Of Forest Fires and Systemic Discrimination:
A Review of British Columbia
(Public Service Employee Relations
Commission) v. B.C.G.S.E.U.
Colleen Sheppard,
This case comment addresses the recent contributions
to human rights law developed in the Supreme Court of
Canada’s decision British Columbia (Public Sen’ice Em-
pLoyee Relations Commission) v. B.C.G.S.EU. The Court
held that the aerobic standard for evaluating the fitness of
forest firefighters was discriminatory towards women. The
Court ordered the reinstatement of Tawney Meiorin, a fe-
male forest firefighter who had lost her employment by
reason of failing the mandatory provincial fitness testin&
The author maintains that the Court significantly advances
human rights analysis by articulating a unified approach to
human rights defences that is not premised on any prelimi-
nary classification of the discrimination as either direct or
adverse effect. The Court also highlights the importance of
an employer’s duty to accommodate as an integral dimen-
sion of equality. The author suggests, nonetheless, that
further elaboration of certain aspects of discrimination law
will be required in future cases. More specifically, the con-
cept of adverse effect discrimination should be retained and
clarified to ensure that hidden and institutionalized forms
of inequality are identified and remedied. Furthermore,
there remains a need to ensure that discriminatory stan-
dards, rules, or policies are fully scrutinized and potentially
revised before assessing individual accommodation strate-
gies. Finally, the approach to health and safety risks in the
context of human rights adjudication deserves further dis-
cussion.
Cett chmnique aborde les contributios r,1-cn=- am-
drolts de la personne dlhbors par la Ceur supz~Er du
Canada dans la dLcision Co!vmb k-ia-nniTe (PWblic
Senice Enplovce Relations Cemmioien) c. BCG.SEU.
La Cour a conclu qu’ura des nnres mininrales de
condition physique d..blies par J gonm===uant pour scs
pompiers forestiers, dtait discriminatoire 4 I’dard des
femmes. Elie a par consrquzat ordoann ha rSwrgmration de
Tawney Meiorin. une pomp:,re forestihre qui avait &.t
cong&lide apr s avoir &lioul Ic test obligatoire. L’auteura
soutient que la Cour a 2ppod unre contribution signfimii:v
au dlveloppement de l’analse des droits dz la porA=sn
en
adoptint une approche unifiie qui net pas b=& sur u=
classification de la discriminton dans la c
.anroz d la
discrimination diirecte.> on dans cele d- la &irimion
.,par suite d’un effet prn~judiciab!e.
insi sm
dgalement sur limpoitance de l’obliption d’=conmmder
de l’emplo-eur en tant qua pattie intrgranroz du droit a
l’dgalit, Lauteure suggre toutefois qu’il s
r.%-c ire.
dans des d&isions futures, d’d&.borer devanmge certains
concepts du droit relatif h la discrimina tion: normn’.n
la
notion de discrimination par suite d’rn effet Vjrud.Eab!-
re retenue et &arifi e pour ponnttre didmntifier et
deait
d’6liminer des
intiutio:nalibes
d’in-galit& Il faut dgalement skasterer quo Its
rIgles on politiques discrimmaoires s *=et minuie n.mon
analysis et 6entuellement rdvis s avat que la pertinence
suit
de strwZ-gies d’accommodment
dvalue. Par ailleurs. I’approclm rmr- nx
qulm do
am
sanrt, et de s&duitd dans l coatexte des d
.cisions do droit do
la personne appelle une discussion plus approforffie.
irdiidellzs no
La Cour
formes cac
s et
“ar:Aards.
“Faculty of Law, McGill University. The author would like to thank Alison Gray for her research
assistance and Matthew Garfield for his helpful comments on an earlier draft of this comment.
McGill Law Journal 2001
Revue de droit de McGill 2001
To be cited as: (2001) 46 McGill W. 533
Mode de rf6rence: (2001) 46 RD. McGill 533
534
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
Introduction
I. Facts and Judgments
II. A Unified Approach and an Extended Duty to Accommodate
Ill. The Continuing Relevance of Adverse Effect Discrimination
A. Eliminating Intent from Discrimination Analysis
B. When Are Disparate Effects Discriminatory?
C. Recognizing Different Types of Adverse Effect Discrimination
IV. The Duty to Accommodate versus Institutional Change
V. Assessing Health and Safety Risks in Human Rights Cases
Conclusion
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.O.G.S.E.U.
535
Introduction
The case of British Columbia (Public Service Employee Relations Commission) v.
British Columbia Government and Services Employees’ Union (B.C.G.S.E.U.)’ in-
volved allegations of gender discrimination against women working in a domain tra-
ditionally occupied by men, forest firefighting. Tawney Meiorin challenged the valid-
ity of an employment fitness test on the grounds that it resulted in sex discrimination.
In exploring the issues of systemic discrimination raised in this case, forest firefight-
ing can provide some insights. Forest firefighters know that addressing a problem
while it is small is fundamental to success, that vigilance is critical, that fighting fires
cannot be done alone, and that seemingly small human errors can cause unthinkable
harm and damage. Systemic discrimination-pervasive and embedded in institutional
practices and policies-represents the forest fire that has not been prevented. And yet,
while we all recognize the urgency of stopping forest fires, there is no similar urgent
call to quell systemic inequality. It seems apt therefore that the Supreme Court of
Canada should elaborate some critical ideas about systemic discrimination and human
rights law in the context of forest firefighting.
In this case comment I review the significant and positive contributions to human
rights law developed in the Meiorin case. The most notable contribution that the case
makes to human rights law is the articulation of a unified approach to human rights
defences which applies regardless of whether the discrimination is direct or indirect.
The Court effectively eliminates the confusion and unfairness that sometimes resulted
from the previous bifurcated approach. While Meiorin is to be applauded for inte-
grating the law of human rights defences, it raises new concerns that deserve consid-
eration. I highlight three aspects of the judgment that remain in need of further elabo-
ration or clarification. First, I consider the continuing relevance of adverse effect dis-
crimination, despite the Court’s rejection of the relevance of classifying discrimina-
tion for the purpose of deciding which human rights defences and remedies are avail-
able. Second, I examine the interplay between individual accommodation and a trans-
formative and substantive approach to equality. It is important to ensure that the in-
stitutional change the Court endorses in Meiorin is not inadvertently overlooked in the
subsequent application of its new integrated test. Third, I discuss the appropriate stan-
dard for assessing health and safety risks. A review of the Court’s application of the
integrated test in its subsequent decision, British Columbia (Superintendent of Motor
Vehicles) v. British Columbia (Council of Human Rights), reveals that the issue of
risk remains complex and unresolved. It also raises questions about the clarity of the
‘[1999] 3 S.C.R. 3, 176 D.L.R. (4th) 1 [hereinafterAfeiorin cited to S.C.R.].
2 [1999] 3 S.C.R. 868, 181 D.L.L (4th) 385 [hereinafterGrismercited to S.C.R.].
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
new test itself in terms of whether the more expansive applicability of the duty to ac-
commodate actually promotes a more substantive and transformative approach.
I. Facts and Judgments
Tawney Meiorin was employed by the Province of British Columbia as part of an
Initial Attack Forest Fighting Crew. The crew’s responsibility was to suppress forest
fires while they were still small and relatively easy to contain. Ms. Meiorin worked on
the crew for three years and had received positive job evaluations. As a result of a
coroner’s report, the province decided to introduce a series of fitness tests for forest
firefighters in 1993. The fitness tests included a running test, an upright rowing exer-
cise, and a pump carrying/hose dragging exercise. Each test was to be completed
within a stipulated time. As part of the mandatory fitness testing, Ms. Meiorin was re-
quired to run 2.5 kilometres in eleven minutes. She failed to do so despite four at-
tempts. The running test was designed to measure aerobic capacity.’
As a result of her failure to succeed in the fitness tests, Ms. Meiorin lost her em-
ployment. Her union then filed a grievance on her behalf claiming discrimination on
the basis of sex.’ The union argued in particular that the aerobic standard constituted a
form of adverse effect discrimination. The union suggested that women should be
tested according to a lower aerobic standard. The union also maintained that even
though Ms. Meiorin did not pass the running test, given her positive work appraisals,
her experience on the job, and the fact that she almost met the required standard she
should be accommodated and reinstated to a firefighting position, or alternatively, re-
located to some other job within the ministry.’
Deciding in Ms. Meiorin’s favour, the arbitrator concluded that the aerobic stan-
dard resulted in adverse effect discrimination because of its disproportionately nega-
tive effect on women. He also held that the Province of British Columbia had failed to
prove that accommodating Ms. Meiorin by continuing her employment would pose a
safety risk to herself, her co-workers, or the public. The employer, therefore, had not
satisfied its duty to accommodate Ms. Meiorin, and the arbitrator ruled that she should
be reinstated.’
The B.C. Court of Appeal quashed the arbitrator’s award and concluded that the
province had not discriminated against Ms. Meiorin. The court maintained that the ar-
bitrator had concluded that the running test was necessary to the safe and efficient
3Meiorin, supra note 1. See review of the facts at paras. 4-12.
4 See British Columbia (Public Service Employee Relations Commission) and B.C.G.S.E.U.
(Meiorin) (1996), 58 L.A.C. (4th) 159 (B.C.), Chertkow.
‘For a summary of the union’s arguments, see ibid. at 161.
6Ibid. at 208.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.C.G.S.E.U.
537
performance of the work. The court also concluded that Ms. Meiorin had been ac-
commodated because she had received individualized testing. Finally, the court noted
that to lower the aerobic standard for women only would result in reverse discrimina-
tion against male employees
The Supreme Court of Canada unanimously allowed the appeal, concluding that
the aerobic standard was discriminatory and that the government had failed to prove
that it was necessary to ensure the safe performance of the work. The Court therefore
ordered the reinstatement of Ms. Meiorin with compensation for lost wages and bene-
fits.’ In making its decision, the Supreme Court took the opportunity to reassess its
previous approach to the key concepts of human rights defences, most specifically the
bona fide occupational requirement and the duty to accommodate. It is in its elabora-
tion of these two concepts that the judgment makes its most significant contribution to
human rights law.
II. A Unified Approach and an Extended Duty to Accommodate
Prior to Meiorin, the Court had articulated separate human rights defences de-
pending on whether the alleged discrimination was direct or adverse effect. For cases
involving direct discrimination, the applicable human rights defence was the bona fide
occupational requirement or qualification (“BFOR” or “BFOQ”)? Pursuant to the Su-
preme Court’s earlier jurisprudence,
To be a bona fide occupational qualification and requirement a limitation …
must be imposed honestly, in good faith, and in the sincerely held belief that
such limitation is imposed in the interests of the adequate performance of the
work involved [the subjective element] … In addition it must be related in an
objective sense to the performance of the employment concerned, in that it is
reasonably necessary to assure the efficient and economical performance of the
job without endangering the employee, his fellow employees and the general
public [the objective element].”
In applying this test the Court had also found that where individualized testing was
feasible, employers should not rely on general standards to exclude categories of
7 British Columbia (Public Senice Employee Relations Commission) v. B.C.G.S.EU. (1997), 149
D.L.R. (4th) 261,37 B.C.L.R. (3d) 317 (C.A.).
‘ Meiorin, supra note 1.
9 The emergence of a bifurcated approach can be traced to Wilson J.’s decision in Central Alberta
Dairy Pool v. Alberta (Human Rights Conunission), [1990] 2 S.C.R. 489, 72 D.LR. (4th) 417 [ere-
inafterAlberta Dairy Pool cited to S.C.R].
” Ontario (Human Rights Comnission) v. Etobicoke (Borough ol), [1982] 1 S.C.R. 202 at 203, 132
D.LR. (3d) 14 [hereinafter Etobicoke cited to S.C.R.].
MCGILL LAW JOURNAL / REVUE DE DROITDE MCGILL
[Vol. 46
workers.” Of note is the absence of any duty to accommodate individual employees
pursuant to the BFOQ test. 2
In cases of adverse effect discrimination, the BFOQ defence was held not to ap-
ply. Instead, an employer was required to show “(1) that there is a rational connection
between the job and the particular standard, and (2) that it cannot further accommo-
date the claimant without incurring undue hardship:”‘ Wilson J. suggests in Alberta
Dairy Pool that, in most adverse effect discrimination cases, the neutral standard
should remain in place and individual employees adversely affected by the standard
should be accorded special accommodation.”
Numerous critiques were made of this bifurcated or conventional approach.
McLachlin J. (as she then was) reviews these critiques in outlining what she refers to
as “seven difficulties with the conventional approach:”‘ One of the most fundamental
components of McLachlin J.’s analysis is her critique of the malleability of the direct
versus adverse effect discrimination distinction. In many cases it is possible to con-
ceptualize a problem as either direct or indirect discrimination. The example
McLachlin J. uses is a mandatory pregnancy test before commencing employment.
One could argue that the employment test constitutes a form of direct discrimination
against women. Alternatively, one could argue that the employment test “is a neutral
rule because it is facially applied to all members of a workforce and its special effects
” See Saskatchewan (Human Rights Commission) v. Saskatoon (City o), [1989] 2 S.C.R. 1297 at
1313-14, 65 D.L.R. (4th) 481.
2Me idea that the duty to accommodate was conceptually inconsistent with the BFOQ emerged in
McIntyre J.s judgment in Bhinder v. Canadian National Railway Co., [1985] 2 S.C.R. 561 at 589,23
D.L.R. (4th) 481 [hereinafter Bhinder cited to S.C.R.]:
[Where s. 14(a) applies, the subsection in the clearest and most precise terms says that
where the bonafide occupational requirement is established, it is not a discriminatory
practice. To conclude then that an otherwise established bona fide occupational re-
quirement could have no application to one employee, because of the special charac-
teristics of that employee, is not to give s. 14(a) a narrow interpretation; it is simply to
ignore its plain language.
See also Large v. Stratford (City o), [1995] 3 S.C.R 733, 128 D.L.R. (4th) 193.
“Meiorin, supra note 1 at para. 22.
“Supra note 9 at 505-506. In some jurisdictions (e.g. Ontario, the federal government), an inte-
grated approach was required by legislative reform in the wake of Bhinder, supra note 12.
” Meiorin, supra note 1 at para. 26. The seven difficulties reviewed in paras. 27-49 are (a) the artifi-
ciality of the distinction between direct and adverse effect discrimination; (b) different remedies de-
pending on the method of discrimination; (c) the questionable assumption that the adversely affected
group is always a numerical minority; (d) difficulties in the practical application of employers’ de-
fences; (e) the legitimation of systemic discrimination; (0 the dissonance between conventional
analysis and the express purpose and terms of the human rights code; and (g) the dissonance between
human rights analysis and analysis under the Charter, infra note 22.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.C.G.S.E.U.
539
on women are only incidental” 6 McLachlin J. proceeds to point out that the malle-
ability of the distinction between direct and adverse effect discrimination is particu-
larly troubling from a human rights perspective because, pursuant to the conventional
approach, the initial classification had significant repercussions in terms of the nature
of the employer’s defences and the kinds of remedies available to the claimant. As
noted above, if the discrimination were considered direct, the BFOQ analysis applied
and there was no duty to accommodate. If it were adverse effect discrimination, the
BFOQ analysis was not applicable and the employer had a duty to accommodate to
the point of undue hardship.
Both defences were critiqued by human rights advocates for different reasons.
With respect to the BFOQ defence in direct discrimination cases, it was criticized for
not imposing on employers any duty to accommodate.” An employment policy that
overtly excluded individuals from certain groups, therefore, could be upheld provided
it passed the subjective and objective prongs of the BFOQ test. The employer was not
obligated to find creative and accommodating solutions to make the workplace more
inclusive for individuals. Accordingly, if, for example, an individual was directly dis-
criminated against on the basis of his or her disability, the employer would be under
no obligation to accommodate the individual’s needs provided the BFOQ test was
satisfied.’ The accommodation of differences was increasingly recognized as a central
component of substantive equality in other contexts, and yet it was not applied to
cases of direct discrimination.
The defence set out for adverse effect discrimination was also critiqued for pre-
cisely the opposite reason. The focus of the critique was that the preliminary rational
connection requirement was too lenient and that the analysis shifted too quickly and
exclusively to individual accommodation issues.” Earlier jurisprudence appeared to
support a presumption that where the discrimination resulted from the adverse effects
of a facially neutral policy or practice, in most cases the appropriate response would
be to leave the policy or practice in place while providing accommodation to the indi-
vidual or group affected. The major justification for this presumption was the belief
that in most cases the group adversely affected would be a minority group, as demon-
Commission” (1993) 1 Can. Lab. W. 283.
16 Meiorin, ibid. at para. 27.
‘7 See eg. M.FI Yalden, “The Duty to Accommodate-A View from the Canadian Human Rights
” See discussion in A. Molloy, “Disability and the Duty to Accommodate” (1993) 1 Can. Lab. U.
23; M. Lynk, “Accommodating Disabilities in the Canadian Workplace” (1999) 7 Can. Lab. U. 183.
‘9 See e-g. MID. Lepofsky, ‘The Charter’s Guarantee of Equality to People with Disabilities-How
Well Is It Working?” (1998) 16 Windsor YB. Access Just. 155; S. Day & G. Brodsky. “The Duty to
Accommodate: Who Will Benefit?” (1996) 75 Can. Bar Rev. 433; B. Etherington, “Central Alberta
Dairy Pool: The Supreme Court of Canada’s Latest Word on the Duty to Accommodate” (1993) 1
Can. Lab. U. 311.
540
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
strated in many of the cases involving religious-based discrimination.” In addition, it
was generally presumed that the policy, rule, or standard should continue to apply to
the majority. Thus, in a number of cases, the neutral rule, policy, or practice that
caused the adverse effect discrimination was not subject to any real or effective scru-
tiny. The analysis focussed instead on whether the individual could be accommodated
despite the continued applicability of the policy, practice, or standard to others. As
McLachlin J. suggests, this tendency in the jurisprudence reinforced the institutional
status quo, legitimizing, in effect, the sources of systemic discrimination:
Under the conventional analysis, if a standard is classified as being “neu-
tral” at the threshold stage of the inquiry, its legitimacy is never questioned. The
focus shifts to whether the individual claimant can be accommodated, and the
formal standard itself always remains intact. The conventional analysis thus
shifts attention away from the substantive norms underlying the standard, to
how “different” individuals can fit into the “mainstream”, represented by the
2′
standard.
While such an approach ensured that accommodation would be considered and re-
quired to the point of undue hardship, the defence seemed to presume the validity of
the underlying policies, practices, or standards.
Beyond reviewing the inadequacies of each of the tests in terms of substantive
human rights law, McLachlin J. also suggests that the tests were manipulated to the
point where there was considerable confusion about their proper interpretation and
application. Further, she suggests that the equality rights analysis emerging with re-
spect to the Canadian Charter of Rights and Freedoms’ does not rely on a prelimi-
nary categorization of the discrimination as either direct or indirect. The focus instead
is on the harmful effects of the discriminatory practice or law.
The significant difficulties enumerated by McLachlin J. convinced her of the need
to articulate a new integrated test. Accordingly, she sets out a single test that applies to
cases of both direct and adverse effect discrimination and that appears to embrace the
positive features of both defences. The three-step test requires an employer to prove
the following on a balance of probabilities:
(1) that the employer adopted the standard for a purpose rationally connected
to the performance of the job;
‘0 See e.g. Alberta Dairy Pool, supra note 9; Commission scolaire rdgionale de Chambly v. Ber-
gevin, [1994] 2 S.C.R. 525, 115 D.L.R. (4th) 609.
2 Meiorin, supra note 1 at para. 40. McLachlin J. suggests that not scrutinizing a neutral standard,
rule, or practice because it likely affects only a minority is neither a convincing nor a principled ap-
proach. In addition to being “easily manipulable”, it should not justify a lesser level of scrutiny,
2′ Part I of the Constitution Act, 1982, being Schedule B to the Canada Act, 1982 (U.K.), 1982, c.
11 [hereinafter Charter].
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATIN IN B.C.G.S.E.U.
541
(2) that the employer adopted the particular standard in an honest and good
faith belief that it was necessary to the fulfilment of that legitimate work-
related purpose; and
(3) that the standard is reasonably necessary to the accomplishment of that le-
gitimate work-related purpose. To show that the standard is reasonably
necessary, it must be demonstrated that it is impossible to accommodate in-
dividual employees sharing the characteristics of the claimant without im-
posing undue hardship upon the employer
One can see in this test the remnants of the Etobicoke” test supplemented by an inte-
grated duty to accommodate and a preliminary rational purpose test.’ Accordingly, in
cases of either direct or indirect discrimination, employers are required to prove the
validity of the standard or policy itself, and in so doing, to show that individual ac-
commodation would impose undue hardship.
III. The Continuing Relevance of Adverse Effect Discrimination
The first concern raised by Meiorin is one of potential misinterpretation. The
Court emphasizes how it is problematic to premise available human rights defences
on the initial classification of discrimination as direct or adverse effect, when the clas-
sification itself is malleable. And yet there is a risk that McLachlin J:s judgment will
be interpreted as rejecting the relevance of the conceptual distinction between adverse
effect and direct discrimination. This would be a mistake and a step backwards in our
understanding of human rights issues. Recognition of adverse effect discrimination
represents a powerful and important breakthrough in human rights law, and we should
be careful not to jettison the concept too quickly simply because it overlaps with di-
rect discrimination in some contexts.
The potential conclusion that the distinction between direct and adverse effect
discrimination is no longer pertinent is reinforced by the Supreme Court’s decision in
Grismer, where McLachlin J. writes, “Meiorin announced a unified approach to adju-
dicating discrimination claims under human rights legislation. The distinction be-
tween direct and indirect discrimination has been erased.’: There is no doubt that the
distinction is no longer relevant to the question of which human rights defence is to be
applied. But it does not necessarily follow that the distinction is no longer relevant to
2’Meiorin, supra note 1 at para. 54.
2 Supra note 10.
25The test set out by McLachlin J. is similar to that suggested by Dickson CJ.C. in his dissenting
reasons in Bhinder, supra note 12. It also resonates with the decision of Beetz J. in Brossard (Town aoJ)
v. Quebec (Conwdssion des droits de la persomne), [1988] 2 S.C.IL 279, 53 D.LR. (4th) 609. The
first prong of the test seems inspired in part by the test set out in R. v. Oakes, [1986] 1 S.C.R. 103 at
138-40,26 D.L.R. (4th) 200.
26 Sup ra note 2 at para. 19.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
anti-discrimination law. Indeed, the distinction remains critically important for identi-
fying the adverse effects of apparently neutral policies, practices, and procedures on
the socially disadvantaged groups accorded protection against discrimination in hu-
man rights documents. Moreover, despite its path-breaking recognition of adverse ef-
fect discrimination in Ontario (Human Rights Commission) v. Simpsons Sears,’ the
Court has not provided any significant elaboration of how to identify adverse effect
discrimination in its more recent jurisprudence.28 Rather than moving away from the
albeit challenging task of elaborating on the concept of adverse effect discrimination,
more in-depth legal analysis of the concept is needed.
Why is it so critical to expand on our understanding of adverse effect discrimina-
tion? If we do not, there is a significant risk that discrimination embedded in appar-
ently neutral institutional policies, rules, or procedures will not be recognized as dis-
criminatory. This risk is accentuated by the necessity in anti-discrimination law to
connect the experience of exclusion, harm, prejudice, or disadvantage to a recognized
ground of discrimination. Thus, for example, it is not enough to claim that one is ex-
cluded because of a height and weight standard; it is necessary to link the apparently
neutral standard to the recognized ground of sex or race discrimination.’ While the
analogous grounds in subsection 15(1) of the Charter mean that direct discrimination
against domestic workers, for example, could be recognized as constituting discrimi-
nation against a new analogous ground (that of domestic workers), adverse effect
analysis allows us to make the conceptual link between discrimination against do-
mestic workers and sexism and racism.” While making that conceptual link may be
self-evident in some cases, in others it will be invisible and unacknowledged, espe-
cially for those who do not experience the disadvantage. We need a sophisticated and
coherent theory of adverse effect discrimination to assist claimants, lawyers, and ad-
judicators with the complexities of the manifestations of systemic discrimination.
While I do not endeavour here to articulate a complete theory of adverse effect dis-
crimination, I offer a few preliminary ideas about potential avenues for clarification.
[198512 S.C.R. 536,23 D.L.R. (4th) 321 [hereinafter O’Malley].
28 See D. Pothier, “M’Aider, Mayday: Section 15 of the Charter in Distress” (1996) 6 NJ.C.L. 295.
Nor has the U.S. Supreme Court provided significant guidance on how to identify disparate impact
discrimination. See e.g. J.D. Cummins, “Refashioning the Disparate Treatment and Disparate Impact
Doctrines in Theory and in Practice” (1998) 41 How. L.J. 455; R. Belton, “The Dismantling of the
Griggs Disparate Impact Theory and the Future of Title VII: The Need for a Third Reconstruction”
(1990) 8 Yale L. & Pol’y Rev. 223; C. Sheppard, “Equality in Context: Judicial Approaches in Can-
ada and the United States” (1990) 39 U.N.B.LJ. 111.
29 See e.g. discussion of Re Complaint of Ann J. Colfer (12 January 1979) (Ont.) in B. Vizkelety,
Proving Discrimination in Canada (Toronto: Carswell, 1987) at 47-48.
‘0 See discussion in Egan v. Canada, [1995] 2 S.C.R. 513 at para. 78, 124 D.L.R. (4th) 609,
L’Heureux-Dub6 J. See also discussion, below.
2001]
C. SHEPPARD- SYSTEMIC DISCRIMINAT7ON IN B.C.G.S.E.U.
543
A. Eliminating Intent from Discrimination Analysis
First, it is necessary
to dispense with the intent problem. When anti-
discrimination laws first emerged, the general understanding was that discrimination
was rooted in the intentional exclusion or mistreatment of individuals on the basis of
their group membership. Indeed, early forays into the legal prohibition of discrimina-
tion relied on a criminal law model, including the requirement of a mens rea or a dis-
criminatory intent? This approach to anti-discrimination law was premised on deeply
rooted ideas in criminal law and tort law that individuals should only be held legally
accountable if they are at fault or if they intentionally cause harm to others. When the
concept of adverse effect discrimination first emerged, it ushered in a growing recog-
nition of the possibility of legal accountability based on the victim’s experience of
harm even in the absence of any intent to discriminate. Human rights law was to be a
remedial law that compensated victims of discrimination. Its focus diverged from its
criminal law antecedents and shifted away from the earlier perpetrator perspective. As
Dickson CJ.C. notes in Action travail des fenunes, “The purpose of the [Canadian
Human Rights] Act is not to punish wrongdoing but to prevent discrimination:”
Adverse effect discrimination, as an emerging category of discrimination, was
understood as the dichotomous opposite of direct discrimination. Direct discrimina-
tion involved differential treatment. Adverse effect discrimination resulted in unequal
effects in the wake of similar or facially neutral treatment. Intent was explicitly elimi-
nated as a component of adverse effect discrimination. By default, intent appeared to
be a necessary component of direct discrimination. Surely, so the argument went, if
individuals were being treated differently based on their group membership there was
an intent to discriminate.
Intent, however, is a problematic concept that is difficult to apply in direct dis-
crimination cases as well. In Action travail desfenmnes Dickson CJ.C. recognizes the
analytical quagmire of the intent requirement, observing “a continuing confusion of
-3 Intent was understood in anti-discrimination
the notions of ‘intent’ and ‘malice’
cases to entail some moral blameworthiness rather than the “simple willing of a con-
sequence.”‘ Adverse effect discrimination often arose inadvertently and did not signal
moral blameworthiness on the part of the policy-makers who introduced the facially
-‘ See wV.S. Tamopolsky & NV. Pentney, Discrindnation and the Lawv, looseleaf, vol. 1 (Toronto:
Carswell, 1994) at 2-5.
2’2 Canadian National Railway Co. v. Canada (Humnan Rights Commission), (1987] 1 S.C.R. 1114
at 1134, 40 D.L.R. (4th) 193 [hereinafter Action travail desfenunes cited to S.C.R.J. See also Robi-
chaudv. Canada (Treasury Board), [1987] 2 S.C.R. 84 at 89-90,40 D.LR. (4th) 577, La Forest J.
N’Ibid. at 1135.
Ibid.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
neutral rule.” The hard hat policy at issue in Bhinder’6 is a good example. It was
clearly not introduced to exclude Sikhs, but rather to advance occupational health and
safety. Direct discrimination may or may not involve an intent to discriminate. It
clearly involves an intent to treat individuals differently based on group membership;
however, the differential treatment may not have been designed to cause discrimina-
tory effects. If it does, it should be actionable under human rights laws regardless of
whether or not the discriminator realized or had an intent to discriminate. While it
may be important to recognize intent to underscore the human agency behind both di-
rect and adverse effect discrimination, or possibly as an element that would justify
some aggravated damages for deterrence reasons, intent should not normally be ac-
corded legal significance in either direct or indirect discrimination cases.
McLachlin J. observes that, in the context of equality claims arising under the
Charter, the Court has expressly rejected any requirement of an intent to discriminate
or of a discriminatory legislative purpose. She quotes from Iacobucci J.’s judgment in
Law v. Canada (Minister of Employment and Immigration),”3 where he states, “What
is required is that the claimant establish that either the purpose or the effect of the
legislation infringes s. 15(1), such that the onus may be satisfied by showing only a
discriminatory effect.”38 Neither the defences nor the available remedies vary depend-
ing on whether the discrimination is “overt or covert”.” Thus, while there may remain
a belief that direct discrimination is usually intentional and therefore “more loath-
some” in contrast to the “unwitting, accidental” nature of most adverse effect dis-
crimination,’ “the fact that a discriminatory effect was unintended is not determina-
tive of its general Charter analysis.'” ‘
B. When Are Disparate Effects Discriminatory?
Second, we need an understanding of when and how adverse effects cause dis-
crimination. Our paradigm for thinking about adverse effect discrimination is religion
and employment. 2 Many of the leading cases on adverse effect discrimination involve
discrimination against employees of minority religions. In these cases, the adverse ef-
fect analysis is relatively straightforward because of the complete (or 100 percent) ex-
clusion or disadvantage that results. That is, everyone in the minority religion is po-
3 It is important to recognize that intent can exist despite the facially neutral appearance of the
source of discrimination.
Supra note 12.
3 [1999] 1 S.C.R. 497, 170 D.L.R. (4th) 1 [hereinafter Law].
“Meiorin, supra note 1 at para. 47, quoting Law, ibid. at para. 80 [emphasis in original].
“Meiorin, ibid. at para. 48.
Day & Brodsky, supra note 19 at 457, quoted by McLachlin J. in Meiorin, ibid. at para. 49.
41 Meiorin, ibid.
42See Day & Brodsky, supra note 19 at 465.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.C.G.S.E.U.
545
tentially affected. The individual litigant provides a living example of the absolute ex-
clusionary effects and harm caused by employment policies, rules, or practices not
designed for the minority religion. For example, in O’Malley,; an employment rule
that requires store clerks to work on Saturdays has an absolute exclusionary impact on
Seventh Day Adventists who cannot work on their Saturday Sabbath. Disability cases
also entail virtually absolute exclusion in their discriminatory effects.
It is in adverse impact cases dealing with race or gender that the disparate effects
are often not so straightforward. In this domain, the effects may not be absolute or re-
sult in 100 percent exclusion of women or racial minorities. And yet, in many cases
the patterns of exclusion and disadvantage can and should be understood to result
from sex or race discrimination. The leading decision in the United States, Griggs v.
Duke Power Corporation,” involved not an absolute exclusion of black employees
from promotions and hiring, but a disproportionate exclusion resulting from a stan-
dardized employment test.
Meiorin provides us with an interesting factual basis for understanding adverse
effect discrimination that is based predominantly on a statistical disparity. In explain-
ing the gender inequity of the aerobic standard, McLachin J. writes:
Evidence accepted by the arbitrator demonstrated that, owing to physio-
logical differences, most women have lower aerobic capacity than most men.
Even with training, most women cannot increase their aerobic capacity to the
level required by the aerobic standard, although training can allow most men to
meet it. The arbitrator also heard evidence that 65% to 70% of male applicants
pass the Tests on their initial attempts, while only 35% of female applicants
have similar success. Of the 800 to 900 Initial Attack Crew members employed
by the Government in 1995, only 100 to 150 were female.”
Although there was expert evidence put forward by the government that “most
women can achieve the aerobic standard with training:’ McLachlin J. notes that the
arbitrator considered this evidence to be “anecdotal” and “not supported by scientific
data.’ McLachlin J. then concludes that Ms. Meiorin had established a prima facie
case of discrimination. One finds no discussion in the judgment of the amount of sta-
tistical disparity required to satisfy an allegation of sex discrimination. Of note, how-
ever, is that 35 percent of women pass the test on their initial attempt. And yet, 65
percent of women fail in contrast to the 30 to 35 percent failure rate for men-a dis-
crepancy that discloses a fairly stark gender-based disparity. It would have been help-
ful if McLachlin J. had elaborated even slightly on the Court’s conclusion that the
facts disclosed a prima facie case of discrimination. When does a statistical disparity
Supra note 27.
” 401 U.S. 424,91 S. Ct. 849 (1971).
41 Meiorin, supra note 1 at para. 11.
Sbid. at para. 69.
546
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
reveal discrimination? Is an adverse impact affecting only 51 percent of a disadvan-
taged group enough to ground a claim of discrimination, or would it be necessary to
show a greater disproportionate impact of 55 or 60 percent? In Meiorin the Court en-
dorses a finding of prima facie discrimination based on a 65 to 70 percent disparity.”
From a statistics perspective, it is generally thought that discrimination can be de-
duced from a disparate pattern of exclusion or harm that is statistically significant and
not simply the result of chance. In her book Proving Discrimination in Canada, Bea-
trice Vizkelety reviews the various approaches to statistical proof of adverse effect
discrimination in anti-discrimination law.” She highlights both the positive and nega-
tive dimensions of reliance on statistics in anti-discrimination cases. In terms of as-
sessing what constitutes a “sufficient disparity to establish discrimination”, Vizkelety
notes that “the tendency has been to decide intuitively what is significant and what is
not-‘ Drawing on the more extensive experience in the United States, Vizkelety
points to three approaches to determining a discriminatory statistical disparity. First,
there is the .05 level of statistical significance test, which Vizkelety explains means
that the “courts regard as significant, values that have only a one-out-of-twenty or five
per cent probability of occurring by chance’ Second, there is the “two or three stan-
dard deviation” approach.” Third, she refers to the four-fifths or 80 percent rule which
“presumes adverse impact where the disparity is greater than twenty per cent but not
otherwise””
While statistical clarity is important, the hidden biases of statistics and the extent
to which statistical evidentiary requirements seem to increase the complexities and
thus the costs of human rights litigation may make us reluctant to elaborate any fixed
quantitative requirements for proving adverse effect discrimination. Moreover, it is
often difficult to isolate the quantitative dimensions of discrimination from the larger
qualitative aspects of inequality. A lower statistical disparity combined with signifi-
cant qualitative factors may well substantiate a discrimination claim in some cases.”
“Moreover, recall that the Court’s approach to discrimination requires the finding of a distinction,
followed by a finding that the distinction causes harm, prejudice, or disadvantage.
48 Supra note 29 at 173-92.
49 Ibid. at 190.
Ibid. at 191.
“, Ibid. Vizkelety explains that two standard deviations correspond roughly to the .05 standard.
Three standard deviations are equivalent to a “one-out-of-a-hundred probability of occurring by
chance.”
52 Ibid.
” See Pothier, supra note 28 at 317, emphasizing that adverse effect discrimination virtually always
entails both qualitative and quantitative dimensions.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATON IN B.C.G.S.E.U.
547
In Symes v. Canada,” one of the few judgments to discuss adverse effect dis-
crimination, Iacobucci J. also raises a causation requirement. The adverse impact must
be directly caused or related to the impugned provision and not be due to other fac-
tors:
If the adverse effects analysis is to be coherent, it must not assume that a
statutory provision has an effect which is not proved. We must take care to dis-
tinguish between effects which are wholly caused, or are contributed to, by an
impugned provision, and those social circumstances which exist independently
of such a provision.”
In Symes Iacobucci J. concludes that both parents had an obligation to pay for child
care under law and that there was no proof that women assumed a greater proportion
of this financial burden. The proof adduced focussed on gender inequalities in the as-
sumption of child care work.
Thus, beyond questions of the degree of disparity, the proof must reveal a causal
link or connection between the discriminatory effects and the rule or policy. In
Meiorin the aerobic standard is directly linked to the gender disparities.’
A second issue that Iacobucci J. briefly addresses in Synes is the question of
those who are adversely affected by a provision, but who do not belong to the group
that is most affected by the neutral policy or classification. For example, if we take the
scenario in Thibaudeau v. Canada,” while 73.2 percent of custodial parents may be
women, 12.3 percent are men.’ Or in Meiorin, while a disproportionate number of
women fail the fitness tests, some men fail them as well. Iacobucci J. acknowledges
this possibility in Symes:
Mhe important thing to realize is that there is a difference between being able
to point to individuals negatively affected by a provision, and being able to
prove that a group or subgroup is suffering an adverse effect in law by virtue of
an impugned provision … If a group or subgroup of women could prove the
adverse effect required, the proof would come in a comparison with the rele-
vant body of men. Accordingly, although individual men might be negatively
“‘ [1993] 4 S.C.R 695, 110 D.L.R. (4th) 470 [hereinafter S.)nes cited to S.C.R.].
“5Ibid. at 764-65.
-Eldridge v. British Columbia (A.G.), [1997] 3 S.C.R. 624, 151 D.L.R. (4th) 577, held that pre-
existing disadvantages are clearly relevant to an equality analysis and that the state’s failure to take
them into account can ground a claim of discrimination. Thus, lacobucci J’s concerns should be lim-
ited to ensuring a causal link between the quantitative and qualitative evidence and the impugned
policy, practice, rule, or law.
[1995] 2 S.C.R. 627, 124 D.L.R. (4th) 449 [hereinafter Thibaudeau].
“These statistics are based on results of 1990 court proceedings. See ibid at para. 211.
548
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
affected by an impugned provision, those men would not belong to a group or
subgroup of men able to prove the required adverse effect. 9
It is important to clarify that the fact that some individual men may be disadvantaged
by a facially neutral policy or classification does not render the classification dis-
criminatory vis-A-vis men as a group. Nevertheless, individual men can experience
sex discrimination in the same way that women do if their life situation is similar to
many women. For example, male nurses may experience gender-based pay inequities
related to their situation in a predominantly female job. In other cases, despite the ad-
verse effect, individual men cannot be included in a group deserving of a remedy in a
human rights complaint. For example, if some men fall the fitness test their individual
failure cannot be connected to the larger problem of the fitness standard’s having been
shaped in accordance with most men’s abilities.
C. Recognizing Different Types of Adverse Effect Discrimination
A third, further dimension of our understanding of adverse effect discrimination
relates to the need to distinguish between a facially neutral rule or policy applied to
everyone and a facially neutral distinction. Some cases involve differential treatment
based on a category that does not correspond to any enumerated or recognized analo-
gous ground of discrimination. For example, differential treatment of an employee
group (domestic workers) would not directly implicate any of the prohibited grounds
of discrimination in human rights laws. It is only once we document the gender and
race composition of the classification-domestic workers-that we can identify an
apparently neutral employment classification as implicating potential problems of sex
and race discrimination. ‘ Another example arose in Thibaudeau, where it was argued
that direct differential treatment of custodial parents, a majority of whom are women,
resulted in gender-based discrimination.” Again, the category, custodial parent, is
linked to sex discrimination through an analysis of the disparate impact of the catego-
rization on women.
While in some cases the overwhelming correspondence between certain catego-
ries and the gender or racial composition of the category makes the sex or race dis-
crimination claims relatively easy to substantiate, in other cases the statistical prepon-
derance may be less marked. In such cases it may also be important to consider the
qualitative components of the harm that constitutes discrimination. For example, a
disproportionate percentage of women immigrants arrive in Canada through the
spousal sponsorship program. Nevertheless, a large number of men seek entry to Can-
ada through this program. If we combine a statistical gender disparity with a qualita-
59Symnes, supra note 54 at 770-71.
6 See supra note 30.
61 See supra note 57 and accompanying text.
2001]
C. SHEPPARD – SYSTEMIC DIscRIMINATIoN IN B.C.G.S.E.U.
549
tive appreciation of how men and women experience the spousal sponsorship pro-
gram, we can develop a deeper understanding of potential inequalities.” The tradi-
tional division of labour in the family tends to impose greater domestic work obliga-
tions on women and isolates them in the home. Women also tend to be paid lower
wages in the workforce and they are at greater risk of domestic violence. These addi-
tional gendered realities accentuate the problems women experience under the spon-
sorship program-a program that reinforces familial dependency. Thus, it is important
to be attentive to both the quantitative and larger contextual factors that combine to
cause adverse effects.
Adverse effect discrimination that is connected to an apparently neutral classifi-
cation also gives rise to the issue of discrimination against a subgroup within a larger
category. As the above examples illustrate, discrimination against domestic workers,
custodial parents, or sponsored wives affects only a subgroup of women. And yet, on
numerous occasions the Supreme Court of Canada has reiterated that discrimination
against a subgroup represents part of the larger problem of discrimination against the
particular group in society. As Iacobucci J. observes in Synes, “That an adverse effect
felt by a subgroup of women can still constitute sex-based discrimination appears
clear to me from a consideration of past decisions:”
As noted, the numerous complexities that surround the concept of adverse effect
discrimination require more, not less, legal analysis. One starting point for discussing
the various components of adverse effect discrimination would be a more precise ty-
pology of adverse effect discrimination. It could include the following:
1. Facially Neutral Policy, Practice, or Standard: A standard rule, test, practice,
or policy that is applied to everyone in the same way. The adverse effects
may affect (a) all members of a group protected by human rights law (100
percent exclusion or harm–e.g. religion or disability); or (b) a dispropor-
tionate number of members of a group protected under human rights law;
2. Facially Neutral Distinction or Categorization: Harmful differential treat-
ment, categorization, or direct distinction that corresponds to a prohibited
ground of discrimination in human rights law because of the preponderance
of individuals from particular social groups affected by the facially neutral
distinction or category (e.g. domestic workers, immigrant spouses).
For a fuller discussion of this argument, see C. Sheppard, “Women as Wives: Immigration Law
and Domestic Violence” (2000) 26 Queen’s W. 1 at 32-36.
0 Supra note 54 at 769.
550
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
IV. The Duty to Accommodate versus Institutional Change
A second area of concern regarding Meiorin is the need to clarify the relationship
between individual accommodation and a transformative and substantive approach to
equality. One of the key dimensions of Meiorin is its conclusion that the duty to ac-
commodate is an integral component of the BFOQ test. In other words, there is a duty
to accommodate with respect to all types of discrimination, both direct and adverse
effect discrimination. At the same time, McLachlin J. develops an extensive critique
of the concept of the duty to accommodate. She suggests that in the past when an ini-
tial finding of adverse effect discrimination was made, adjudicators jumped too
quickly to the question of individual accommodation without adequately assessing the
validity of the institutional policy or procedure. This critique is an important one be-
cause it emphasizes the need to change the institutional status quo, rather than simply
according special or exceptional treatment to individuals otherwise excluded or disad-
vantaged by institutional norms, policies, or practices.
In her discussion of the duty to accommodate, McLachlin J. cites with approval
the following lengthy passage from an article by Shelagh Day and Gwen Brodsky,
where they critique the conventional approach to individual accommodation in human
rights law:
Accommodation … appears to be rooted in the formal model of equality. As a
formula, different treatment for “different” people is merely the flip side of like
treatment for likes. Accommodation does not go to the heart of the equality
question, to the goal of transformation, to an examination of the way institu-
tions and relations must be changed in order to make them available, accessi-
ble, meaningful and rewarding for the many diverse groups of which our soci-
ety is composed. Accommodation seems to mean that we do not change proce-
dures or services, we simply “accommodate” those who do not quite fit. We
make some concessions to those who are “different”, rather than abandoning
the idea of “normal” and working for genuine inclusiveness.
In this way, accommodation seems to allow formal equality to be the
dominant paradigm, as long as some adjustments can be made, sometimes, to
deal with unequal effects. Accommodation, conceived of in this way does not
challenge deep-seated beliefs about the intrinsic superiority of such character-
istics as mobility and sightedness. In short, accommodation is assimilationist.
Its goal is to try to make “different” people fit into existing systems.Y
The inclusion of this citation, which articulates a transformative ideal for equality law,
is quite remarkable in a Supreme Court judgment. Indeed, earlier in her judgment
McLachlin J. uses the example of women to highlight how “the adverse effect analy-
sis may serve to entrench the male norm as the ‘mainstream’ into which women must
6′ Day & Brodsky, supra note 19 at 462, cited in Meiorin, supra note I at pam. 41.
20011
C. SHEPPARD – SYSTEMIC DISCRIMIAATION IN B.C.G.S.E.U.
551
integrate” In specifically applying her critique to Ms. Meiorin’s situation, McLachlin
J. emphasizes the importance of “rigorously assessing” the aerobic standard itself,
rather than simply assessing “whether the ‘mainstream’ can afford to confer proper
treatment on those adversely affected, within the confines of its existing formal stan-
dard.’
If the law’s role is limited to the latter it risks endorsing “the edifice of sys-
temic discrimination”
How, then, are we to understand the apparent contradiction between McLachlin
J.’s insistence that the duty to accommodate be extended to direct discrimination cases
and her extensive critique of the duty to accommodate? It would seem important to
understand this tension as follows. McLachlin J.’s concerns about an exclusive focus
on the duty to accommodate individuals reflects her commitment to a more transfor-
mative approach to equality rights. It is absolutely critical for McLachlin J. that insti-
tutional norms, policies, and practices be scrutinized and revised whenever possible as
part of the project of securing greater substantive equality. Nevertheless, individual
accommodation remains essential as -well. McLachlin J.’s integration of individual ac-
commodation into the BFOQ test reflects its importance as a component of substan-
tive equality. It should not be an excuse to avoid scrutinizing the underlying validity
of institutional norms, but rather a mechanism to ensure inclusion when revision of
the underlying practices is not possible. Both components are necessary and impor-
tant.
To understand the interplay between institutional change and individual accom-
modation, it is important to examine the third part of McLachlin J.’s integrated test
more closely. It is at this third stage that the reasonable necessity of exclusionary in-
stitutional policies or standards is assessed, and it is also at this stage that the accom-
modation requirement must be met. As mentioned, pursuant to the third part of the
test, an employer must prove that “the standard is reasonably necessary to the accom-
plishment of a legitimate work-related purpose.” In so doing, “it must be demon-
strated that it is impossible to accommodate individual employees … without impos-
ing undue hardship on the employer.”
McLachlin J. sets out a series of questions to consider with respect to this third
prong of the BFOQ test. They include the following:
(a) Has the employer investigated alternative approaches that do not have a
discriminatory effect, such as individual testing against a more individually
sensitive standard?
Meoirin, ibid. at para. 36.
bit at para. 42.
67 bid.
‘a Ibid. at para. 54.
552
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
(b) If alternative standards were investigated and found to be capable of ful-
filling the employer’s purpose, why were they not implemented?
(c) Is it necessary to have all employees meet the single standard for the em-
ployer to accomplish its legitimate purpose or could standards reflective of
group or individual differences and capabilities be established?
(d) Is there a way to do the job that is less discriminatory while still accom-
plishing the employer’s legitimate purpose?
(e) Is the standard properly designed to ensure that the desired qualification is
met without placing an undue burden on those to whom the standard ap-
plies?
(f) Have other parties who are obliged to assist in the search for possible ac-
commodation fulfilled their roles6 9
These questions are helpful and reinforce the importance of assessing the validity of
the standard itself before exploring individual accommodation or special treatment.
They prompt us to consider the extent to which an employer made a significant effort
to try to develop alternative non-discriminatory standards and approaches.
There remains, nevertheless, a potential source of confusion aggravated in some
ways by integrating a duty to accommodate into the BFOQ defence. The integrated
duty to accommodate tends to blend the analysis of the BFOQ and individual accom-
modation in such a way that one risks jumping to the individual accommodation
component of the test rather than beginning by fully assessing the reasonable neces-
sity of the standard, policy, or practice. A comprehensive analysis of the reasonable
necessity of an exclusionary practice, standard, or policy entails an assessment of
whether the policy or standard is necessary for health and safety, business necessity
reasons, or both, and whether a different standard that does not cause adverse effects
could be introduced instead (i.e. institutional change) without imposing undue hard-
ship. It would only be in the face of the impossibility of a more general policy change
that the individual accommodation question would be addressed.
What needs to be clarified, therefore, is the idea that there are in fact two types of
accommodation that should perhaps be overtly distinguished-institutional policy
change accommodation and individual accommodation. The first type of accommo-
dation entails changing the general policy, standard, or norm for everyone to remove
the adverse effects.”0 For example, if an employer has traditionally held staff meetings
from 4:00 P.M. to 6:00 P.M., employees with young children who must pick up their
‘0id. at para. 65. McLachlin J. is inspired by the insights of David Lepofsky, whose scholarship
emphasizes the need to analyze undue hardship from both a procedural and a substantive perspective.
See e.g. M.D. Lepofsky, “The Duty to Accommodate: A Purposive Approach” (1993) 1 Can. Lab.
LJ. 1.
‘o Using the terminology of accommodation may itself be problematic in this context.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.C.G.S.E.U.
553
children from daycare before 6:00 P.M. are adversely affected by the meeting time.
Two strategies of accommodation are possible. First, the meeting time could be
changed to an earlier time so that the staff meetings would finish by 5:00 P.M. This is
an accommodation that entails institutional transformation. Its feasibility could be as-
sessed in terms of whether the change would impose an undue hardship on the em-
ployer. Alternatively, those employees affected could be allowed to leave the staff
meeting early. Again the accommodation could be assessed in terms of the various
factors of undue hardship. It is apparent, however, that the second strategy is less
helpful to those employees with small children. They are allowed to leave early and
do not face any disciplinary sanctions for so doing, but they miss part of the meeting,
are denied the opportunity to participate fully in staff deliberations, and are thus less
likely to be considered for leadership positions within the organization.
Meiorin appears to endorse the institutional change model of accommodation,
since McLachlin J. challenges the validity of the aerobic standard itself rather than
simply exempting Ms. Meoirin from it while leaving it in place.” Moreover,
McLachlin J. appears to emphasize that accommodating diversity should be consid-
ered at the time institutional policies and practices are set up. Thus, we should be
cognizant of the effect of a late meeting time on employees with small children when
we decide upon the meeting time. In so doing we avoid the adverse effects in the first
place, rather than having to redress them in the wake of a human rights complaint.
Greater sensitivity and awareness of the exclusionary effects of institutional policies
and practices are best ensured through greater participation of individuals from
groups historically excluded in the processes of institutional policy-making.?
V. Assessing Health and Safety Risks in Human Rights Cases
A third concern not fully resolved in Meiorin is the issue of health and safety
risks in human rights cases. On first blush, one might have thought that questions of
risk and safety would be most pertinent to the third part of the integrated BFOQ test
proposed by McLachlin J. It is at this stage that the employer must prove that the
“standard is reasonably necessary” to the accomplishment of a legitimate work-
related purpose. In so doing, the employer must demonstrate that “it is impossible to
7 In contrast to the Supreme Court of Canada, the arbitrator inMeiorin did not rigorously scrutinize
the validity of the aerobic standard. He simply assessed whether it was rationally connected to forest
firefighting and then proceeded to consider whether individual accommodation could be made. See
supra note 4.
See C. Sheppard & S. Westphal, “Equity and the University: Learning from Women’s Experi-
ence” (1992) 5 CJ.WJ. 5 at 28-34. See also MJ. Matsuda, “Looking to the Bottom: Critical Legal
Studies and Reparations” (1987) 22 Harv. C.R.-C.L L Rev. 323.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
accommodate individual employees sharing the characteristics of the claimant without
imposing undue hardship upon the employer.”
Prior jurisprudence had addressed risk issues in the earlier equivalents to the third
prong of the Meiorin test. For example, in Etobicoke, which set out the old BFOQ
test, a standard had to be “reasonably necessary to assure the efficient and economical
performance of the job without endangering the employee, his fellow employees and
the general public.”‘ In assessing a mandatory retirement age for firefighting in that
case, the Court writes:
In an occupation where, as in the case at bar, the employer seeks to justify the
retirement in the interests of public safety, to decide whether a bonafide occu-
pational qualification and requirement has been shown the board of inquiry and
the court must consider whether the evidence adduced justifies the conclusion
that there is sufficient risk of employee failure in those over the mandatory re-
tirement age to warrant the early retirement in the interests of safety of the em-
ployee, his fellow employees and the public at large.!6
The test for assessing risk, therefore, was framed in terms of “sufficient risk” and the
burden of proof was on the employer. In Etobicoke itself the Court concludes that the
employer had not satisfied the burden of proof, relying on what was called “impres-
sionistic” evidence and general assertions that “firefighting was a ‘young man’s
game’:'” Accordingly, in that case the mandatory retirement rule was not accepted as
a BFOQ.
The safety risk issue was also addressed in Alberta Dairy Pool in the context of
factors to consider in assessing undue hardship and the duty to accommodate. Wilson
J. notes that “[w]here safety is at issue both the magnitude of the risk and the identity
of those who bear it are relevant considerations ” ‘8 The facts of that case, which in-
volved accommodating an individual employee who wanted to take Easter Monday
off for religious reasons, did not require any further analysis of risk as a component of
undue hardship.
In the wake of Etobicoke and Alberta Dairy Pool, some lower courts were con-
fronted with cases where health and safety risks were prominent. The case law reveals
a lack of consensus, with one strand of jurisprudence equating “sufficient risk” with
any increase in health and safety risks in the workplace-including minimal risks-
” Meiorin, supra note 1 at para. 54.
,4Supra note 10.
” Ibid. at 208.
7″Ibid. at 209-10 [emphasis added].
7Ibid. at 210.
” Supra note 9 at 521.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.C.G.S.E.U.
555
and another strand of jurisprudence articulating a higher standard of risk necessary to
justify exclusion from the workplace.”
In light of the facts in Meiorin, which directly raised health and safety concerns,
one might have anticipated some resolution of the diverging strands of jurisprudence
on the question of safety risks, the BFOQ, and the duty to accommodate. McLachlin
J. says very little, however, about adjudicating safety risks in her decision. She does
impugn the validity of the research methods used to establish the aerobic standard,
concluding that they reveal methodological flaws that result in gender bias. She fur-
ther accepts the arbitrator’s conclusion that “the employer has presented no cogent
evidence … to support its position that it cannot accommodate Ms. Meiorin because of
safety risks. “” To decide the case, however, she does not find it necessary to resolve
the lingering jurisprudential uncertainties about the meaning of “sufficient risk”.
Of further concern is the Court’s treatment of the risk issue in Grismer, which like
Meiorin, raised safety issues.8′ Terry Grismer alleged discrimination when his driver’s
licence was cancelled by the British Columbia superintendent of motor vehicles be-
cause of his limited peripheral vision. In deciding the case, the Supreme Court applied
its new unified test set out in Meiorin.
In so doing, McLachlin J. states at the outset that it is necessary to begin the
BFOQ analysis by identifying the superintendent’s purpose or goal in setting the pe-
ripheral vision standard.’ This appears to be an additional step not directly enumer-
ated in Meiorin. What is surprising in Grisner is the discussion of risk in McLachlin
J.’s preliminary assessment of the B.C. government’s goal or purpose in setting vision
standards for licensing:
The Superintendent’s goal in this case was to maintain highway safety. But
what kind of safety? What degree of risk would be tolerated? Where did the
Superintendent draw the line between the need to maintain highway safety and
the desirability of permitting a broad range of people to drive? The possibilities
range from absolute safety, to a total lack of concern for safety, in which case
everyone, regardless of their lack of ability, would be allowed to drive. Between
For a discussion of the conflicting case law, see Canada (A.G.) v. Rosin, [199111 FC. 391, 131
N.R. 295 (CA). See also Canada (Han Rights Conmission) v. Canada (Armed Forces), [199413
EC. 188, 114 D.L.IL (4th) 721 (CA); Canadian Pacific v. Canada (Human Rights Commission),
[1988] 1 F.C. 209,40 D.L.R. (4th) 586 (CA).
0 Meiorin, supra note 1 at para. 79. As in Etobicoke, supra note 10, McLachlin J. finds that the em-
ployer has failed to satisfy its burden of proving any increased risk to health and safety.
s” See Grismer, supra note 2.
‘Whether a goal is ‘rationally connected’ to a function, and whether a standard is ‘nmde in good
faith’ and ‘reasonably necessary’ can only be assessed in relation to a defined goal” (ibid. at para. 24).
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 46
these two extremes lies the more moderate view that reasonable safety suf-
fices.83
McLachlin J. then proceeds to conclude that the superintendent adopted a reasonable
safety standard:
The Superintendent thus recognized that removing someone’s licence may
impose significant hardship. Striking a balance between the need for people to
be licensed to drive and the need for safety of the public on the roads, he
adopted a standard that tolerated a moderate degree of risk.8′
It is following her discussion of the superintendent’s reasonable safety purpose that
McLachlin J. then turns to her three-part test.
With respect to the first prong of the test, the rational connection between the
purpose and the driver’s licensing function, McLachlin J. concludes, “Highway safety
is indubitably connected to the licensing of drivers”8 Arguably, McLachlin J. could
have readily reached this conclusion without having to assess the level of risk, since
maintaining safety generally would be rationally connected to licensing responsibili-
ties. Second, McLachlin J. concludes that the superintendent of motor vehicles clearly
adopted the peripheral vision standard in good faith.
It is with respect to the third
prong of the BFOQ test that the Court’s reasoning is more extensive.
Pursuant to the third part of the unified test, the superintendent is required to
prove that the peripheral vision standard is reasonably necessary. Integral to this in-
quiry is the need to show that it is impossible to accommodate individuals without
undue hardship being imposed upon the employer. The Court begins its analysis of
this part of the test in a curious way. McLachlin J. writes:
Risk has a limited role in this analysis. It is clear from Meiorin that the old no-
tion that “sufficient risk” could justify a discriminatory standard is no longer
applicable. Risk can still be considered under the guise of hardship, but not as
an independent justification of discrimination.”
While it is no doubt true that proof of “sufficient risk” is not the end of the inquiry
and that the question of individual accommodation is critical, it is still of fundamental
importance that the validity of the standard itself be assessed from a risk perspective
before turning to the individual accommodation option. If “sufficient risk” cannot be
shown, the policy or practice cannot be justified. If the standard can be changed for
everyone so as to eliminate its adverse effects, surely this is a better solution that ad-
Ibid. at para. 25.
Ibid. at para. 27.
85 Ibid. at para. 28.
Ibid. at para. 29.
Ibid. at para. 30.
2001]
C. SHEPPARD – SYSTEMIC DISCRIMINATION IN B.C.G.S.E.U.
557
vances institutional transformation and substantive equality. Thus, risk analysis does
not have a limited role; in cases where health and safety are at issue, it has a central
role.
Consider the example of exposure to lead in a workplace. Occupational health
and safety standards set a maximum exposure level for lead in the workplace. Re-
search on occupational health hazards facing pregnant women reveals that the maxi-
mum lead exposure level for pregnant women is lower than the general workplace
standard. Thus, pregnant women may well get screened out of occupations where
they would be exposed to unsafe levels of lead. The ideal solution would be to lower
acceptable lead levels for all employees to a level that would be safe for everyone, in-
cluding pregnant employees. It is only after an assessment of the validity of the gen-
eral standard that questions of individual accommodation in the face of an otherwise
exclusionary standard should arise. Thus, pregnant women working in situations
where they could be exposed to unhealthy lead levels should be relocated for the du-
ration of the pregnancy. This accommodation strategy, however, should only be ap-
plied once it has been shown that the standard is necessary and that a lower standard
for everyone would impose an undue hardship on the employer.
In Grismer, however, McLachlin J. does not assess the peripheral vision standard
itself in any depth. There is little discussion of whether the peripheral vision standard
was reasonably necessary to ensure health and safety. Instead the validity of the stan-
dard is presumed and the focus of the analysis is on whether additional individual
testing of Grismer should have been done and on why Mr. Grismer should be poten-
tially accorded a driver’s licence if he can prove that he compensates adequately for
his lack of peripheral vision in other ways. Thus, the Court appears to focus on the
question of individual accommodation rather than on scrutinizing the validity of the
general standard itself. A more transformative outcome for Mr. Grismer and for others
with peripheral vision problems would have been a revised standard for everyone, if
justifiable from a safety perspective. Risk analysis is central to this inquiry, as is an
elaboration of the meaning of “sufficient risk” in human rights law. Taking into ac-
count both the magnitude and likelihood of harm in framing a notion of “significant
risk” would help to give direction to the growing number of cases implicating health,
safety, and human rights!’
What Grismner and Meiorin leave us with is a sense that the Court is in favour of
balancing human rights and safety concerns, that the employer or service provider
See discussion in K. Swinton, “Regulating Reproductive Hazards in the Workplace: Balancing
Equality and Health” (1983) 33 U.T.LJ. 45.
8″ For a discussion of the need to articulate a significant risk standard in Canadian human rights law,
see DJ. Jones & N.C. Sheppard, “AIDS and Disability Employment Discrimination in and beyond
the Classroom” (1989) 12 Dal. L.J. 103 at 120-23.
MCGILL LAW JOURNAL / REVUE DE DROITDE MCGILL
[Vol. 46
must prove its defence with cogent evidence, and that the level of risk that is accept-
able is not minimal risk, but some higher level of risk that assures not absolute but
reasonable safety. What Grismer fails to emphasize, however, is that it is critically
important to analyze whether the standard or policy itself is necessary to ensure rea-
sonable safety. If it is not, the most inclusive and transformative remedy remains
eliminating or changing the standard across the board, rather than providing individ-
. ap-
ual accommodation while leaving the standard in place. In Grismer McLachlin
pears to have forgotten the forceful critique she articulated in Meiorin about how in-
dividual accommodation can reinforce or leave in place standards of systemic exclu-
sion. That is precisely the outcome in Grismer. Moreover, if risk analysis should be
central to the third prong of the integrated test in cases where safety and human rights
interface, it would have made more sense in Grismer to consider the appropriate level
of safety or risk at this stage rather than as a component of the preliminary inquiry
into the purpose or goal being pursued. McLachlin J.’s treatment of the risk issue in
Grismer thus raises a concern about whether an integrated duty to accommodate will
promote transformative equality or rather reduce scrutiny of institutional norms and
standards.
Conclusion
Meiorin provides an important contribution to our understanding of legal equality.
The decision reinforces the importance of redressing systemic inequalities that result
in exclusion and prejudice through institutional transformation and not merely by in-
dividual special treatment. Recognizing how inequality is embedded in apparently
neutral rules and standards is the first step to developing a more transformative and
. expressly endorses such a vision and
substantive vision of equality. McLachlin
thereby encourages us to use law to challenge institutionalized inequities. McLachlin
J. is also to be applauded for endeavouring to simplify the morass of legal tests and
defences in the domain of statutory human rights law by articulating an integrated ap-
proach to human rights defences. Human rights law should not be plagued by inac-
cessible distinctions and categories.
Despite its strengths, however, Meiorin does not resolve some of the human rights
issues that it raises. The Court does not elaborate on its conclusion that a prima facie
case of discrimination had been proved on the basis of gender-based statistical dis-
parities. Further elaboration of the meaning and continued relevance of adverse ef-
fects analysis awaits future cases. Second, the Court does not discuss the level of
health and safety risk required to justify exclusion from the workplace. Nor does the
subsequent Grismer case provide any further clarification on risk analysis. In Grismer
the Court does endorse a reasonable risk level, rejecting absolute safety as an appro-
priate standard; however, it does not review the risk issue in a comprehensive way. Fi-
nally, while endorsing institutional change as the most inclusive strategy for remedy-
ing systemic inequality, the Court fails to distinguish clearly in Grismer between in-
stitutional change and individual accommodation.
Beyond its jurisprudential contributions, Meiorin leaves us with the forest fire
metaphor. If we understand the harm of systemic discrimination through the metaphor
2001]
C. SHEPPARD – SYSTEMIC DISCRIMIATI7ON IN B.C.G.S.E.U.
559
of the forest fire, perhaps its destructive power and the damage it causes will be un-
derstood. So too might we better appreciate that in both firefighting and equality ad-
vocacy, prevention remains the most critical strategy.
