Administrative Law

Vavilov (Part 2 of 2): Did the Court Fix Canadian Judicial Review?

This Part is for those interested in law and how it develops but is not a technical law journal article. For anyone wanting a detailed legal analysis I suggest reading Paul Daly’s 5 blog posts starting here.

A Bit of History

Changing Courts, Changing Attitudes

Until about the early 1980’s most Canadian judges were men, usually appointed from law firms representing businesses and governments.  Judges appeared generally sympathetic to litigants like their former clients and less sympathetic to unions, women and the less fortunate in society. As a broad generalization, judges were to the political right of the average Canadian.

Gradually, judicial appointments became more diverse and judicial attitudes evolved. The attitude change was substantially influenced by administrative law professors like Bora Laskin (who eventually became Chief Justice of the Supreme Court of Canada). Some law professors (then and now) serve as labour arbitrators or members of administrative tribunals like Labour Relations Boards or Human Rights Commissions. Unions will reject the appointment of arbitrators unless they believe them to be at least somewhat sympathetic to unions. Hence these arbitrators and labour board members led the way, both in law journal articles and as judges, to offset the perceived judicial bias against unions and labour relations adjudicators.  The attitude moved leftwards and has remained there, but whether you see this as having moved to the political centre or to the left of centre depends upon your political views.

Judges protected labour adjudicators by declaring that labour boards had labour relations expertise equal to or greater than generalist judges. In practice this was usually true.  “Expertise” became shorthand for “stop picking on them” and show some deference. Considerable respect for such decision-makers has been the judicial policy for decades.

The expertise argument is an appeal to authority, not reason: if Agency X is an expert body its decisions are presumed reasonable so a judge should leave them alone, barring something highly unreasonable.  At the top of the deference pyramid are the labour boards and some well-known labour arbitrators.  However, at the base of the pyramid, human rights commissions (among others) were deemed to have little or no expertise, for reasons rarely supported by any actual evidence of expertise.   In short, “expertise” was a judicially created stereotype based on the type of tribunal rather than an empirical assessment of the expertise of the actual decision-makers.


“Our administrative law is a never-ending construction site where one crew builds structures and then a later crew tears them down to build anew, seemingly without an overall plan.”  Justice David Stratas (writing in a law journal in his non-judicial capacity

The central issue in judicial review (JR) is: when should a court overrule the decision of a government agency to which the legislature has delegated decision-making authority?  This goes to the heart of the relationship between the courts and the executive branch of government in a democracy.  JR is an important, if limited, judicial control over government decision-makers.

Despite Justice Stratas’ hoped for improvement, this endless construction/destruction cycle for JR will probably continue after Vavilov because there is still no overall plan.  Indeed, no such plan may be possible, for the reasons described below.  And, unfortunately, the project has been for the benefit of the construction crew (judges), with little benefit to the current and future occupants of the building.

The Producers and the Consumers of JR

Most of the scholarly writing about administrative law is a dialogue between law professors and judges.  While professors and judges have focused on helping the courts, the consumers of judicial decisions – those bound by them – have been forgotten. In fairness, I’m not sure what could or should be done to help consumers, but making outcomes more predictable would be a good start.

The SCC’s Vavilov decision wanted to reconstruct the law.  Will Vavilov be more durable than several previous SCC reconstructions? That depends upon how well it has resolved the problems created by numerous inconsistent SCC decisions, without creating too many new problems. And there are lots of problems.  David Mullan, one of the leading names in Canadian administrative law, has written about a mere fifteen: “Unresolved Issues on Standard of Review in Canadian Judicial Review of Administrative Action—The Top Fifteen!” [(2013) 42 Advocates Quarterly 1].

Unfortunately, much of the very lengthy Vavilov decision merely restates the obvious, using a lot of vague and hypothetical language. The words “could be” appear 22 times; “may be” appears 36 times; and “for example” appears 36 times.  Vavilov provides an imprecise recipe for JR judges: take a large dollop of deference and add a pinch of reasonableness, to taste.

Deference Expanded

The ‘expertise’ shield created for labour boards was expanded over decades to include others, based on the judges’ assertion of their relative expertise. If the tribunal was deemed more expert than judges, at least on issues of policy, its decisions should be given judicial deference. “Deference” was code for “unless it is clearly unreasonable, let it go”.

Before “deference“ the cases used the “jurisdiction” test, asking “Was the decision within the tribunal’s jurisdiction?”  If not, it would be quashed.  However, even in most “jurisdiction” cases the underlying issue was the same: is this decision bad enough that I should quash it? If so, I will find that it fell outside jurisdiction.

Eventually, the frequency with which “jurisdiction” was held lacking became an embarrassment.  By 1979, in C.U.P.E. v. N.B. Liquor Corporation, the SCC held:

“The question of what is and is not jurisdictional is often very difficult to determine. The courts, in my view, should not be alert to brand as jurisdictional, and therefore subject to broader curial [court] review, that which may be doubtfully so.”

Next, the jurisdiction test was abandoned in favour of “deference”, with which the courts introduced the current JR approach: determine the “standard of review” and then apply that standard. There are two standards: reasonableness (used in most cases) and correctness.

Under reasonableness the decision under review does not have to be correct, it only has to be “reasonable”.  If so, it is entitled to some unspecified degree of deference.  Under correctness there is no deference; the decision reviewed has to be correct. Correctness usually applies when the issue is interpretation of a law that is not within the specialized expertise of the tribunal (again, with some dispute about what is/is not within the tribunal’s expertise).

The majority in Vavilov changed this approach slightly.  It held that expertise is no longer used to determine the standard of review (reasonableness versus correctness).  But it is still relevant in conducting the review itself, to ensure respect for initial decision-makers.  The two dissenting judges in Vavilov sharply criticized this minor reduction in the use of expertise.  How much practical difference it will make, if any, remains to be seen.

Is JR Good Law after Vavilov?

Among the criteria for judging a law is whether it has predictable results.

Predictability permits those subject to it to know beforehand what they can/cannot do, thereby enabling government agencies to avoid making decisions likely to be quashed.  Similarly, parties investing hundreds of millions of dollars on a pipeline application before agencies like the National Energy Board will know beforehand what the Board can/cannot legally do.

On the other hand, laws with highly unpredictable results serve primarily to review done deeds, imposing considerable delay, expense and disruption.  A recent high profile example is the repeated judicial reviews of Cabinet decisions approving the Trans Mountain Pipeline.

Another high profile case, decided by the SCC together with Vavilov, involved the CRTC decision about commercials on the Superbowl program. The CRTC has for years required Canadian broadcasters to remove US commercials from US shows and replace them with Canadian commercials.  However, for the Superbowl, the CRTC wanted to leave in the US commercials.  Before deciding whether to do this the CRTC held a public hearing and consultations, and learned that Canadians wanted to see the very creative US ads in the Superbowl.  But running the US ads would deprive Bell (the owner of the CTV Network, the Canadian carrier of the Superbowl) of millions in Canadian advertising revenues, and the NFL would lose its share of that revenue. Hence these parties would oppose that decision.

Imagine that you are the CRTC’s lawyer and the CRTC members ask you “If Bell and the NFL take us to court will our decision be quashed? Your answer: “toss a coin”. You could start by saying that the SCC has again emphasized in Vavilov that judges should show deference towards the Commission’s decisions, provided that the decision is “reasonable”. But how much deference is enough for this particular decision? We can’t measure “deference” or “reasonableness” with a tape measure.  If it seems reasonable to the CRTC, would it seem reasonable to judges who may know next to nothing about broadcasting or commercial substitution?  (Spoiler alert: the SCC quashed the decision.)

In Vavilov, as in earlier cases, the SCC provided a nuanced description of the mental process that judges should go through, and the appropriate attitude, when reviewing a decision.  That attitudinal guidance may be helpful to judges, but it is still only a judge-centric guide to judicial conduct. It cannot help either the government agency or parties before it to predict JR outcomes.

A decision under JR has only two possible outcomes: pass or fail. The SCC’s exhortations to adopt the virtuous judicial attitude of deference is not a law.  Unfortunately, there is no “law” of JR for those who have to eat the result, only a recipe for those who dish it out.  This is not a criticism of the SCC; that’s just the way things are when judges are judging tribunals deciding facts and policies outside of the normal judicial hierarchy.

Legislatures often delegate decision-making authority in very vague language that is capable of widely varying reasonable interpretations.  An excellent example of such meaningless feel-good fluff is section 3 (1)(d) of Canada’s Broadcasting Act, setting out the policy within which the CRTC must make its decisions:

3 (1) It is hereby declared as the broadcasting policy for Canada that ….

(d) the Canadian broadcasting system should

(i) serve to safeguard, enrich and strengthen the cultural, political, social and economic fabric of Canada,

(ii) encourage the development of Canadian expression by providing a wide range of programming that reflects Canadian attitudes, opinions, ideas, values and artistic creativity, by displaying Canadian talent in entertainment programming and by offering information and analysis concerning Canada and other countries from a Canadian point of view,

(iii) through its programming and the employment opportunities arising out of its operations, serve the needs and interests, and reflect the circumstances and aspirations, of Canadian men, women and children, including equal rights, the linguistic duality and multicultural and multiracial nature of Canadian society and the special place of aboriginal peoples within that society, and

(iv) be readily adaptable to scientific and technological change;”

Would anyone disagree with any of that?  If you can’t disagree with something it really isn’t saying much.  And how would a reviewing court decide whether a CRTC decision would or would not serve to safeguard, enrich and strengthen the cultural, political, social and economic fabric of Canada… etc.?

JR Theory versus Practice

The practice of JR is often quite different from the theory.  In theory, courts must start a JR by determining the “standard” of review: the closeness of the look at what was done (i.e. reasonableness versus correctness).  JR reasons for decision often contain many pages on this issue. In borderline cases the standard of review might affect the outcome, but in practice, if a judge seriously dislikes the decision below she will quash it, and explain that it was unreasonable. JRs are seldom determined on standard of review alone; that lengthy discussion is often a waste.

Respect for a government agency is an attitude. How do higher courts measure the sufficiency of the respect shown by lower courts? If she disagrees with the agency must the judge just hold her nose and approve it? Or can she write lots of pages demonstrating ‘all due respect’ and then overrule it?

The Logical Problem With Deference

In practice, if a judge agrees with the decision below she will usually say she is exercising deference; if she disagrees she will also say she is exercising deference, but, because the decision is unreasonable she must overrule it. In practice, how helpful is the deference mantra?  Does it explain or obscure the judge’s real reasons for decision? Some of each.

Once a conspicuous display of deference became mandatory it became heretical to challenge its usefulness.  The late SCC Justice John Sopinka stood alone in his criticism, for which he was quickly criticized.  He argued that when he agreed with a tribunal’s decision he wasn’t deferring, he was agreeing.  When two people agree on something deference is irrelevant, a misuse of language.  Deference only begins when there is some degree of disagreement.  And in that case how strongly does one need to disagree before deference becomes rubber-stamping a decision one believes is wrong?

Critics responded that “reasonableness” meant that the decision under JR should fall within a range of reasonable outcomes.  This begs the further question of why a generalist judge is qualified to decide the range of reasonable outcomes of a 10,000 page multidisciplinary NEB decision on pipelines.

In response to the latter objection Vavilov and some of its predecessors have shifted the focus of review to whether  the tribunal’s reasons for decision (if, as usual, written reasons are required) demonstrate “justification, transparency and intelligibility”.  The obvious problem with this definition of reasonableness is that it is largely about the style of the decision reviewed.  If it is well-written and not unintelligible, even legally wrong decisions can be held reasonable because legal correctness doesn’t matter.  However the courts rationalize it to make it appear objective, reasonableness review and deference will always require a substantial degree of subjective judgment in deciding the pass or fail question.

Despite the theory, as can be seen from both Vavilov and Bell/CRTC, the SCC decisions were actually based on a close review of a few words in a statute.  Vavilov’s parents were held not to meet the definition of a single word: “representative”.  The decision in Bell/CRTC also rested on interpretation of a few statutory words.  An observer reading the numerous pages of explanation and the actual basis of the outcome might well comment that “the more you explain it the less I understand it.”







6 replies »

  1. The same thing has recently happened with another judicial review of a CRTC decision in regards to the relief of wholesale internet rates to third party providers. Incumbents went to court for a stay, and the bloody courts stuck their noses into CRTC matters and granted the stay. Sort of defeats the purpose of a quasi independent body.


  2. Sorry, I see no other way of contacting you. <> Should be: <>

    I value the dispassionate analysis I find in your articles and share them routinely.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s