Can they? Must they? Should they?

Three questions to answer about the UCP's use of the notwithstanding clause

When governments reach for the notwithstanding clause, three questions often get blurred together. Can they? Must they? Should they?

They are related, but each demands a different kind of answer. Can they is a constitutional question about legal authority: does the Charter permit this move? Must they is a question about necessity: were there viable alternatives that could have addressed the problem without overriding rights? Should they is a normative question about judgment and democratic culture: even if it’s lawful and possible, is it wise or justified in a free and democratic society?

Too often, advocates jump from possibility to inevitability to endorsement in a single breath, as if legal permission settled the matter. It doesn’t.

Just because something is possible doesn’t mean we have to do it, and just because we do it doesn’t make it a good idea.

This post applies that three-part lens to the UCP’s use of the notwithstanding clause (Section 33) in Bill 2 to end the teachers’ strike.

TL;DR

• Can they? Yes. S. 33 is a constitutional power, and pre-emptive use is allowed.

• Must they? No. Canada has multiple recent examples of back-to-work laws without s. 33.

• Should they? Only if other remedies won’t do the job. Bill 2’s design (imposing rejected terms, rushed passage, and a multi-year strike ban) makes it hard to see how the notwithstanding clause was a “last resort”.

1) Can they?

Yes. The notwithstanding clause is part of the Charter itself. It lets legislatures temporarily legislate “notwithstanding” certain rights (s. 2, 7–15), subject to a five-year sunset and renewal, and the override must be explicit. Using s. 33 is therefore not “unconstitutional”. Quite the opposite. It’s a constitutional power with guardrails.

Some argue you must wait for a court ruling before using s. 33. That’s not how the clause has operated in practice. Political scientist Dave Snow has shown most invocations have been pre-emptive (not in response to a fresh court loss). That’s the predominant historical pattern according to his analysis. Howard Anglin agrees, noting it is only “logical” for governments to pre-empt the courts if they know judges will stand in the way of their policy designs.

What courts will say about limits on how pre-emptive use can occur is an active question now before the Supreme Court of Canada in litigation over Quebec’s Bill 21 (the secularism law). Ottawa is urging the Court to set limits on s. 33 and some provinces (including Alberta) are defending broad pre-emptive use. That case isn’t about strikes, but it does go to the constitutional ground rules for invoking s. 33.

I, for one, agree with opponents of the federal position: ruling out pre-emptive use entirely closes off potentially valid uses of the notwithstanding clause.

Unless and until the Court imposes limits on pre-emptive use going forward, invocations like the UCP’s appear constitutional. For now, can they use it? Yes.

2) Must they?

No. Governments do not have to invoke the notwithstanding clause to end public sector strikes. The UCP had a real choice in this matter.

Just because the courts have established a Charter right to strike under S. 2 of the Charter does not mean that right is absolute. No rights are. They are all subject to S. 1 of the Charter — also known as the “reasonable limits” clause.

This s. 1 route requires the government to pass a narrowly tailored statute and justify any rights limits under the Oakes test.

Under this test, the government must show a pressing and substantial objective, a rational connection, minimal impairment of the rights, and proportionality of overall effects. Back-to-work laws featuring short timelines, essential-service carve-outs, and independent arbitration have successfully been defended on that basis.

In 2021, the federal government introduced Bill C-29 (Port of Montreal Operations Act) without invoking s. 33. The Department of Justice’s Charter Statement openly acknowledged that s. 2(b) expression and s. 2(d) association were engaged, then set out the s. 1 justification: (1) pressing national economic harms; (2) a mediator-arbitrator process that kept bargaining pressure but limited harm; (3) a time-limited, neutral, dispute-resolution structure; and (4) extensive failed attempts at voluntary settlement beforehand.

Likewise in 2018, Ottawa passed Bill C-89 (Canada Post) using a mediator-arbitrator model, again staking its case on s. 1 rather than s. 33.

In legislating the teachers back to work, Alberta avoided the s.1 route and invoked s. 33. This allowed them to impose a four-year contract on teachers — a move that courts have previously struck down in other cases because heavy-handed remedies fail the minimal impairment test.

Critics argue Alberta could have continued bargaining or chosen a binding arbitration model — non-preferred to the government from a policy perspective, perhaps, but well-trodden, s. 1-friendly routes.

I, for one, think the government could have made a solid s. 1 case. Breaching s. 2 rights could be justified as a means of safeguarding students’ schooling, stabilizing the system, and preventing cascading harms.

To put it plainly, the Alberta government had options beyond the notwithstanding clause. Most would have required them to pursue arbitration, though. They chose the path of least legal resistance — one with far-reaching implications for labour rights and collective bargaining in the province.

Said more pointedly: if resort to s. 33 (paired with an imposed settlement) was the plan all along, the preceding “collective bargaining” risks looking largely performative. If that becomes the norm, future rounds may devolve into box-checking exercises rather than genuine negotiations, something labour and management alike should want to avoid.

Whether this is the path of least political resistance remains to be seen. This brings us to the normative element of the debate.

3) Should they?

This is ultimately a values question about means versus ends. Some will applaud the outcome (ending a disruptive strike or crushing a union) and discount the means. Others will condemn the outcome (imposing terms on a union and gutting collective bargaining rights) and so reject the means out of hand.

A useful consistency check is the counterfactual: would you endorse a (different) government invoking s. 33 to impose its preferred outcome on a divisive rights issue you oppose—say, mandatory vaccination during the pandemic?

Beyond a simplistic ends-justify-means approach, it is possible to support the policy outcome (end to the strike) without supporting the process.

For principled proponents of the notwithstanding clause, s. 33 is a constitutional safety valve that preserves legislative supremacy in a system of responsible government. Used sparingly, it lets elected governments — accountable to legislatures and answerable to voters — resolve tensions between rights and public objectives, especially when judicial interpretations don’t reflect local priorities. Its safeguards (an explicit declaration, time limits, open debate, and electoral accountability) deter casual use and remind courts they are arbiters, not policymakers.

For principled critics, the notwithstanding clause lets governments override core Charter protections, like expression and equality, through ordinary legislation, weakening safeguards for minorities and concentrating power in the executive. Bypassing judicial review weakens incentives to craft rights-respecting policy and can license discriminatory laws, heightening social division. Routine or pre-emptive use risks making fundamental rights contingent on transient political majorities.

Alberta’s own conservative tradition (Peter Lougheed’s “last resort” mantra) and the legal community’s warnings point us to a middle road: use s. 33 sparingly, transparently, and only after narrower options are truly exhausted.

In other words, use s. 33 only after you have failed to make a s.1 claim and only with full legislative debate. In ending the teachers’ strike, the UCP fast-tracked the bill through all stages in a matter of hours and pre-empted judicial review. That’s hard to square with “last resort.”

This reasoning drives some to call for explicit rules to be put in place for governments’ use of the notwithstanding clause. I’m not opposed to that approach, but would resist efforts to create a nationwide set of rules. If governments wish to bind themselves to certain guidelines, they can. But they ought not impose those rules on others (unless everyone agrees to a constitutional amendment). Here, I find common ground with political scientist Geoffrey Sigalet.

---

The “can they” and “must they” questions are far more straightforward, focusing our attention on what is really at stake in this debate.

Reasonable people can disagree about whether the UCP should have invoked the notwithstanding clause to break the teachers’ strike. I wish more folks focused on the means than the ends, but even principled people on both sides have valid points.

Ultimately, S. 33 assigns the public the role of arbiter. If Albertans don’t agree with the usage, they can call on their MLAs to put pressure on the government to repeal the legislation, seek to topple the government through the same sort of pressure, and ultimately send a new set of MLAs to the legislature following an election.