SCC: Don't Mess with the Three Part Tests

The good news is that the SCC revived the crucial ratio/obiter distinction.

Still, with the Court nothing is simple. Binnie still insists that "much of the Court’s work (particularly under the Charter) required the development of a general analytical framework which necessarily went beyond what was essential for the disposition of the particular case. In those circumstances, the Court nevertheless intended that effect be given to the broader analysis."

Binnie gives the example of Chief Justice Dickson (photograph from Supreme Court of Canada collection, credit to Michael Bedford)'s decision in R. v. Oakes. Oakes set out the "test" for a "reasonable limit" under section 1 of the Charter.

Narrowly,Oakes decided that the government could not require somebody found with a large amount of drugs to prove that they weren't intending to traffick. But the larger doctrine Dickson lays down is about the hoops the government has to go through once somebody shows a presumptive violation of a Charter right. First, it has to show that the law it enacted fulfilled a "pressing and substantial objective." Second, it has to show that the law is "rationally connected" to the objective. Third (and this is where the action almost always is), the government has to show that it "minimally impaired" the right, i.e., it couldn't have achieved what it was trying to do in a way less costly to the applicant's Charter rights. Finally, it has to show proportionality between the beneficial and detrimental effects of the regulation.

Because Dickson loved three-part tests, he divides his test into two stages, with the second one consisting of three parts. You had to be there.

Lawyers are familiar with a number of other "general analytical framework" cases. There is the two part Anns test for negligence, the three part (by Dickson) Pettkus test for unjust enrichment, and the godawful "pragmatic and functional" test for judicial review of administrative decision making (man, if you gotta ask, you don't want to know).

Binnie doesn't want lower courts ignoring this broad, analytical statements. At their best, these tests are just directions to lower courts to balance interests. Sometimes, they create endless verbal formulae without making the law any more determinate. But they actually do positive damage when they are held up as "new law", which overturns the previous "categorical" approach. Typically, this proves to be destabilizing and unworkable, and the Court eventually releases a judgment saying they didn't mean to overturn existing law: for the Anns test, see Cooper v. Hobart; for Pettkus, see Peel. But the Court keeps purporting to overrule all the old, "categorical" cases in particular areas of the law, most recently and importantly on hearsay and statutory interpretation.

To return to Oakes, the reality is that Dickson's statements in the original case give a distorted sense of what the actual law of section 1 really is. For instance, Dickson claimed that the government would have to come up with "clear and convincing" proof (a standard between the normal civil balance of probabilities and criminal reasonable doubt) about both its objective and whether it met the minimal impairment test. However, since then, hardly any "objectives" set out by the government have been rejected, and in most policy areas, the courts recognize the reality that social science doesn't deliver the sort of proof Dickson seemed to be calling for. There are lots of reasons to criticize the Oakes test, which seems to me like a simplistic form of cost-benefit analysis. The reality, though, is that it is sufficiently flexible that it neither poses much of a problem for further development, nor provides much guidance.

Stare decisis should be limited to the actual decisions. Courts that think they have found a unifying principle typically have to retreat later on, and have abandoned the relatively firm ground of case-by-case adjudication for abstract quasi-statutory language.