the undertaking. (emphasis added)
see Blessing, supra, note 71 at 1187.
13SRestatement, ibid. at 144, 324A, comment (e). For a sample of cases employing the rule,
139Supra, note 11. Reliance as a basis for liability for a negligent rescue was recognized for indi-
viduals in Zelenko v. Gimbel Bros., 287 N.Y.Supp. 134 (N.Y. Sup. Ct. 1935). Causation was a
major factor. the Court was impressed by the fact that the “rescuer,” by undertaking to provide
medical assistance, precluded other efforts to provide medical assistance.
REVUE DE DROIT DE McGILL
[Vol. 37
exist between individuals, and that a private law duty of care would arise in
such a relationship. The fact that the coast guard had undertaken to facilitate
navigation by marking hazards did not by itself place it under a general liability
for injuries due to the continued existence of navigational hazards. Such a
requirement would be so open-ended as to resist precise definition. It has, more-
over, no analogue in private law: there is no general private law duty to warn
others that they are heading into danger. Rather, the governnient’s liability was
founded on the much more specific fact that the coast guard had for some time
operated a light in a particular place, that it had caused charts to be printed that
indicated the location of the light, and that mariners had come to rely on the
light.
Liability was based on the principle of specific detrimental reliance based
on the duty to warn. A private law duty to warn can be raised once persons have
assumed the duty to warn about a particular danger so that others rely on such
persona’ continued warnings. They must continue to warn until such time as
they inform others that they intend to abandon the duty. When a duty to warn
is based on specific detrimental reliance, proximity is established between the
person who warns and the victim by: (1) the conduct of the defendant (the past
warnings); (2) the fact that the plaintiff has altered her or his conduct in antic-
ipation of future (or continued) warnings; and (3) the fact that the victim relies
on the specific warnings.14 Elements 1 and 2 supply the relationship of proxim-
ity, including the requirement that the perspective of both parties be taken into
account, and the third element provides the causal link between the defendant’s
conduct and the plaintiff’s injuries. The focus of the cause of action based on
detrimental reliance is the relationship of proximity between the particular par-
ties. The question of foreseeable harm arises only after the relationship has been
established. In this respect, the case anticipated the re-interpretation of the prin-
ciple in Donoghue v. Stevenson, as expressed in recent decisions of the House
of Lords and the High Court of Australia.14′
Since Indian Towing, the Restatement has been adopted by the American
Law Institute, 42 and the cause of action based upon the duty to warn has been
14Compare to Sheppard, supra, note 19. Local authority sued when plaintiff injured himself
because he could not see in a street where the lights had been deliberately extinguished at 9:00 pm.
The authority had a power, but not a duty, to light the street. It placed a lamp at a dangerous point
in the street (the danger not of its own making). Although it could be foreseen that extinguishing
the lights would be dangerous, the authority was held not liable on the ground that it did not create
the danger, that it had a power but not a duty to set up the lights in the first place, and since it had
no duty to set up the lights, it had no duty to continue illuminating the lights.
141Some might argue that this type of approach risks returning the law of torts to the “bad old
days” before Donoghue v. Stevenson, when plaintiffs were obliged to show that the cause of action
they alleged fit within a previously recognized category. It is true that this type of approach would
involve a retreat from grand principles, recognizing more differentiated categories of negligence.
This, however, is probably unavoidable. See the authors cited supra, note 99.
142On 25 May 1963 and 22 May 1965.
1992]
LIABILITY OF PUBLIC AUTHORITIES
pleaded in safety inspection cases. If all the conditions listed above are met, the
inspecting agency comes under a duty to “warn” the plaintiff about dangers that
it has or ought to have discovered through its inspections. 143
The cause of action has not, however, been easy to establish. In Patentas
v. United States,'” the owners and employees of a Greek tanker, in United
States waters, sued the Coast Guard because an allegedly negligent inspection
failed to detect the cause of an explosion that occurred shortly after the inspec-
tion. The Third Circuit expressly adopted the language of the Restatement. The
Court held that while it was possible to base a claim on detrimental reliance
induced by the conduct of the agency, in this case the necessary elements had
not been established. The most important missing element was proof that the
plaintiffs had, as a result of the conduct of the Coast Guard, been induced to
forego their own inspections.
In Blessing,145 plaintiff employees sued the Occupational Health and Safety
Administration for alleged negligence in inspecting manufacturing facilities.
The Court held that the inspections were intended to supplement, not supplant,
the employers’ inspections. The employees would be able to establish that they
had relied on the inspections to their detriment if they could show that the
employers had been induced to reduce their own safety programs as a result of
the inspectors’ conduct. As the inspectors had done nothing to suggest either to
the employers or the employees that they were assuming responsibility for
ensuring a safe workplace, they had no private law duty to the employees.
In Clemente v. United States,'” the relatives of airline crash victims alleged
that the victims had relied on the duty of air traffic controllers to protect their
safety by ensuring that overweight planes would not be permitted to take off.
The Court held that the plaintiffs had not established that the victims had altered
their conduct on the basis of the agency’s conduct. In addition, the First Circuit
would also have struck out the cause of action because the agency was not in
fact under a duty to prevent overweight planes from taking off, and it did not
act as if it did. 47
A claim based on the duty to warn is regarded with particular suspicion
when the defendant is a regulatory agency. As Burger C.J. stated in Varig
Airlines,
143Both Just and Rothfield may be seen as inspection cases. The aspect of Laurentide Motels
dealing with fire hydrant inspections could also be considered an inspection situation.
144687 F.2d 707 (3d Circ. 1982) [hereinafter Patentas].
145Supra, note 71.
146567 F.2d 1140 (1st Circ. 1977).
147A statutory duty to act does not create a private law right: Lacey v. U.S., 98 FSupp. 219 (D.
Mass 1951) (the Coast Guard has no private law duty to rescue even though it has a statutory
responsibility to do so); In Re Silver Bridge Disaster Litigation, 381 F.Supp. 931 (S.D.W.Virg.
1974) (statutory obligations to inspect bridges are public in nature, and thus do not give rise to pri-
vate rights of action).
McGILL LAW JOURNAL
[Vol. 37
[w]hatever else the discretionary function exception may include, it plainly was
intended to encompass the discretionary acts of the Government acting in its role
as a regulator of the conduct of private individuals… .148
This suspicion has two main bases. First, the creation of a procedure for
policing compliance does not alter the fact that the responsibility for compliance
rests with the regulatory targets named in the enabling statute: the “police” try
to make sure the law is obeyed, but they are not themselves guilty when the law
is broken. 49 This reasoning is particularly forceful when the plaintiff is himself
or herself charged with compliance, as in Rothfield. Even if the inspections were
regarded as operational, thus outside the ambit of sovereign immunity, the plain-
tiffs would fail because instituting the inspections did not create new private law
duties on the inspectors, or displace existing ones on the regulatory targets. 5′
The second reason reflects judicial reluctance to interfere with decisions
establishing “philosophical priorities.”
Congress wished to prevent judicial “second guessing” of legislative and admin-
istrative decisions grounded in social, economic, and political policy through the
medium of action in tort.1 52
Decisions on “compliance review” involve a trade-off between economy
on one side, and full compliance on the other. Thus, a compliance inspection
regime that would foreseeably permit a certain number of violations is not for
that reason unreasonable: it may be that reasonable concern with thrift makes
it necessary to accept a less than perfect system of inspections. It follows there-
fore that the mere fact of a particular violation (which injures the plaintiff) says
nothing whatsoever about the private law liability of the inspecting authority.
These principles are illustrated in Blessing. 51 In addition to the reasons
outlined above, the Blessing court refused to recognize a cause of action
because the inspection system of the health and safety authority was in the
nature of policing compliance. It was not reasonable to hold the “police”
responsible for the conduct of the offenders. This is consistent with the view
expressed by the English Court of Appeal in Acrecrest Ltd v. W.S. Hatrell &
Partners,”4 a post-Anns building inspection case.
148Supra, note 71 at 813.
1491bid. See also Smith v. Leurs (1945), 70 C.L.R. 256 at 262 (H.C. Aus.): ‘The general rule is
that one man is under no duty of controlling another man to prevent his doing damage to a third.”
150Varig Airlines, supra, note 71.
151See text accompanying notes 24ff.
152Restatement, supra, note 9 at 412, 895D(b).
153Supra, note 71.
154[1983] Q.B. 260, [19831 1 All E.R. 17, [1982] 3 W.L.R. 1076 (C.A.). The local building
authority’s function is not to restrain the wrongdoer from persisting in his wrongdoing.
1992]
LIABILITY OF PUBLIC AUTHORITIES
2.
Australia
The leading Australian case, Sutherland Shire, 5′ arose from facts which
are nearly identical to those in Rothfield’56 The High Court had before it the
same body of precedent cited by the Canadian Supreme Court in its trilogy of
government liability decisions, but it reached very different conclusions.
In Sutherland Shire, land owners submitted an application to a local
authority for a permit to erect a house on their land. The plans did not include
a detail of the foundations, but the permit was issued under the condition that
the owners submit a survey of the foundations to the Council, and that the
Council be given notice at various stages of construction. The Council’s records
did not indicate whether the inspections had in fact been made. There was no
evidence that the owners had given notice to the Council, or that the building
had in the end been passed for occupation. Each of the judges who wrote opin-
ions agreed that when an authority acts in such a way that the plaintiff is
induced, to her or his detriment, to rely on the skill of the authority, an action
may lie for damages.’57 Mason J. (as he then was) made the reliance thesis the
centre of his judgment:
Generally speaking a public authority which is under no statutory obligation to
exercise a power comes under no common law duty of care to do so… But an
authority may by its conduct place itself in such a position that it attracts a duty
of care which calls for the exercise of the power. 5
8
He agreed with his colleagues in preferring Peabody to Anns, holding that the
ability of an authority to foresee that a failure to exercise a power would harm
the plaintiff would not create a duty to exercise the power. However, a duty can
arise where
a person, by practice or past conduct upon which others come to rely, creates a
self-imposed duty to take positive action to protect the safety or interests of
another, or at least to warn him that he or his interests are at risk.’ 59
He cited’and adopted a number of the American cases. As in the American
cases, the Australian judges were of the opinion that the authority’s conduct and
reliance on the part of the plaintiff had to be taken seriously. Here, however,
where the builder did not give the required notice to the Council it could not be
said that he had relied on the judgment of the Council’s employees. The plaintiff
was therefore denied recovery.
155Supra, note 40.
15 6Supra, note 3.
157See Sutherland Shire, supra, note 40, Gibbs C.J. at 17, Brennan J. at 47, Deane J. at 65, and
Mason J. at 29.
158Ibid. at 28.
159Ibid. at 29.
REVUE DE DROIT DE McGILL
[Vol. 37
Since Sutherland Shire, courts have been cautiously receptive to the idea
of a cause of action basing detrimental reliance upon continued inspections. In
McCauley v. Hamilton Island Enterprises Pty. Ltd,6 the plaintiff was injured in
a helicopter accident at a floating helipad off the Great Barrier Reef. He sued
the government agency with authority over the Great Barrier Reef on the ground
that the helicopter accident which injured the plaintiff would have been avoided
if the authority had properly regulated aviation in the area. The plaintiff argued
that the Commonwealth knew or should have known about the helicopter
flights, that the government knew or should have known that they were unsafe,
and that he relied on the government to prohibit activities which it knew were
unsafe. On a preliminary motion the Commonwealth sought to have the action
dismissed as disclosing no cause of action. The trial judge refused to do so, say-
ing that on the basis of Mason J.’s judgment in Sutherland Shire, the plaintiff
was entitled to try and prove his case. He was of the opinion, however, that “the
plaintiff is confronted with a difficult task in demonstrating a case of general
reliance, especially in terms of causation.” ’61 In McDonogh v. Commonwealth
of Australia,62 the plaintiff was injured when the truck he was driving over-
turned on a road on federal land. The way the road was maintained created the
appearance that it was level and firn, but in fact the outer portion was inade-
quate to support the plaintiff’s truck. At trial, the plaintiff sued on the grounds
that (1) the authority had been negligent in maintaining the road and (2) the
authority failed in its duty as occupier. The Federal Court of Appeal dismissed
the second ground. On the issue of negligence, the Court found that while a pub-
lie authority cannot, in general, be held liable for the state of repair of the road,
it can be held liable if, by its conduct, it induces drivers to believe that it is
actively ensuring that the road is adequate as a road. The plaintiff succeeded on
the ground that the authority had so conducted itself, and that he had relied on
that conduct.
On the other hand, the courts insist that the plaintiff’s reliance be reason-
able, especially in light of the general duty to the public owed by public agen-
cies. In San Sebastian Pty Ltd v Minister Administering the Environmental
Planning and Assessment Act,’63 developers sued the state government and the
City of Sydney. The developers alleged that the governments had induced them
to purchase and consolidate land holdings through the publication of a zoning
plan that allowed high density housing. The High Court held that statements and
conduct of government authorities could, in general, be the basis for an action
in detrimental reliance. Significantly, however, it held that such reliance must
take into account the nature of the public authority, and in particular the fact that
160(1987), 75 A.L.R. 257 (Fed. Ct. Gen. Div.).
6 1The law reports do not indicate that a trial has been held.
1
162(1985), 73 A.L.R. 148 (Fed. Ct. Gen. Div.).
163(1986), 68 A.L.R. 161 (H.C.).
19921
LIABILITY OF PUBLIC AUTHORITIES
the general purpose of public authorities is to act in the public interest. It would
be unreasonable to rely on an authority to act in a private interest if doing so
would fetter its discretion to act in the public interest (in whatever manner pub-
lic interest be defined).”6 In Skuse v. Commonwealth,’65 the Federal (Appeal)
Court held that a duty to warn based on detrimental reliance could not be rec-
ognized where it would interfere with other important administrative concerns.
In Skuse, a lawyer attending court was shot by a man who confused the lawyer
with another lawyer. Prior to the incident, the local police had been warned by
state police that the assailant had threatened to kill his lawyer, but they did not
act. On the day of the incident the assailant’s daughter informed the police that
he was headed to the courthouse with a gun, but they were unable to intercept
him. The lawyer sued on the ground that he had relied generally on the author-
ities to keep the courts safe. The Federal Court of Appeal held for the Common-
wealth on the ground that recognizing a private law duty of care would interfere
with a matter of general public concern; namely, whether the police or the courts
held ultimate authority over security within the court buildings. The claim might
also have been rejected on the ground that security staff had not conducted itself
in relation to the particular plaintiff so as to lead him to alter his conduct.
3.
Specific Detrimental Reliance: Summary
Specific detrimental reliance, based on an analogy to the Good Samaritan
duty to continue past warnings which have induced reliance by the plaintiff, is
a promising basis for developing a rational private law duty on the part of public
authorities to regulate or inspect. When an agency induces a particular plaintiff
to rely on its inspections, and when the plaintiff does rely on the inspections and
reasonably alters his or her conduct because of that reliance, a private law duty
to inspect carefully may arise.
Specific detrimental reliance takes seriously the conduct of the agency
towards the plaintiff, and the reliance of the plaintiff on the agency, so it incor
porates the sort of mutual regard implied by the terms “neighbourhood” and
“proximity.” The requirement that the plaintiff have altered his or her conduct
supplies the necessary element of causality. Put crudely, the “neighbourhood”
provides the moral basis for a duty, and, together with causation, provides a rea-
son for insisting that the agency compensate the plaintiff. The duty is, of course,
bounded by foreseeability in that the agency is not culpable if it could not have
164See also L Shaddock & Associates Prop. Ltd v. Parramatta (City) (1981), 150 C.L.R. 225
(H.C. Aus.). A council that gave information upon which it was reasonable to expect a home pur-
chaser to rely could in principle be liable if the purchaser in fact relied on that information to his
or her detriment. However, it was not reasonable for a purchaser to rely on anonymous telephone
information, and it would not be reasonable for a council telephone information officer to reason-
ably expect that such telephone information would be acted on without further verification.
’65(1985), 62 A.L.R. 108 (Fed. Ct. Gen. Div.).
McGILL LAW JOURNAL
[Vol. 37
foreseen the plaintiff’s reliance or if the plaintiff’s reliance was so unreasonable
as to be beyond anticipation. In these ways the cause of action is consistent with
the re-interpretation of tort duties in Caparo Industries and Sutherland Shire.
The cause of action has been recognized in the United States and Australia,
although plaintiffs have had limited success. Recognizing the possibility of the
cause of action does not make the government liable in private law generally for
injuries suffered in relation to activities over which the government has a mea-
sure of regulatory control. On the other hand, if the government induces reliance
by undertaking a specific safety measure, it will be held liable to compensate
persons who are injured if it fails to do so carefully.
Finally, it must be recognized that the detrimental reliance analogy, which
imposes liability on the government in circumstances where the government
acts like a person, may not go far enough in some instances. We may wish to
impose additional duties on government agencies that we are not willing to
accept as individuals. In some cases, concerns for the liberty of the individual
prevent the recognition of duties which a generous morality and more commu-
nitarian spirit would willingly embrace. For example, the common law’s hesi-
tation in accepting affirmative duties to rescue is often justified on the ground
that such a duty would unduly impose upon the individual’s freedom of
action.”6 Since individual liberty is not an issue where agencies of the state are
concerned, the moral duty to rescue could become a legal duty for certain gov-
ernment agencies. We might, therefore, wish to recognize duties on government
agencies which are broader than those imposed between persons. This analysis
might prove helpful in dealing with police and fire protection cases.’67 But even
in these cases, where it is not helpful to analogize government activity to per-
sonal activity, the focus remains the relationship between the parties. A specific
detrimental reliance-based analysis was at one time considered persuasive by
166See supra, note 134.
167Jane Doe v. Metropolitan Toronto Police Commission (1989), 58 D.L.R. (4th) 396, 48
C.C.L.T. 105 (Ont. H.C.J.) (potential cause of action based on duty of police to warn persons
known to be potential target of serial rapist); Air India Flight 182 Disaster Claimants v. Air India
(1987), 44 D.L.R. (4th) 317, 62 O.R. (2d) 130 (H.C.) (suit against government and officers for neg-
ligent investigation of terrorist threat); Lafleur v. H.M.TQ. (1990), 48 B.C.L.R. (2d) 180, 4
C.C.L.T. (2d) 78 (S.C.) (duty of police to arrest drunk motorcyclist to protect him from riding and
crashing his motorcycle); Schacht, supra, note 133 (duty to mark accident site); Suarez v. Dosky,
407 A.2d. 1237 (N.J. Super. Ct. App. Div. 1939) (police assisting children across a cross-walk);
Schuster v. City of New York, 154 N.E.2d 534 (N.Y. 1954) (duty to protect persons who identify
dangerous criminals); Crouch v. Hall, 406 N.E.2d. 303 (Ind. Ct. App. 1980) (police investigation
failed to apprehend known rapist who raped and killed daughter of plaintiff; no liability on “public
duty doctrine” and discretionary function exception); Ketchum v. Alameda County, 811 F2d 1243
(9th Circ. 1987); Balistreri v. Pacifica Police Dept., 855 F.2d 1421 (9th Circ. 1988); Jensen v. Con-
rad, 747 F.2d 185 (4th Circ. 1984), cert. denied, 105 S.Ct. 1754 (1985) (duty of police to protect
where victim’s vulnerability known to police).
1992]
LIABILITY OF PUBLIC AUTHORITIES
the Supreme Court of Canada. 6 ‘ However, in the most recent cases where it
would have been particularly illuminating, it was not considered. While the
action in detrimental reliance based on the analogy to a duty to warn or rescue
is a promising analytical tool, it need not be the sole basis for imposing govern-
ment liability. The chain of reasoning does, however, provide a strong model.
In contrast, the Anns foreseeability test remains an obstacle to reaching a prin-
cipled basis for imposing liability.
C. The Supreme Court Trilogy
In the recent cases involving inspection and regulation by public authori-
ties, the Supreme Court, after employing the policy-operational distinction,
relied on an Anns-inspired foreseeability approach in analyzing the issue of duty
of care. This section will focus on the Supreme Court’s treatment of the duty
issue in the two recent common law cases, Rothfield and Just, critiquing the
foreseeability approach and demonstrating the superiority of a specific detri-
mental reliance analysis. In adopting a specific detrimental reliance approach,
we would get a different and far more satisfactory result. As the three cases
demonstrate, the policy-operational distinction and the Anns foreseeability
approach lead the court to find prima facie duties of care. In contrast, a specific
detrimental reliance approach would force consideration of the actual relation-
ship between the parties, thus staying truer to a traditional tort law analysis.
1.
Rothfield v. Manolakos
It will be recalled that in Rothfield, homeowners who wished to enjoy a
level backyard employed contractors to build a retaining wall to support the
built-up slope behind their house. 69
a. Justice La Forest
La Forest J., writing for the majority, concluded that once the City made
a policy decision to enact a house inspection by-law, it “owed a duty of care to
all who it is reasonable to conclude might be injured by the negligent exercise
168See R. v. Nord-Deutsch Verischerungs Gesellschaft, [1971] S.C.R. 849 at 863,20 D.L.R. (3d)
444, Ritchie J. (ship collision due in part to breach of Department of Transport’s duty to maintain
range lights “upon which mariners were entitled to place reliance”). See also Grossman v. R.,
[1952] 1 S.C.R. 571, 2 D.L.R. 241, Kellock J. (when the Crown undertakes to mark obstructions
on airport runways, pilots are entitled to rely on the assumption that that duty will be discharged
carefully). The specific detrimental reliance thesis has also been explicitly accepted in England.
One of the earliest instances of a plaintiff succeeding on the basis similar to the Good Samaritan
test of the Restatement, supra, note 9 was the judgment of Slesser L.J. in Kent v. East Suffolk,
[1939] 4 All E.R. 174 at 175 (C.A.). The decision was reversed by the House of Lords, supra, note
18. Surprisingly, Lord Atkin was particularly hard on the reliance thesis.
169See above, subsection I.D.3.
REVUE DE DROIT DE McGILL
[Vol. 37
of those powers.”’17 Surely La Forest J. was not saying that making a policy
decision is itself grounds for imposing a private law duty of care. Such a posi-
tion would have no basis in principle or precedent. Perhaps the source of the
duty of care is to be supplied by the foreseeability theory of the existence of
duties of care: the City owes a duty to those it reasonably foresees will be
injured. As we have already seen, that thesis as relating to public authorities has
been roundly rejected in England and Australia, and for good reason. But even
if foreseeability were generally an acceptable basis for private law duties of care
for public authorities, it would not explain the City’s duty in this case. Compare
two hypothetical situations. In the first, the City codifies rules of prudence by
enacting a regulation to govern a certain activity. It foresees that persons who
do not observe the code of conduct might suffer. In the second hypothetical, the
City does not establish the code, but it is still able to foresee that persons who
do not observe the rules of prudence will suffer. No one would impose a duty
of care on the City in the second case, yet the element of foreseeability is the
same as in the first hypothesis. 7′ Foreseeability, therefore, is not the critical cri-
terion upon which the existence of a duty of care exists. The duty rests almost
entirely on the imputed consequences of its making a “policy” decision to cod-
ify the rules of prudence. As we have seen, it is all too easy to draw the false
conclusion that the government activity which follows from a “policy” decision
is automatically operational. Hence a version of the policy-operational distinc-
tion was used as a prima facie basis for imposing a duty.
In short, La Forest J. assumed, rather than proved, the existence of a pri-
vate law duty of care. However, he did discuss the reliance-based cause of
action in the course of rejecting Cory J.’s reasons for judgment. While Cory J.
stated that third parties might be entitled to rely on the City, owner-builders’
reliance on the City could not be reasonable because of the regulatory context
of the by-laws. La Forest J. decided that reliance was reasonable.
As a preliminary matter, it is not clear to me how owner- builders, unless pos-
sessed of a high degree of technical knowledge, are supposed to see to it that their
contractors comply with the technical aspects of building by-laws. Doubtless,
owner-builders can choose their contractors, and it is incumbent on them to hire
reputable tradesmen. But I fail to see how, having done that, they are in a position
to ensure that construction actually proceeds according to standard. Owner-
builders can hardly be expected to serve as their own inspectors. It can, I think,
safely be assumed that the great majority of those who engage building contractors
170Supra, note 3 at 1266-67. See text accompanying note 92.
171It might be argued that in Rothfield the City did not voluntarily undertake to establish the
code: it was obliged to do so by statute. This obligation, it might be said, creates the obligation.
But this does not follow. There is nothing in the policy-operational distinction, in the concept of
foreseeability, or indeed in the canons of statutory interpretation that transforms a public duty to
regulate the conduct of persons into a duty to those persons when they injure themselves by ignor-
ing the regulations.
1992]
LIABILITY OF PUBLIC AUTHORITIES
to undertake a project must rely on the disinterested expertise of a building inspec-
tor to ensure that it is properly done. In that respect, owner-builders are in a posi-
tion similar to third parties who may be affected by the construction. Like them,
they are, in my respectful opinion, entitled to rely on the municipality to properly
inspect construction to see that it conforms to the standards set out in the muni-
cipality’s building by-laws.1 72
The reliance that La Forest J. considered adequate to form the basis of a duty
of care was not at all like the specific detrimental reliance of the American and
Australian cases. All this passage shows is that the owner-builders would ben-
efit if the City were to supplement the inspections of owner-builders or their
consultants.
First, La Forest J. did not refer to any specific conduct of the City in rela-
tion to the particular plaintiff that would permit the latter to conclude that the
City intended its inspections to “supplant, rather than supplement” the inspec-
tions that prudent owner-builders would conduct on their own behalf.173 The fact
that an owner-builder may not personally possess the technical knowledge is
irrelevant. He or she can, and should, retain architects or engineers to oversee
the construction project, both in their own interests and in the interests of their
neighbours. 74 In other words, the owner-builder had alternative courses open to
ensure that the construction was sufficient. Second, there is no evidence in this
case to suggest that the owner-builders actually relied on the City’s inspections
for their own purposes. In fact, the evidence suggests that the plaintiffs here,
like the plaintiff in Sutherland Shire, regarded the by-laws as regulatory duties
owed by builders to the City rather than as duties owed by the City to builders.
Third, there is no suggestion that the plaintiffs altered their conduct to their own
detriment: there is no suggestion that they would have acted differently in the
absence of the building inspection scheme. Fourth, as La Forest J. himself
observed, “[t]he by-law squarely imposes the duty on the owner at particular
stages of construction.”‘7 By making the City liable to the owner-builders when
the latter fails to ensure that his or her project satisfies the by-laws, the court
shifts the statutory duty without justification. In so doing it makes the “police”
responsible for the conduct of the “offenders.” 176 Finally, and most importantly,
the obligations on the owner-builders may well be predicated on the justifiable
assumption that a person who undertakes the initiative to effect new construc-
172Supra, note 3 at 1267-68.
173See the discussion of Blessing, text accompanying note 71; and Patentas, text accompanying
note 144.
17 That they cannot afford to do so, or that the project is so trivial that it seems to them excessive
to do so, is no reason to shift the consequences of their frugality to the City. Cory J. indicates that
the Manolakoses eventually did hire an engineer, so they were evidently aware that engineers were
available for consulting.
175Supra, note 3 at 1270.
176See text accompanying note 149.
McGILL LAW JOURNAL
[Vol. 37
tions should bear the risks inherent with those constructions, at least as between
themselves and the public at large. 177
For La Forest J., the fact that the owner-builders failed to notify the City
amounted to contributory negligence. Accordingly, he reduced the liability of
the City. If he had adopted a reliance-based test for the existence of a duty of
care he would have seen the owner-builder’s failure to summon the City as evi-
dence that they did not rely on the City. The City owed them no duty of care
because they did not rely on the City. That was the view of the judges in Suth-
erland Shire and, with respect, it seems correct.
b. The Dissent of Justice Cory
The better view is that of Cory J. (Lamer J., concurring), in dissent. Cory
J. saw the question of liability between the City and the owner-builder as a mat-
ter of negligent misstatement. 7′ His decision, then, turned on whether the City
had engendered reasonable reliance, and whether the owner-builder was justi-
fied in relying on the City. He concluded that the owner-builders’ reliance, if
they did in fact rely on the City, was not reasonable. According to Cory J.,
whether the City had a duty of care to the owner-builder had to be determined
according to a two-stage test. 179 The first stage is to see whether the parties were
in a position of proximity in the sense of being within a “‘special relationship’
which ‘would arise in circumstances where the defendant, [was] so placed that
others would reasonably rely on his … statements.'”180 He concluded that,
although the City’s chief inspector was not acting in a private capacity, “there
[was] still such a close relationship that in the reasonable contemplation of Mr.
Phillips, carelessness on his part might cause damage to the Manolakoses.”‘ 181
The second stage was to see whether there were considerations of policy that
ought to negate or diminish the duty of care or the damages that would be recov-
erable upon its breach. Cory J. wrote:
It must be remembered that Mr. Phillips, although a professional engineer, was a
senior employee fulfilling certain specific duties on behalf of the City. He was nei-
’77Richard Buxton says:
The attraction of making the local authority liable appears to spring from its assured
solvency and general responsibility to the public at large, and not because reliance is
placed on the council’s inspection in any particular case by any particular plaintiff. It
is a pity that this rationale, apparently accepted by Lord Wilberforce in Anns, was not
more stringently reviewed in that case, since it is (at least) not self-evident that the rate-
payers should act as insurers to house-purchasers who make bad bargains ….
R. Buxton, “Built Upon Sand” (1978) 41 Mod. L. Rev. 85 at 87. See similar comments, infra, note
181.
178Supra, note 3 at 1287.
179The source of this methodology in Anns is obvious.
I80Supra, note 3 at 1287.
18’Ibid.
1992]
LIABILITY OF PUBLIC AUTHORITIES
ther dispensing advice to the respondents nor was he guaranteeing that the res-
pondents’ home improvement (the retaining wall) would be a success. The respon-
dents had hired a contractor to take care of their interests in that regard. Further,
the granting of a building permit did not and could not relieve the respondent own-
ers of their responsibility to have the work on the retaining wall carried out in
accordance with the City’s by-laws. Nor did Mr. Phillips’ willingness to allow this
relatively small and inexpensive project to proceed without requiring the respon-
dents to incur the cost of obtaining drawings completed by a professional engineer
relieve the respondents of that responsibility.’ 82
Cory J.’s approach has the merit of focusing on the particularities of the rela-
tionship between the City and persons in the place of the plaintiff. The City and
the plaintiff stand in the position of regulator and regulatory target. Like the
courts in the American and Australian specific detrimental reliance cases, Cory
J. concluded that it did not matter whether the City could foresee that the plain-
tiff would be better off if the inspections were made. The point was that the City
inspector was acting in his capacity as a public official enforcing by-laws. The
inspection office was not a publicly funded architecture or engineering service
intended to provide free consultations to displace those that prudent owner-
builders would have conducted on their own behalf.
Despite the merits of Cory J.’s approach, specific detrimental reliance
based on an analogy to the duty to warn provides a more appropriate analytical
scheme than Cory J.’s approach based on the law of negligent misstatement. An
analysis based on negligent misstatement asks whether a person should be held
accountable for what he or she did say. Here the question is whether the City
is liable for something it did not do: make a proper inspection. The first inquires
into the responsibility flowing from a positive act. The second asks whether the
defendant is under a duty to act. The issues are very different. Had Cory J.
employed the reliance test derived from the duty to warn, instead of the reliance
test derived from the law of negligent misstatement, he would not have held that
the City owed a primafacie duty of care, and the regulatory context of the rela-
tionship of the parties would not have been a “policy” element that negates the
primafacie duty. Rather, it would go to show that no private law duty to inspect
arose in the first place.
c. Summary of Rothfield v. Manolakos
The Supreme Court’s decision in Rothfield is questionable in principle and
in policy. The Court appears to believe that it is watching out for the interests
of the small home owner by making the City a virtual guarantor of a warranty
of fitness of new constructions.’ 3 The concern is misplaced. The decision will
1821bid.
183Todd, supra, note 99 at 392 (commenting on the liability imposed in Anns, which was less
strict than the liability imposed in Rothfield). Todd says that the liability in Anns, which resembles
a contractual warranty of fitness, is particularly problematic because unlike true contracting parties
REVUE DE DROIT DE McGILL
[Vol. 37
have the effect of increasing the costs of small projects generally because cities
will be loathe to approve any plans without engineering certificates, or other-
wise to allow any shortcuts in the compliance review process. More fundamen-
tally, there is no good reason why the risk of home improvements should be
shifted from those who undertake them to the residents of a municipality as a
group. The owner-builders are the ones who benefit and they are the ones who
should bear the risk. This is particularly so when homeowners can protect their
own interests by obtaining insurance against third party claims and by selecting
contractors who have demonstrated competence.
The reason the elaborate arguments of the Supreme Court are necessary in
the first place is quite simple: neither the by-laws nor the statute obliging the
municipality to enact the by-laws mentions civil liability anywhere. The Court
therefore had to infer a “common law” duty of care. The Court did not even ana-
lyze the statutory scheme to articulate the reasons for reading in private liability.
As Duff J. stated as early as 1925:
[T]he object and provisions of the statute as a whole must be examined with a
view to determining whether it is a part of the scheme of the legislation to create,
for the benefit of individuals, rights enforceable by action; or whether the reme-
dies provided by the statute are intended to be the sole remedies available by way
of guarantees to the public for the observance of the statutory duty, or by way of
compensation to individuals who have suffered by reason of the non-performance
of that duty.’ 84
This passage suggests an approach very much like the one suggested in Paten-
tas and Blessing.”5 The court should focus on the relationships between the reg-
ulatory targets, the problems the statute seeks to redress, and the means chosen
by the statute for that purpose. By focusing on the relationship between the par-
ties within the context of the regulatory scheme, it is also consistent with
Caparo Industries. The policy-operational distinction and the foreseeability test
ignore these questions.
To argue that owner-builders should not have a right of action against the
City does not imply that there should not be private law liability arising out of
negligent inspections. Rather it is only to argue that owner-builders have no
right of action. The approach suggested by Caparo Industries says that we
should not look at duties “at large,” but look at them in the context of the spe-
cifics of the relationship between the parties. Owner-builders, who are able to
the municipality is not in a position to expressly limit its exposure by the contract, or control the
work of the contractors. See also, Wallace, supra, note 99 (conclusion similar to Todd’s). See also
Buxton, supra, note 177.
‘8Orpen v. Roberts, [1925] S.C.R. 364, 1 D.L.R. 1101 at 1106. See also Ostash v. Sonnenberg
(1968), 67 D.L.R. (2d) 311, 63 W.W.R. 257 (Ont. C.A.); G. Williams, “The Effect of Penal Leg-
islation in the Law of Tort” (1960) 23 Mod. L. Rev. 233.
185Supra, notes 144, 77, respectively.
1992]
LIABILITY OF PUBLIC AUTHORITIES
protect their own interests, should not have a right of action, but it might be oth-
erwise with remote purchasers. Perhaps remote purchasers, who did not partic-
ipate in the construction of the building, can be said to rely on the integrity of
the regulatory scheme. Whether that reliance is reasonable remains an open
question. After all, remote purchasers are also able to protect their interests by
hiring inspection engineers before they purchase.
These issues are complicated, and reasonable persons will differ on how
they should be resolved. The Supreme Court’s decision, however, does not even
acknowledge that these questions exist.
2.
Laurentide Motels v. Beauport
It will be recalled that in Laurentide Motels,”8 6 Beetz J. held that the divid-
ing line between the public law duties and the private law duties of municipal-
ities is itself a matter for public law. The public law of Quebec is derived from
the common law, so the policy-operational distinction applies. Applying (or,
misapplying) the policy-operational distinction in this case, Beetz J. concluded
that the fire department was subject to the private law duties set out in art. 1053
C. C.L. C. It is beyond the scope of this paper to analyze in detail the scope of
the civil law delictual liability of municipalities. I will simply assert my opinion
that the judgment of L’Heureux-Dub J. is inconsistent with authority.”
Let us consider, however, whether the facts in Laurentide Motels would
justify compensation in a jurisdiction which employed the detrimental reliance
analysis based on the duty to rescue suggested in this paper.
If the plaintiff is suing for personal injuries suffered because of negligent
fire-fighting, there may be strong justifications for recognizing such a cause of
action. The case is much less compelling when the plaintiff is suing for property
damage for which insurance is available. When the plaintiff claims that she or
he has relied on the government agency for protection it must be shown that that
reliance was reasonable, taking into account all the circumstances, and that the
conduct of the agency must have been such that the plaintiff was induced to
forego other remedies or precautions against the risk. There is little doubt that
it is reasonable for one person to rely on another to prevent personal injury even
if the rescuer did not create the risk. That message is conveyed by the common
law rescuer cases, by the Quebec Charter of Human Rights and Freedoms, and
by the Vermont Good Samaritan statute reproduced above. 8 The duty to inspect
and regulate, developed by analogy from the duty to rescue, thus focuses nat-
urally on injury to the person. It is another matter altogether where property
186Supra, note 3.
187See supra, note 55.
‘8SReproduced supra, note 136.
McGILL LAW JOURNAL
[Vol. 37
owners seek restoration of their property from a government agency which did
not cause the damage to the property; and where the specific property damage
was insurable. First, it is not at all clear that the communitarian values that
oblige each of us to assist others in personal peril also oblige us to restore, from
the public purse, the asset position of those who have acquired property. Sec-
ond, and more fundamentally, the fact that governments set up fire-fighting ser-
vices should not be regarded as excusing individuals and corporations from pur-
chasing their own insurance, a “remedy or precaution” against the risk of
property loss.
The Caparo Industries approach to the existence of duties of care invites
us to distinguish between categories of plaintiffs and defendants, and the type
of injury suffered. A reliance-based duty to inspect or regulate, derived from the
duty to rescue, may therefore distinguish between types of injury (personal
injury or property loss) and types of plaintiff (individual or corporation). I sub-
mit that in firefighter cases like Laurentide Motels the law may deem it to be
unreasonable to rely on the government for compensation for insurable property
losses.
3.
Just v. British Columbia
In Just, the activities of a roadside inspection crew were found to be oper-
ational in nature, and thus subject to private law duties of care. When analyzing
the private law duty of care, Cory J. immediately adopted the Anns foreseeabil-
ity test. 9 He mentioned in passing the doubts expressed by the House of Lords
in Yuen Kun-yeu,'” but he stated,
[n]evertheless, it is a sound approach to first determine if there is a duty of care
owed by a defendant to the plaintiff in any case where negligent misconduct has
been alleged against a government agency. 19 1
He may, however, have been uncomfortable relying on foreseeability alone,
because he went on to suggest another basis for a duty of care:
In the case at bar the accident occurred on a well used major highway in the prov-
ince of British Columbia. All the provinces across Canada extol their attributes
and attractions in the fierce competition for tourist business. The skiing facilities
at Whistler are undoubtedly just such a magnificent attraction. It would be hard to
imagine a more open and welcoming invitation to use those facilities than that
extended by the provincial highway leading to them. In light of that invitation to
use both the facilities and the highway leading to them, it would appear that, apart
from some specific exemption arising from a statutory provision or established
common law principle, a duty of care was owed by the province to those that use
its highways. That duty of care would extend ordinarily to reasonable maintenance
189Supra, note 3 at 1235.
19Supra, note 43.
191Supra, note 3 at 1235.
19921
LIABILITY OF PUBLIC AUTHORITIES
of those roads. The appellant as a user of the highway was certainly in sufficient
proximity to the respondent to come within the purview of that duty of care. In this
case it can be said that it would be eminently reasonable for the appellant as a user
of the highway to expect that it would be reasonably maintained. For the Depart-
ment of Highways it would be a readily foreseeable risk that harm might befall
users of a highway if it were not reasonably maintained. That maintenance could,
on the basis of the evidence put forward by the appellant, be found to extend to
the prevention of injury from falling rock. 192
The foreseeability theory of the duty of care employed by Cory J. in Just would
have the effect of making the government liable for losses suffered in any of a
myriad of activities in which the government facilitates the private objectives of
residents and visitors.
One hopes that Cory J. was not suggesting that the province, by inviting
persons to use tourist attractions (not even owned or operated by it), became
responsible for the tourists’ safety while they were en route to those attrac-
tions. 3 That explanation is not at all satisfactory. First, as a practical matter, this
explanation leaves unanswered the question about what duty, if any, is owed
people using the road for purposes wholly unconnected with these so-called
provincial inducements. Second, and more fundamentally, the analogy between
persons inviting others onto their private property and a person using public
roads in response to government promotions of tourism is patently false. 4
a. Standard of Care
“Duty of care” and “standard of care” are often spoken of as logically sep-
arate concepts. In practice, however, they are inseparable. One cannot say that
a defendant had a duty of care to do a particular thing, that is, inspect roadside
slopes at certain intervals, without having already defined the scope of the
standard of care. Cory J.’s difficulty in articulating the standard of care casts
doubt on whether the province could have the duty of care he imposed.
Finding, as the Supreme Court did, that the province has a duty to take
“reasonable” steps to ensure that the roads are “safe” throws no light on what
precise steps would be sufficient. “Safety” may mean different things to differ-
ent reasonable persons, especially at the margin. Moreover, the relative impor-
192It is interesting that the British Columbia judges, all of whom were familiar with the Whistler
highway, and who would not have imposed a private law duty of care, stressed the natural causes
of the rock fall. See McLachlin J., supra, note 64 at 350, and Hinkson J.A., supra, note 66 at 224.
For Cory J. the possibility of rock falls was simply a matter of “reasonable maintenance of the
roads.”
193A similar analogy is used later. He suggests that the duty of a person to maintain his or her
driveway is similar to the duty of the province to maintain its highways, although he notes that dif-
ferences in the scope of the duty must be recognized even as we recognize their essential similarity.
Supra, note 3 at 1246.
194See W. Friedmann, “Liability of Visitors to Premises” (1943) 21 Can. Bar Rev. 79.
REVUE DE DROIT DE McGILL
[Vol. 37
tance of various perils is a subject on which reasonable people will differ.
Should money be spent on guard-rails? Should sand barrels be placed in front
of bridge abutments? Should the province check for potential rock falls? Should
more attention be paid to preventing rock falls or removing snow? Should
checks be made on a dangerous portion of road on which only a few people
travel, or a heavily travelled portion which is by its nature safer? Cory J. offered
an example of an obviously bad safety program or, more accurately, an example
where there was no program at all. However, he did not suggest a test for
assessing a program that is not blatantly inadequate.
Cory J. stated in several places that a possible test would be to determine
whether the authorities have wisely spent the money allotted to them. With
respect, this could not provide a workable test. For example, what could consti-
tute the relevant budget? If the province defends by saying that its inspections
were adequate considering that it only had a limited amount allocated to the
rock-scaling budget, could the plaintiff assert that the relevant budget is in fact
the entire maintenance budget, and that within that budget the province
unwisely allocated too little to rock-scaling? Similarly, could the province
defend by claiming that the cut-back in its rock-scaling efforts was necessitated
by a premature exhaustion of its quarterly budget because of the unforeseeable
cost of repairing flood damage? Further, could the plaintiff counter by saying it
was imprudent to so rigidly compartmentalize quarterly budgets, and that a rea-
sonable plan would have allowed flexibility throughout the fiscal year? Obvi-
ously, there are no objective answers to these questions. In practice, courts will
be tempted to call otherwise reasonable programs which they deem inferior
“unreasonable” in order to invoke the power to review this unreasonable deci-
sion. The court would simply be substituting its opinion of what is reasonable
for that of the province. 95
b. Specific Reliance
Perhaps Cory J. did not intend the invitee analogy to be taken literally.
Maybe all he meant was that the government builds and maintains roads, so the
people who use them are entitled to assume that the government will ensure that
rocks will not fall from the mountain slopes onto the road. Could the Justs claim
compensation on this ground under the specific detrimental reliance thesis dis-
cussed above? Probably not. The specific detrimental reliance thesis requires
that: (1) the authority have conducted itself so as to suggest to persons in the
place of the plaintiff that it would take specific steps (for example, continue
195There is precedent for this. Consider the history of judicial review of administrative adjudi-
cative bodies, where decisions of administrative tribunals which might be open to debate among
reasonable persons were consistently held to be “manifestly unreasonable” so as to open the door
to judicial review. See Y-M. Morissette, “Le contrrle de la compdtence d’attribution: th~se, anti-
these, et synthese” (1986) 16 R.D.U.S. 591.
19921
LIABILITY OF PUBLIC AUTHORITIES
warnings); (2) that it is reasonable for the plaintiff to rely on the authority to
take those steps; and (3) that the plaintiff has altered his or her conduct in reli-
ance on the anticipated conduct of the authority. First, the highway has signs at
the beginning of the mountainous passage, and at frequent intervals along the
way, indicating the danger of falling rocks. It cannot be asserted that the prov-
ince invited drivers to rely that they would be safe from falling rocks. Second,
in a case like Just, it would be difficult to argue that the province had by its con-
duct induced persons to rely on it taking specific steps: its commitment to safety
generally along an entire highway is not analogous to the Coast Guard’s com-
mitment to light a specific lighthouse at a specific place in the Mississippi
river.196
Many people would feel that denying recovery to plaintiffs in the position
of the Justs is hard-hearted. If the specific reliance thesis cannot accommodate
such compensation, it ought to be set aside in favour of an alternative that better
reflects the communitarian values of a progressive society. I cannot dispute this
intuition. It may be that we are not satisfied with the corrective justice of neg-
ligence law and expect something more. As John Finnis has written:
Consider the common law of torts. … [T]his set of rules and principles was
regarded as an instrument of justice, indeed of corrective justice in Aristotle’s
sense. … In [a more recent view] the question is not “What are the standards of
conduct which one person must live up to in relation to his ‘neighbours’?” or
“What should be the extent of liability of one who fails to live up to those stand-
ards of conduct?,” or even “How should someone injured by the wrong of another
be restored to his former condition?” Those are questions central to the theory of
commutative justice. But in the newer view, the question is held to be “How
should the risks of common life be apportioned, especially the risks of essentially
collaborative enterprises as travel and traffic by road?” Injury to one of the partic-
ipants is then treated as an incidental loss to be set against the gains which accrue
to all who participate in this sphere of common life. … The question whether the
injury was caused by any fault becomes substantially irrelevant.197
If road travel is indeed an area where “common enterprise” justice ought to pre-
vail it would be better if that were stated outright. The present approach of the
Supreme Court, while intending to “do justice” in an individual case like Just,
does a disservice to the development of negligence law generally and the liabil-
If we are willing to adopt the com-
ity of government agencies in particular.’
mon enterprise thesis we must then decide which activities qualify as common
enterprises, and what sorts of classes of plaintiffs can claim compensation for
which injuries. For example, it does not seem unreasonable to require private
drivers and commercial carriers to insure their vehicles and cargoes, nor does
196See Indian Towing, supra, note 11.
1971j.M. Finnis, Natural Law and Natural Rights (Oxford: Clarendon Press, 1980) at 180 (empha-
198Sopinka J.’s decision reflected similar concerns, see above, subsection I.D.2.c.
sis added to final sentence).
McGILL LAW JOURNAL
[Vol. 37
it seem unreasonable to prevent their insurance companies from passing on
claims to the general public via suits against the government. Insurable perils
of this sort might, then, be ineligible for assistance based on “common enter-
prise” justice or other communitarian values. But as long as the true basis of
compensation is concealed behind the policy-operational and foreseeability
tests, these questions will not receive the attention they deserve.
It may be that Canadian society expects governments to cushion the blow
of tragedies like that which befell the Just family. It is not inconceivable that the
members of a progressive society would expect the state to compensate the vic-
tims of natural calamities arising from the common enterprises of life. If that is
so, rules will have to be developed to segregate justified claims from the unjus-
tified. Those rules would not be based on the fault of the government, but on
the victim’s loss. Whether such compensation is well founded is an open ques-
tion. But it is clear that it should not be implemented through the back door by
bending the rules of negligence law.
c. Summary of Just
The decision of Cory J. in Just illustrates how the policy-operational dis-
tinction combines with the foreseeability thesis of the duty of care to impose
novel private law duties on government without explicit justification or expla-
nation. Using the policy-operational distinction as a test for immunity implies
that government activity is prima facie subject to private law duties. The con-
cept of discretion, upon which the policy-operational distinction is based, is so
vague that it invites circular reasoning. As a result, the presumption that the spe-
cific government activity, here, roadside safety inspections, can be judged by
private law standards becomes virtually irrebuttable.
If Cory J. had employed the approach suggested by Brennan J. in Suther-
land Shire and Lord Bridge in Caparo Industries, he would have had to take a
second look at the relationship between the government agency and the specific
plaintiffs. He would have recognized that the province’s role as facilitator of
transportation generally has no persuasive analogy to the activities of persons. 99
He would also have recognized that the duty he would impose on the province,
a duty to prevent injury caused by nature, is not one normally imposed on per-
sons in favour of the public at large. In short, the false presumption created by
the policy-operational distinction would have been rebutted: he would have con-
cluded that government in this case cannot be judged by ordinary private law
199One might be tempted to compare the government as road builder and maintainer to private
corporations which act as common carriers. The analogy is false. The relationship between a com-
mon carrier and a passenger (or freight shipper) is contractual. The common carrier is bound by
the terms of the contract, and charges a fee sufficient to carry out the terms of the contract.
1992]
LIABILITY OF PUBLIC AUTHORITIES
duties. Of course, Cory J. may have gone on to conclude that the government
should compensate persons in the position of the plaintiffs in this case, but he
would have been obliged to say why they should be compensated, and he would
have had to outline the limits of that compensation.
Instead, Cory J. relied on the Anns foreseeability test. Like the discretion-
based policy-operational distinction, the foreseeability test invites circular rea-
soning. When one looks with hindsight at an accident that has actually hap-
pened, it is easy to jump to the conclusion that it should have been foreseen.
Having concluded, perhaps erroneously, that the accident was foreseeable, it
seems to follow that someone should have done something to prevent it. From
there, fingers naturally point to the government. The result is an ad hoc asser-
tion that the government ought to compensate persons who are injured by nat-
ural forces while using public highways.”
Conclusion
Lord Atkin saw tort law as the legal expression of the moral injunction,
“Love thy neighbour.”‘ The ideas of love (or more prosaically, mutual respect)
and neighbourhood are distinctly human. The concepts of foresight and proxim-
ity, through which the moral maxim is translated into legal rules, are no less
human. People “foresee” the consequences of their acts, and have “relation-
ships” with other people. We try to make corporations human by calling them
“persons.” We do not go so far with government, but we say that it should be
held civilly responsible as if it were a person. We say, by analogy, that govern-
ment agencies can “foresee,” but we really mean that it is possible to create
administrative procedures which assign to a member of the bureaucracy the task
of effecting or avoiding certain results. We say that the government is in a rela-
tionship of proximity with a person. But a government agency is merely a polit-
ical concept. What we really mean is that the government has come to affect the
person in a way that sets that person apart from the rest of society.
The policy-operational distinction and the use of the foreseeability test for
public agencies appear to affirm the precept that governments can, in principle,
be treated as if they were human. “Policy” seems abstract and legislative, while
“operational” evokes the image of actual people doing things. And as I have
observed, attributing foresight to a government agency is anthropomorphic.
In practice, however, these two tools have allowed the Supreme Court of
Canada to impose liability on government agencies that has no analogue in the
duties owed by individuals to their neighbours. Discretion, which is said to
2See supra, note 64.
20 Supra, note 99.
REVUE DE DROIT DE McGILL
[Vol. 37
divide policy decisions from operational functions, is hopelessly vague. The
distinction yields only ad hoc assertions. The Supreme Court has increasingly
chosen to find that the government’s activities are operational. With the benefit
of hindsight courts can find “foresight” whenever they feel the result is
desirable.
The result is incoherence. Whether one agrees with the results or not, the
decisions must be criticized because it is impossible to say how the results came
about or, more importantly, what will happen in future cases. In addition, gov-
ernment liability cases affect tort liability generally. Consider the effect of Anns
in the jurisprudence. The broad-based foreseeability approach is as problematic
in many cases involving actual people as it is in cases where the defendant is
a government agency. The repetition of the Anns test in the Supreme Court tril-
ogy obstructs the development of more principled bases of civil liability.
It is likely that some members of the Supreme Court believe the analogy
to human activity to be too limiting. We expect more of our government than
we expect of our human neighbours. Imposing liability on the government when
it fails in its role as public guardian is consistent, they might say, with commu-
nitarian social values. In the abstract, there is considerable merit to that argu-
ment. But it is by no means clear that the ad hoc creation of duties practiced by
the Supreme Court has achieved or will achieve the values of an increasingly
communitarian society. Payments from the public purse can go to the undeserv-
ig as easily as to those in genuine need. If the Supreme Court believes that tort
law should recognize new duties to compensate, it should articulate why the
plaintiffs deserve compensation. The policy-operational distinction and the fore-
seeability test ignore the plaintiffs. It does not distinguish those who have the
resources to protect their own interests from those who are vulnerable and thus
have a greater need for the protection of government.
The reconsideration of the concept of proximity proposed by the Australian
High Court in Sutherland Shire and the House of Lords in Caparo Industries
would permit consideration of these questions. Proximity and the duties that
arise from it are a matter of context and individual circumstances. New duties
can be created if they can be justified by analogy to existing duties. It is there-
fore not necessary to limit new duties to ones which persons already owe to one
another. But if new duties are to be created, the basis for the duty must be artic-
ulated expressly.
A duty to inspect and regulate, based on specific detrimental reliance, is a
type of duty that has been created based on close examination of all the circum-
stances in which the parties came together. It is consistent with the goal of
1992]
LIABILITY OF PUBLIC AUTHORITIES
143
imposing liability on government agencies “as if’ they were persons, but it is
not unduly restricted by that goal. If the policy-operational distinction and the
foreseeability test are laid to rest, other similar duties may be recognized.
