In an important new ruling released Wednesday, the Ontario Court of Appeal has declined to certify a class action by problem gamblers against the Ontario Lottery and Gaming Corporation (“OLG”). The decision in Dennis v. Ontario Lottery and Gaming Corporation, 2013 ONCA 501 marks only the seventh time since the introduction of the Class Proceedings Act, 1992 (“CPA“) twenty years ago that the Ontario Court of Appeal has denied certification based on one of the criteria in ss. 5(1)(b)-(e) of the CPA (the other cases being Hollick v. Toronto (City) (1999), 46 O.R. (3d) 257, aff’d,  3 SCR 158; Chadha v. Bayer Inc. (2003), 63 O.R. (3d) 22; Kumar v. Mutual Life Assurance Company Of Canada (2003), 226 D.L.R. (4th) 112; Zicherman v. Equitable Life Insurance Company Of Canada (2003), 226 D.L.R. (4th) 131; Lacroix v. Canada Mortgage and Housing Corporation, 2012 ONCA 243; and McCracken v. Canadian National Railway Company, 2012 ONCA 445). Like its decisions in Lacroix and McCracken, the Dennis decision continues the Court’s recent trend towards a rigorous scrutiny of the certification record.
At the root of the Dennis class action was the allegation that OLG failed to exercise its best efforts to exclude the putative class members from its gambling facilities, as required under the voluntary self-exclusion forms they signed with OLG in their capacities as problem gamblers. The action – which involved breach of contract, negligence, occupiers’ liability and dependant’s relief claims – was denied certification at both first instance and by the Divisional Court on the ground that all of the significant issues of liability turned on an individual issue, i.e., proof that each class member was a vulnerable problem gambler who returned to OLG facilities despite signing the self-exclusion form.
The Court of Appeal, per Sharpe J.A. (Weiler and Rouleau JJ.A. concurring), agreed with both levels of court below that certification should be denied, finding that ”the claims advanced in this case and the allegations of fault against OLG are so heavily infused with the issues of individual vulnerability that resolution of those allegations in terms of a generalized systemic wrong would not significantly advance the claims of the individual class members”. (para. 58) Three aspects of the Court’s reasons are particularly noteworthy.
First, Sharpe J.A. found that the plaintiffs failed to satisfy s. 5(1)(b) since their proposed class definition was “fatally over-inclusive”:
I agree with the motion judge that the proposed class definition of Class A Members is fatally over-inclusive. It includes all individuals who signed self-exclusion forms over a period exceeding five years. That class will include many individuals who have no claim, even if a potentially actionable failure on the part of OLG to enforce the self-exclusion form is made out. It is apparent that the problem of the class definition raises very similar issues to the question of common issues. (para. 61)
Second, after citing the common issues test for s. 5(1)(c) which the Court of Appeal had recently endorsed in Fulawka v. Bank of Nova Scotia, 2012 ONCA 443, Sharpe J.A. found that the issues proposed by the plaintiffs failed to meet virtually all of the Fulawka criteria:
In my view, resolution of the issue of OLG’s alleged systemic wrong fails on all of the listed factors. There is no “rational relationship between the class identified by the plaintiff and the proposed common issues” and the class definition is overly inclusive. Resolving the issue of whether OLG should have done more by way of enforcement of the self-exclusion forms does not make up even “a very limited aspect of the liability question” given the inherently and inescapably individual nature of the claims at their core. The significance of any determination as to OLG’s allegedly wrongful conduct is dwarfed by the need to focus on the individual issues of vulnerability and would not amount to “a substantial ingredient of each class member’s claim” nor would its resolution sufficiently “advance the litigation for (or against) the class”. The answer to the proposed common issue would not “be capable of extrapolation, in the same manner, to each member of the class” as the issues of duty, breach and causation are inextricably bound up with the individual circumstances of the class members. OLG’s alleged wrongdoing is entirely “dependent upon individual findings of fact that have to be made with respect to each individual claimant”. There is no “workable methodology” to determine issues of causation or damage on a class-wide basis. Finally, as I have already explained, OLG’s alleged misconduct is framed in “overly broad terms” and it is inevitable that the action will “ultimately break down into individual proceedings”. (para. 68)
Third, Sharpe J.A. held that the plaintiffs failed to meet the preferable procedure test in s. 5(1)(d) as well:
Even if the class definition and common issue requirements were satisfied, it is my view that a class action is not the preferable procedure. A general finding of “systemic wrong” would not avoid the need for protracted individualized proceedings into the vulnerability and circumstances of each class member. A more efficient and expeditious way to adjudicate these claims would be to proceed directly by way of individual actions as it is inevitable that a class proceeding will break down into individual proceedings in any event. (para. 71)
The Dennis decision is an important reaffirmation of the McCracken principle that claims which lack a “core of commonality” should be denied certification. Where the real issues in the proceeding are likely to be individual ones, it would not advance the goals of the CPA to certify them on the pretext of a “systemic” theory of liability. As the Court of Appeal concluded here:
Rather than providing an effective procedural tool to advance the resolution of the claims of the proposed class members, a class proceeding would amount to little more than a general commission of inquiry into the prevention of problem gambling. …[A] general determination of shortcomings in OLG policies would not address in a meaningful way the narrower specific legal issue of OLG’s liability to individual problem gamblers for the losses they have suffered. The claims advanced inevitably require an individualistic inquiry into the nature, degree and consequences of each class member’s gambling propensity. (para. 59)