By: Jennifer Koshan
Cases Commented On: Stewart v Elk Valley Coal Corporation, 2015 ABCA 225, Quebec (Commission des droits de la personne et des droits de la jeunesse) v Bombardier Inc. (Bombardier Aerospace Training Center), 2015 SCC 39
Last month Shaun Fluker posted a comment on the Alberta Court of Appeal’s standard of review analysis in Stewart v Elk Valley Coal Corporation, 2015 ABCA 225 (here). In this post I will comment on the Court of Appeal’s analysis of the test for discrimination under human rights legislation in Stewart, a matter I have commented on previously in relation to the same case at the Court of Queen’s Bench level (here), as well as in posts on other cases (see e.g. here, here and here). I will include in my analysis the Supreme Court of Canada’s decision from late July in Quebec (Commission des droits de la personne et des droits de la jeunesse) v. Bombardier Inc. (Bombardier Aerospace Training Center), 2015 SCC 39, which also deals with the test for discrimination. I will argue that the ABCA majority (Justices Watson and Picard) affirmed the wrong test in Stewart, particularly in light of the Supreme Court’s subsequent clarification in Bombardier. The approach of Justice O’Ferrall, writing in dissent at the Court of Appeal, is more in keeping with Bombardier and other recent jurisprudence.
Facts / Issues
Stewart involves a claim of discrimination based on an addiction-related disability in a unionized workplace. Stewart was an employee at Elk Valley Coal Corporation, which had an Alcohol and Drug Policy providing that employees “with a dependency or addiction” could proactively seek rehabilitation without fear of disciplinary consequences. If they did not disclose and had a workplace accident related to drugs or alcohol, they could be disciplined or terminated. Stewart was advised of this policy in a training session and signed a form indicating that he understood it. Several months later, he had a workplace accident when the loader truck he was driving hit a stationary vehicle and damaged its mirror. He tested positive for cocaine, admitted to using the drug on his days off, and to “feeling sleepy” at the time of the accident due to cocaine use the night before. However, Stewart told Elk Valley that he did not believe he had a drug problem until after the accident, after which he realized he was addicted to cocaine. Stewart’s employment was terminated, although Elk Valley also advised him that he could return to work after 6 months if there was a vacancy and he had successfully completed a rehabilitation program and agreed to a drug-free lifestyle (2015 ABCA 225 at paras 9-19, 93).
Stewart’s union filed a human rights complaint alleging a violation of section 7 of the Alberta Human Rights Act (AHRA), RSA 2000, c A-25.5, which protects against discrimination in the employment context on grounds including disability. The Court of Appeal indicated that the appeal was moot, given that Stewart is now employed elsewhere subsequent to undergoing rehab, but it exercised jurisdiction to hear the appeal “because the decision of the Tribunal and the chambers judge are both material contributions to the law in this area” (at para 23).
There were two issues for the Court of Appeal in Stewart: (1) whether the Alberta Human Rights Tribunal applied the proper test for discrimination when considering whether Stewart’s termination was contrary to human rights legislation, and (2) whether the Tribunal applied the proper test for bona fide occupational requirement to Elk Valley’s claim that its policy provided a defence to any discrimination. The Tribunal had found against Stewart on his claim of discrimination, and in the alternative, in favour of Elk Valley’s claim of a bona fide occupational requirement (Bish v Elk Valley Coal Corporation, 2012 AHRC 7). The reviewing judge, Justice Peter Michalyshyn, upheld the Tribunal on the first issue, and reversed it on the second (Bish v Elk Valley Coal Corporation, 2013 ABQB 756). This post will focus on the Court of Appeal’s treatment of the first issue, the test for discrimination.
As noted by Shaun Fluker in his post, the Court of Appeal held that the standard of review was correctness for the issue of the appropriate test for discrimination, and reasonableness for the application of that test to the facts.
The majority noted (at paras 6 and 24) that the prevailing test for discrimination is set out in the Supreme Court of Canada’s decision in Moore v British Columbia (Education), 2012 SCC 61 (CanLII) at para 33:
[C]omplainants are required to show that they have a characteristic protected from discrimination under the Code; that they experienced an adverse impact with respect to the service; and that the protected characteristic was a factor in the adverse impact. Once a prima facie case has been established, the burden shifts to the respondent to justify the conduct or practice, within the framework of the exemptions available under human rights statutes. If it cannot be justified, discrimination will be found to occur.
The Tribunal’s ruling in Stewart was rendered before the Supreme Court’s decision in Moore, but some case law prior to Moore had articulated a similar three-part test for discrimination which was referenced by the Tribunal (see 2012 AHRC 7 at para 116). It noted that other cases defined discrimination in terms of stereotypical or arbitrary decision-making (at para 117). The Tribunal accepted that Stewart had an addiction-related disability, but concluded that he was not fired because of his disability, but because of his use of drugs and breach of Elk Valley’s policy (at paras 120-121, emphasis added). Accordingly, the Tribunal found that Stewart’s disability was not a factor in his termination (at para 125, emphasis added). The Tribunal also considered the safety sensitive nature of Stewart’s position, and the policy’s objective of “accountability for an individual who had the capacity to make choices” (at para 126, emphasis added). Stewart was said to be such an individual: the Tribunal found, based on the evidence, that he had the capacity to control his use of drugs and chose not to do so (at paras 108, 118). The Tribunal concluded that “[t]he termination, in this context, did not act, either through its intent or effect, to perpetuate stereotypes or disadvantage suffered by drug addicts” (at para 126, emphasis added), and that a prima facie case of discrimination had not been established.
Justice Michalyshyn’s review decision, released after Moore, found that in spite of that judgment, “the test for prima facie discrimination includes some consideration of whether [the] adverse treatment was based on stereotypical or arbitrary assumptions” (2013 ABQB 756 at para 38). His reasoning was that the Supreme Court had considered factors of stereotyping and arbitrariness as relevant to discrimination in cases such as McGill University Health Centre (Montreal General Hospital) v. Syndicat des employes de l’Hopital General de Montreal, 2007 SCC 4 (CanLII) and Honda Canada Inc. v. Keays, 2008 SCC 39 (CanLII), and had not explicitly overruled those decisions in Moore (at paras 37, 47). Justice Michalyshyn also relied heavily on the Alberta Court of Appeal’s decision in Wright v College and Assn. of Registered Nurses of Alberta, 2012 ABCA 267 (CanLII), leave denied 2013 CanLII 15573 (SCC), in which a claim of addiction-related disability discrimination was denied by the majority, relying on an approach which incorporated considerations of stereotyping and arbitrariness. He concluded that “the Tribunal was correct to find no causal connection or nexus between Stewart’s disability and termination” (at para 45, emphasis added). Moreover, “it was appropriate for the Tribunal to consider, at the prima facie discrimination stage of the analysis, the drug policy context of the appellant’s employment, why for example such concerns may be rationally connected to the impugned treatment in question, and therefore not arbitrary or based on stereotypes” (at para 50, emphasis added).
Before the Court of Appeal, there was no dispute about the first two elements of the test for discrimination from Moore; Elk Valley conceded that Stewart had a protected characteristic (an addiction-related disability) and that he suffered adverse treatment from his employer (termination). The sole issue regarding discrimination was whether the disability had been a factor in the adverse impact (ABCA at para 28). This, in turn, raised the issue of how this element of the test should be interpreted and applied.
The appellant’s main submission on this point was that the “a factor” element should be “unencumbered by extra requirements” such as stereotyping and arbitrariness, which have sometimes been imported into the human rights context by applying case law decided under section 15 of the Charter (at para 60). (As an aside, in support of this argument the appellant cited an article of mine that was unpublished at the time and is referred to by the majority as such; that article is now available in the Canadian Journal of Human Rights).
In finding that the Tribunal had applied a test for discrimination “consistent with … Moore” (at para 6), the Court of Appeal majority emphasized (at para 29) the following statement by the Tribunal:
[P]roof of prejudice or stereotyping are not additional evidentiary requirements for the Complainant in proving prima facie discrimination. Once adverse treatment is shown on the basis of a prohibited ground, an inference of stereotyping, arbitrariness or perpetuation of disadvantage will usually be drawn. … It is not necessary that discriminatory considerations be the sole reason for the impugned actions in order for there to be a contravention of the Act (2012 AHRC 7 at para 117).
However, the majority also stated as follows (at para 6, emphasis in original):
We are not persuaded that the Tribunal imported into the test for discrimination a condition that the discrimination be based on arbitrariness or perpetuation of historical stereotypes inconsistent with our common right to equality. We do find arbitrariness or stereotypical reasoning to be relevant however.
The majority found that the Tribunal decision could be interpreted in two ways: (1) as focusing on whether Stewart’s disability was a “real factor in the adverse impact” (at paras 63, 66), with consideration of stereotyping and arbitrariness in the alternative (at para 67) and (2) as focusing on whether the employer’s actions were tainted by stereotyping and arbitrariness (at para 72).
On the first interpretation, the majority held that the Tribunal was correct in concluding that disability was not a real factor in the adverse treatment experienced by Stewart. It reasoned that “the Policy did not distinguish between people with disability and people without. It distinguished between people who break the Policy and people who do not” (at para 66), essentially confirming the Tribunal’s conclusion that Stewart had been disciplined not for his addiction, but for his use of drugs, which he had the capacity to control and chose not to. In other words, Stewart was not disciplined because of his disability, he was disciplined because of his use of drugs and breach of employer policy (at para 35, emphasis added). The majority stated that any other approach to the test for discrimination “would amend the meaning of s 7 of the AHRA, notably the words “because of … physical disability”” (at para 65).
This formulation of the test seems to elevate the requirement that disability be “a factor” in the adverse treatment to one of causation, and this is where the Supreme Court’s decision in Bombardier becomes relevant. Bombardier involved a claim of racial profiling by Javed Latif, a Canadian citizen born in Pakistan who held both Canadian and U.S. pilot’s licenses, but was refused training by Bombardier at its Dallas location because he had been denied security clearance by U.S. authorities. Latif filed a complaint against Bombardier with Quebec’s human rights commission, alleging discrimination on the basis of ethnic or national origin (see 2015 SCC 39 at paras 5-19). His complaint was upheld by Quebec’s Human Rights Tribunal, which based its finding of prima facie discrimination on circumstantial and expert evidence speaking to racial profiling by U.S. authorities since 9/11. The Quebec Court of Appeal overturned the Tribunal’s decision, finding that there was no proof of a “causal connection” between Bombardier’s refusal to train Latif and a prohibited ground of discrimination (at para 26).
The appeal before the Supreme Court raised the issues of how to define prima facie discrimination and what level of proof is necessary to establish it. In a unanimous decision written by Justices Wagner and Côté, the Court noted the “special quasi?constitutional status” of human rights legislation, and the corresponding requirement to undertake a “liberal, contextual and purposive interpretation”. Moreover, the approach to discrimination “does not change” on the basis of the ground(s) alleged by the claimant – it is always up to the claimant to establish a prima facie case of discrimination, and then the respondent can seek to justify their conduct on the basis of legislative defences or exemptions (at paras 31- 37).
Delving more deeply into the elements of prima facie discrimination, the Court articulated the test somewhat differently than it did in Moore, based on the language of Quebec human rights legislation and case law from that jurisdiction: there must be (1) a ‘distinction, exclusion or preference’, (2) based on one of the [prohibited] grounds, (3) which ‘has the effect of nullifying or impairing’ the right to full and equal recognition and exercise of a human right or freedom (at para 35). However, the elements are essentially the same: there must be adverse treatment, a protected ground, and a connection between the two. The issue in Bombardier, as in Stewart, was what level of connection was required in order to establish prima facie discrimination. The Quebec Tribunal held that it was not necessary for the prohibited ground to be the sole cause of the adverse treatment, echoing the Moore’s approach of whether the ground was “a factor” in the adverse treatment. As noted above, the Quebec Court of Appeal required a “causal connection” between the ground and the adverse treatment (at paras 43-44).
The Supreme Court held that the Tribunal’s approach was the appropriate one: “for a particular decision or action to be considered discriminatory, the prohibited ground need only have contributed to it” and it is “neither appropriate nor accurate to use the expression “causal connection” in the discrimination context” (at paras 48, 51). In keeping with Moore, the Court indicated that “[i]t is more appropriate to use the terms “connection” and “factor” in relation to discrimination” (at para 50; see also para 52). It noted with approval another decision involving a claim of racial discrimination, Peel Law Assn. v. Pieters, 2013 ONCA 396 (CanLII), where the Ontario Court of Appeal found that requiring a causal connection would be too burdensome in some cases given that human rights legislation protects against unintentional and effects-based discrimination in addition to direct and intentional discrimination (at para 49). Another reason the Supreme Court gave for its rejection of a causal connection requirement was particular to Quebec, where the civil code attributes a specific meaning to the term “lien causal.” The “close relationship” required by this standard was also said to “impose too heavy a burden” on human rights claimants (at para 51).
The Court went on to find that Latif had not led sufficient evidence to prove the elements of the test for discrimination on a balance of probabilities (at para 59). I will not comment on this aspect of the Court’s ruling as I am still mulling over its import. (Readers may be interested in Paul Daly’s analysis on Administrative Law Matters, where he argues that the Court did not accord adequate deference to the Tribunal’s application of the test for discrimination to the facts, despite cloaking its review in the language of “reasonableness.”)
For the purposes of commenting on Stewart, the salient part of the Supreme Court’s ruling in Bombardier is that a causal connection between the relevant ground of discrimination and the adverse treatment is not required. Accordingly, it is asking too much of human rights claimants that they prove the adverse treatment was “because of” the ground in question, as the Court of Appeal majority did in Stewart. Although the AHRA uses this language, so does the BC Human Rights Code, RSBC 1996, c 210, which was at issue in Moore, and Bombardier reminds us that human rights legislation should generally be interpreted consistently across jurisdictions (at para 31). In spite of the language of the Alberta and BC statutes, the burden on the claimant to establish a prima facie case of discrimination is simply to prove that the alleged ground was a factor in the adverse treatment they received. To the extent that the majority decision in Stewart suggests otherwise, it is inconsistent with Moore and Bombardier.
The majority decision in Stewart is also concerning for its acceptance of the Tribunal’s reasoning based on “choice.” The notion that a rights claimant’s choices can undermine their claim has recently been questioned in cases such as Canada (Attorney General) v. Bedford,  3 SCR 1101, 2013 SCC 72 (CanLII) (concerning the “choice” to engage in prostitution) and Quebec (Attorney General) v. A,  1 SCR 61, 2013 SCC 5 (CanLII), (concerning the “choice” of marital status). It appears to continue to hold sway in lower court decisions involving addictions, even though the Supreme Court also refuted the relevance of choice in this context in Canada (Attorney General) v. PHS Community Services Society,  3 SCR 134, 2011 SCC 44 (CanLII) (repudiating the use of injection drugs as a “choice”). In an example of slippery slope reasoning, the majority in Stewart expresses concern not only for the implications of allowing claims of discrimination by drug addicts, but also by those addicted to tobacco (see para 65).
As for the second interpretation of the Tribunal’s decision, the majority concluded that its references to arbitrariness and stereotyping were not erroneous. Rather, they amounted to “the Tribunal … effectively saying that the application of the Policy was not a pretext for discriminatory action against Stewart arising from his disability” and “effectively saying that the application of the Policy was not shown to be “discriminat[ion] … because of … physical disability” because neither direct nor indirect discrimination as revealed by arbitrariness or stereotypical thinking was visible in the application of the Policy by Elk Valley (at para 73).
The majority opined that this approach was “not inconsistent with Wright” (at para 73), and elsewhere noted that no leave to challenge the decision in Wright had been sought in Stewart’s appeal (at para 36). Once again, the majority put too much reliance on the language “because of”, an error linked to its consideration of arbitrariness and stereotyping. Where employers develop policies or otherwise act in ways that rely on stereotypes or arbitrary assumptions about (for example) addiction-related disabilities, we would say that the employer’s adverse treatment of the employees is because of their disabilities. This kind of discrimination is certainly prohibited by human rights legislation, absent a bona fide occupational requirement – which it is up to the employer to prove, not for the employee to disprove at the prima facie discrimination stage of analysis.
But human rights legislation also prohibits adverse effects discrimination, which applies where a law or policy is neutral on its face, yet has an adverse impact on persons characterized by a protected ground such as disability (see Bombardier at para 32). In these cases, the discrimination analysis focuses on the impact of the policy on the group identified by the protected ground, rather than the rationale for the employer’s conduct. To require the claimant to prove arbitrariness or stereotyping in discrimination claims makes it very difficult to make out a prima facie case of adverse effects discrimination (see a recent review of adverse effects discrimination cases under the Charter by Jonnette Watson Hamilton and me here).
To be fair to the majority, the Supreme Court has not entirely repudiated the relevance of stereotyping and arbitrariness in human rights claims. The majority cites Justice Abella’s 2007 concurring opinion in McGill University Health Centre, supra at paras 48 to 53, in support of the continued relevance of these concepts. More recently, I have noted that the Court missed an opportunity in Moore to explicitly discount these considerations, and in fact used the language of arbitrariness in some parts of its judgment in that case (though not in its formulation of the test for discrimination). The same is true of the Court’s decision in McCormick v. Fasken Martineau DuMoulin LLP,  2 SCR 108, 2014 SCC 39 (CanLII) (see my critique here), as well as in Charter discrimination cases such as Quebec v A, supra (see here) and Kahkewistahaw First Nation v. Taypotat, 2015 SCC 30 (CanLII) (see here). Although the Supreme Court’s decision in Bombardier does not rely on arbitrariness or stereotyping, and indeed could be seen to refute those concepts as requirements in proving discrimination in light of its causation analysis, it would be preferable if the Court would explicitly confirm this approach once and for all.
Turning to Justice O’Ferrall’s dissent, he agreed with the appellant’s interpretation of the “factor” requirement in the test for discrimination, i.e. that “the protected characteristic need only be part of the causal chain leading to the adverse impact” (at para 99). His reasons for accepting this argument are very much in line with the Supreme Court’s decision in Bombardier. He noted that this interpretation was consistent with Pieters, supra, and other appellate level jurisprudence, as well as with the “actual wording” of the test in Moore (at paras 99, 100, 106). He also noted that the concepts of stereotyping and arbitrariness had arisen under section 15 of the Charter, and were inappropriate in the human rights context, given the difficulty for claimants in knowing the motives behind their employers’ actions – these were for employers to prove as part of the bona fide occupational requirement analysis (at para 102). And, he noted that stereotyping and arbitrariness would be “virtually impossible” to prove in adverse effects discrimination cases (at para 103). Justice O’Ferrall distinguished Wright on the basis that it was decided before Moore and involved culpable behaviour (theft from an employer by employees with addictions) (at para 109), but his reasons for rejecting stereotyping and arbitrariness are similar to those of Justice Berger’s dissent in Wright.
On the question of whether the Tribunal had applied the correct test, Justice O’Ferrall emphasized that while it had initially indicated that proof of stereotyping or arbitrariness was not required, it did go on to apply those as requirements, leading it to the conclusion that a prima facie case of discrimination had not been proved (at para 113). The reviewing judge had also applied these elements, and thus the wrong test (at paras 114-116).
Applying the correct test, Justice O’Ferrall believed that a prima facie case of discrimination was made out. In his view (at para 118), “the complainant’s addiction was not just a factor leading to his dismissal, it was the entire reason for it.” This was made clear by the employer’s termination letter, which indicated that “[t]he reason why the complainant was terminated, rather than given a lesser disciplinary penalty, was because the employer wanted to deter other employees like the complainant, who were also drug dependant, from failing to report that dependency prior to a workplace accident” (at para 122). In contrast to the majority, Justice O’Ferrall found that the employer’s treatment of Stewart actually was because of his disability, even though such a high degree of connection is not required. As for the employer’s argument that Stewart’s addiction “had nothing to do with his dismissal because it would have fired him regardless of the extent of his drug use”, Justice O’Ferrall properly noted that “If an employee is fired for failing a drug test, regardless of the nature and extent of his drug use, the employer is adopting a zero-tolerance work rule that is based upon the stereotypical assumption that any amount of off-duty drug use impairs an employee from working safely.” This assumption “would satisfy even the harsher test for prima facie discrimination on the basis of a perceived disability brought on by the use of drugs” (at para 124). There was thus significant disagreement between the majority and dissent on both the appropriate test for discrimination and whether there was discrimination on the facts.
The Stewart case, like Wright before it, illustrates the difficulties that tribunals and courts have in articulating and applying the test for discrimination. This seems to be a particular problem in cases involving drug and alcohol abuse, as the Court of Appeal has previously applied the Moore test in other contexts (see e.g. Telecommunications Workers Union v Telus Communications Inc., 2014 ABCA 154 (CanLII), which was acknowledged by the majority in Stewart at para 25). In addiction-related disability cases, employers may have valid reasons for responding to employee conduct – whether culpable or non-culpable – with consequences that deter employees from substance use that may affect their job performance and create workplace safety issues. However, the place to consider these reasons is at the bona fide occupational requirement stage of analysis, not by importing causation requirements and considerations of stereotyping and arbitrariness into the test for discrimination. Although Bombardier goes some way toward clarifying this area of law, it is time for full rehabilitation of the test for discrimination, including an explicit rejection of stereotyping and arbitrariness as irrelevant factors.
To subscribe to ABlawg by email or RSS feed, please go to https://ablawg.ca
Follow us on Twitter @ABlawg