Squaring the Public Law Circle

Canadian administrative lawyers keep trying to reconcile parliamentary sovereignty and the Rule of Law; they shouldn’t bother

Ancient Greeks wondered whether it was possible to construct a square of the same area as a given circle using only a compass and a ruler ― to square the circle. The problem occupied some great minds of that age and of the subsequent ones, even Napoleon apparently. It took well over two millennia until it was shown to be impossible to solve. Public law has its own quadrature problem, posed by A.V. Dicey (the first edition of whose Introduction to the Study of the Law of the Constitution came out just a couple of years after the demonstration of the impossibility of squaring the circle): it consists in fitting together, albeit by means of verbal rather than geometrical contortionism, parliamentary sovereignty and the Rule of Law.

Dicey and many others since him have mostly been preoccupied by this problem in the context of fundamental individual rights, and their protection from a legislature unconstrained by a supreme law constitution. Canada eventually abandoned this attempt ― or rather cut back on it significantly, since some rights, such as that to property, remain unprotected by the Canadian Charter of Rights and Freedoms. But, to an extent that Dicey did not imagine and that is arguably without parallel in the rest of the Commonwealth, we have re-deployed our intellectual energies merely to a different application of the same problem, this one in administrative law. We are struggling to reconcile parliamentary sovereignty, which suggests giving effect to legislative attempts to insulate administrative decision-makers from judicial review, and the Rule of Law, which, as Dicey himself suggested, requires courts of justice to apply the law. We are not succeeding.

It is not for lack of trying. The majority opinion in the supposedly still-leading case on judicial review of administrative action,  Dunsmuir v New Brunswick, 2008 SCC 9, [2008] 1 SCR 190, recognized that

[j]udicial review seeks to address an underlying tension between the rule of law and the foundational democratic principle, which finds an expression in the initiatives of Parliament and legislatures to create various administrative bodies and endow them with broad powers. [27]

Dunsmuir and the subsequent cases that have fucked up beyond all recognition refined the framework that it laid down attempted to resolve this tension and to make sure that, as a Russian saying has it, the wolves are sated, and the sheep unharmed. Scholarly commentary has worked, I think, in the same direction.

The most recent example is a thoughtful post on ABlawg by Martin Olszynski. Professor Olszynski seeks to recover what he sees as Dunsmuir’s promise of reconciling parliamentary sovereignty and the Rule of Law. He proposes to achieve this by making

two inter-related changes to the Dunsmuir framework … The first change would be to reverse the presumption of reasonableness on questions of law to a presumption of correctness, which can then be rebutted for the large majority of such questions through the presence of a privative clause (this approach would be similar to that proposed by Justice Deschamps in Dunsmuir). The second related change would be to abandon the overly broad and fundamentally contradictory concept of “expertise” as a basis for deference and to replace it with the potential for democratic accountability, which ultimately is the basis for legislative supremacy.

Although the judiciary has the “training, independence, and impartiality” to claim “the upper hand in the interpretation of the law”, it ought to yield this upper hand to  legislative statements that call for deference to administrative decision-makers. Legislatures “must be respected – because they are democratically elected and accountable”. Provided they make themselves sufficiently clear by enacting “privative clauses” (provisions that typically seek to out judicial review of administrative decisions or to strictly limit it), legislatures can be made to answer for any decision to remove legal interpretation from the purview of the courts. When the legislation includes a privative clause, a reviewing court should, therefore, defer, but not otherwise ― and especially on the pretense that an administrative decision-maker is an expert by virtue of its very existence.

I agree with Professor Olszynski’s criticism of the role that the idea of administrative expertise has come to play in Canadian administrative law (which I have not fully summarized ― you really should read it). Last year I wondered here whether “the Supreme Court is embracing that pop-psychology staple about 10,000 hours of doing something being enough to make one master it”, and I elaborate on my worries about “expertise” in a paper I recently presented at the TransJus Institute of the University of Barcelona. I also agree that courts should not be shrinking violets when it comes to legal interpretation. It’s their job, and it’s the think that they’re supposed to be good at. If legislatures decide to scrap some of the administrative bodies they have set up (a guy can dream, right?), the courts will have to apply the legislation these bodies are now responsible for. They ought to be able to do that.

But I am skeptical of Professor Olszynski’s suggestion that the presumption that questions of law must be addressed by courts should, in the name of democratic accountability, by rebutted by privative clauses. Indeed, I think that the idea of democratic accountability is not readily applicable in this context. Professor Olszynski argues that accountability works by pointing to his own criticism of the application of a privative clause in an environmental law case, and contrasting it with the fact “that few labour or employment lawyers would argue against privative clauses in that context”. With respect, the possibility of academic criticism does not make for democratic accountability; nor does acceptance by a relevant expert community (if indeed “labour and employment lawyers” are the relevant expert community in relation to labour law ― what about economists, for instance?) make for democratic legitimacy. How many voters have ever heard of privative clauses, never mind being able to articulate any thoughts on their desirability? To believe that legislatures can, let alone will, be held accountable for eliminating the courts’ role in legal interpretation unwisely, or even abusively, requires more optimism than I could ever muster.

I am inclined to think ― though my thoughts on administrative law are still tentative ― that in determining the standard of review we should not attempt to reconcile the Rule of Law and legislative supremacy. The reconciliation is never meant to be real in any case. The Rule of Law is, ultimately, the dominant value, because even those who claim that they want to respect legislative will refuse to give effect even to the clearest privative clauses. To take a statutory provision that says “no judicial review” to mean “deferential judicial review” is not to accede to the legislature’s desires, but to impose one’s own principles ― including the principle of the Rule of Law ― on it.

And there is nothing wrong with this. The Rule of Law, as the Justice Rand observed ― in the context of a lawless exercise of administrative power ― in Roncarelli v Duplessis, [1959] SCR 121 at 142, is “a fundamental postulate of our con­stitutional structure”. It is a constitutional principle that can, as the Supreme Court recognized in Reference re Secession of Quebec, [1998] 2 SCR 217, result in “substantive limitations upon government action” ― including, relevantly to us here, in government action aiming at reducing the courts’ powers of judicial review. By contrast, as the Secession Reference also recognized, democracy ― whether direct democracy, which was at issue in that opinion, or representative democracy, and whether accountable or otherwise ― must be confined by constitutional limitations. The Court wrote “that with the adoption of the Charter, the Canadian system of government was transformed to a significant extent from a system of Parliamentary supremacy to one of constitutional supremacy”. [72] But that’s not quite right. The Charter imposed additional restrictions on legislatures, but it did not “transform” the constitutional system, which was already one of “constitutional supremacy” under the Constitution Act, 1867.

To the extent that it is required by the Rule of Law principle, judicial review of administrative action, including correctness review on questions of law, is a constitutional requirement. This extent is the question that Canadian administrative lawyers and judges should be addressing. Virtually everyone, I think, agrees that the Rule of Law requires correctness review in at least some cases. My own inclination is to say that it requires correctness review often, and perhaps always. I might be wrong about that, but if I am, this is because I misunderstand the Rule of Law, not because I fail to account for Parliamentary sovereignty and to give effect to (modified versions of) privative clauses. There is simply no need to bring parliamentary sovereignty into the standard of review equation, thereby making it unsolvable. Unlike in mathematics, the impossibility of squaring the public law circle cannot be conclusively demonstrated (though even in mathematics the demonstration apparently did not stop enthusiasts from trying). But the futility of well over a century’s worth of attempts should, I submit, be a warning to us all.

Law in La-La-Land

The post-truth jurisprudence of Canadian administrative law

Last month, the Supreme Court issued a decision in Edmonton (City) v. Edmonton East (Capilano) Shopping Centres Ltd., 2016 SCC 47, which deals with the evergreen issue of determining the standard on which a court must review the decision of an administrative tribunal. I wasn’t able to comment on this case at the time, because 240 exam papers landed in my office a couple of days thereafter, but I would like to return to it now, because in my view the majority opinion shows quite clearly what a fantastical creature Canadian administrative law has become.

Justice Karakatsanis, writing for the majority, describes the circumstances of the dispute as follows:

Alberta residents may dispute their municipal property assessment before a local assessment review board. When one Edmonton taxpayer did so, the assessment review board decided to increase the assessment the taxpayer had disputed. [1]

The question is whether the board in question had the authority to increase ― or could only decrease or refuse to change ― the assessment disputed before it. The board obviously concluded that it did, but did not explain the decision ― admittedly because the taxpayer did not deny that it had the requisite authority. It is only on a subsequent appeal to the courts that it did so. How, then, must the board’s tacit, unexplained, decision be treated? Is it entitled to judicial deference, or can the courts simply substitute their own view of the matter?

* * *

Defer, says the majority. In the Supreme Court’s administrative law jurisprudence, deference is the presumptive stance, including on questions of law such as the one at issue.

This presumption of deference on judicial review respects the principle of legislative supremacy and the choice made to delegate decision making to a tribunal, rather than the courts. A presumption of deference on judicial review also fosters access to justice to the extent the legislative choice to delegate a matter to a flexible and expert tribunal provides parties with a speedier and less expensive form of decision making. [22]

Only when the question at issue falls within one of a few exceptional categories is the “presumption of deference” rebutted. In the majority’s view, none apply here. In particular, the fact that the legislation involved specifically provides for an appeal from the administrative decision on a question of law, as was the case here, does not matter. The administrative law framework, with its presumption of reasonableness, is to be applied regardless.

As Justice Karakatsanis notes, this framework acknowledges that “[t]he presumption of reasonableness may be rebutted if the context indicates the legislature intended the standard of review to be correctness” [32]. The dissenting judges would have found that this is the case here, because the board’s expertise lies not in the interpretation of its enabling legislation, which is at issue here, but rather in property valuation, which is not. Justice Karakatsanis is unmoved by this:

Expertise arises from the specialization of functions of administrative tribunals like the Board which have a habitual familiarity with the legislative scheme they administer. … [A]s with judges, expertise is not a matter of the qualifications or experience of any particular tribunal member. Rather, expertise is something that inheres in a tribunal itself as an institution. [33]

Having concluded that she must defer to the board’s decision, Justice Karakatsanis then face the question of how to defer to a decision that is entirely unexplained. After all, as the still-leading (although perhaps increasingly from behind) case on reviewing administrative decisions,  held, “reasonableness ‘is … concerned with the existence of justification, transparency and intelligibility within the decision-making process'” [36; citing Dunsmuir v New Brunswick, 2008 SCC 9, [2008] 1 SCR 190 at [47]]. But not to worry: in that same case, the Supreme Court went on to say it is enough to consider the reasons that could have been given for the decision, even though they were not, and to defer to them. This is what Justice Karakatsanis does, for 20 paragraphs. She concludes that board’s decision was reasonable.

* * *

If this is not post-truth jurisprudence, it’s pretty close. The Oxford English Dictionary, which recently chose “post-truth” as its word of the year, says that it describes “circumstances in which objective facts are less influential in shaping public opinion than appeals to emotion and personal belief”. The Supreme Court’s vision of administrative law might not be the product of appeals to emotion, but there is certainly a great deal of personal belief unburdened by a reckoning with objective facts there.

Start with the claim that the presumption of deference is all about respecting legislative intent to grant decision-making authority to administrative tribunals rather than courts. If the Court really cared about legislative intent, wouldn’t it ask itself what the purpose of providing a statutory right of appeal on questions of law, as the legislature had done in this case, is? The majority does no such thing. It is content to affirm the irrelevance of either the text or the purpose of such provisions and holds that even if the statute says “appeal”, the courts must read “judicial review” and defer. Yet as the dissent cogently points out, the statutory text is pretty clear that the legislature wanted some questions of law (those deemed important enough by a judge) to be resolved by the courts. For all its show of deference to the legislature, the majority only cares about its own views about how administrative law should operate.

Next, consider the claim that the administrative decision-maker’s expertise entitles it to judicial deference. The majority does not discuss the expertise of the board whose decision is at issue here. It does not meaningfully respond to the dissent’s argument that the board has no particular expertise on the question before the Court. It is content to state an airy generality: the very existence of a specialized tribunal is enough to make it an expert. Why? It seems almost as if the Supreme Court is embracing that pop-psychology staple about 10,000 hours of doing something being enough to make one master it. Here’s news for the Court: that’s not true. The dissent’s take on this issue is, once again, more realistic:

an administrative decision maker is not entitled to blanket deference in all matters simply because it is an expert in some matters. An administrative decision maker is entitled to deference on the basis of expertise only if the question before it falls within the scope of its expertise, whether specific or institutional. [83]

As mentioned above, the dissent would have found that the question at issue here was not within the scope of the board’s expertise. The majority, once again, is uninterested in facts or, as the dissent also notes, in legislative intent (which in some case may well be to create a decision-maker without legal expertise), and ignores them the better to apply its beliefs about the proper relationships between courts and administrative decision-makers.

And all that for what? Not to actually defer to the supposedly expert board’s reasoning ― but to make up a reasoning that the board may or may not have come up with on its own, and defer to that. As Paul Daly put it (aptly, it goes without saying),

Karakatsanis J. points to multiple features of the elaborate statutory scheme that might be said to support the alternative interpretation and explains how each of them is nonetheless consistent with the Board’s interpretation (if one can call it that), much of which is supported by reference to a decision made by another body that “formerly” had appellate jurisdiction from the Board [44]. This, frankly, is quite bizarre. Who knows what the Board would have said if these points had been made to it? Indeed, who knows what the Board will say in future cases when these points are made to it? (Paragraph break removed)

It’s as if Justice Karakatsanis were playing chess with herself, and contriving to have one side deliberately lose to the other. This, admittedly, is not a post-truth reasoning. It is outright fiction.

* * *

As you can probably guess, I do not like any of it one bit. Perhaps this is not a very educated feeling. I have not thought about administrative law as much as prof. Daly, for instance, or many others. (I do have plans to get more serious about it, but it’s a medium- or even a long-term project.) I hope, however, that there may be some value in even an uneducated person stating the view that the emperor has no clothes, and that his proceeding any further in this state is offensive to public decency. What is more difficult for an uneducated mind is to say what dress the emperor ought to put on. Prof. Daly argues

that the only way to move the law forward within the existing framework without starting again from scratch is to apply reasonableness review across the board, with the important caveat … that the range of reasonable outcomes will be narrower in cases featuring an appeal clause.

I cannot comment on this suggestion, beyond saying that I am, in my uneducated state, extremely uneasy with deferential review of administrative decisions on questions of law. I will say, though, that I’m not at all sure that starting again from scratch is not the right thing to do, or even the only possible thing to do.

The Scope of Smuggling

On Friday, the Supreme Court issued a pair of decisions clarifying the scope of the provisions of the Immigration and Refugee Protection Act (IRPA, among friends) relative to “people smuggling” ― the transportation to or across international borders of consenting individuals who lack the authorization to cross the borders in question. In  B010 v. Canada (Citizenship and Immigration), 2015 SCC 58, the Court held that a person is not inadmissible to Canada on grounds of organized criminality if he or she was not acting in order to obtain “a financial or other material benefit in the context of transnational organized crime.” [5] In R. v. Appulonappa, 2015 SCC 59, the Court held that the penal provision of IRPA relative to smuggling was unconstitutionally overbroad “insofar as [it] permits prosecution for humanitarian aid to undocumented entrants, mutual assistance amongst asylum-seekers or assistance to family members.” [5]

While the legal issues on which the two decisions ultimately turn are different ― B010 is about statutory interpretation while Appulonappa is a Charter case ― they are closely related. Both decisions are unanimous, with the Chief Justice writing for the Court. I will review them together.

* * *

In B010, the issue was the scope of IRPA‘s provision that made foreign nationals inadmissible to Canada “on the grounds of organized criminality” and thus, if they claim asylum, prevented their claims from being determined on the merits, for “engaging, in the context of transnational crime, in activities such as people smuggling.” In the opinion of the Chief Justice, the “ordinary sense” of that provision, and in particular of the phrases “people smuggling” and “organized criminality” did not necessarily suggest that the activities it is aimed had to have a profit motive.

However, the Chief Justice concluded that provision’s “broader statutory context … suggests that [it] targets organized criminal activity in people smuggling for financial or other material benefit.” These considerations included the other grounds of inadmissibility provided by IRPA, which suggest that the provision at issue was specifically intended to target money-making organized crime, as well as the scope of the penal provision at issue in Appulonappa, which the Court held also targeted those who sought to make a profit from people smuggling. Moreover, the definition of a “criminal organization” in the Criminal Code, enacted so as to bring Canada into compliance with an international treaty a protocol to which deals with people smuggling, also refers to the profit motive. The protocol in question and other international agreements provided an “international context” which unambiguously pointed towards an intention to specifically outlaw profit-motivated smuggling, while not penalizing humanitarians and family members helping refugees.

* * *

As for Appulonappa, the issue there was whether IRPA‘s provision that imposed severe penalties, including potentially lengthy terms of imprisonment, on those who “knowingly organize, induce, aid or abet the coming into Canada of one or more persons who are not in possession of a visa, passport or other document required” was overbroad, and thus inconsistent with the principles of fundamental justice enshrined in section 7 of the Canadian Charter of Rights and Freedoms. A law is unconstitutionally overbroad if it penalizes acts that are not inconsistent with its purpose. Here, the problem was said to be that the provision reached the actions of those who did not seek to derive a material benefit from helping would-be refugees come to Canada. While the government argued the provision aimed “to catch all acts that in any way assist the entry of undocumented migrants,” the Court concluded that its purpose was narrower than that.

While the text of the provision in question was “broad enough to catch assistance to close family members and humanitarian assistance,” [36] here too other factors were more revealing of its aims. These factors included, once again, the international law, as well as the statutory context ― notably, the inadmissibility provision interpreted in B010, and the purposes of IRPA as a whole, which balance humanitarian concerns with those related to security. Importantly, IRPA provided that no prosecution for smuggling could be initiated without the agreement of the Attorney General. This “filter” was intended, as records of Parliamentary debates confirmed, to prevent the prosecution of persons participating in activities that fall within the broad definition of smuggling but are motivated by humanitarian or family concerns.

As the Chief Justice pointed out, it was thus clear that

Parliament itself understood … that the provision’s reach exceeded its purpose by catching those who provide humanitarian, mutual and family assistance to asylum-seekers coming to Canada, but argued that this overbreadth was not a problem because the Attorney General would not permit the prosecution of such people.  [72]

In her view, reliance on the Attorney General exercising his discretion in this way was not enough to avoid the overbreadth problem. Whatever Parliament’s intentions, nothing stopped the Attorney General from authorizing a prosecution inconsistent with those intentions. Thus, so long as the provision remained on the books, “people whom Parliament did not intend to prosecute [were] at risk of prosecution, conviction and imprisonment.” [74]

By way of justification under section 1 of the Charter, the government seemed to argue that the provision could not have been less broadly than it was, and was therefore minimally impairing. The Chief Justice, however, observed that “[s]ection 1 of the Charter does not allow rights to be limited on the basis of bare claims, but requires the Crown to provide a demonstrable justification for inconsistencies with Charter rights” [82; emphasis in the original], and concluded that this justification was missing.

Ultimately, this may be of little help to the appellants who had challenged the constitutionality of the provision in question in Appulonappa. Instead of striking it down completely, as they had hoped, the Court read it down “as not applicable to persons who give humanitarian, mutual or family assistance,” [85] and order a trial based on this revised version. Because the appellants were not actually humanitarians at all, but the crew of a ship that brought dozens of refugees to Canada, they might not far any better as a result of their constitutional victory.

* * *

The Court’s reasoning in both decisions seems persuasive to me in both decisions, at least as to the specific questions the cases turn on. I am rather less impressed with the Court’s apparent endorsement of the policy of criminalizing profit-motivated people smuggling. In the introduction to B010, the Chief Justice claims that “[t]he smugglers … cynically prey on these people’s desperate search for better lives to enrich themselves without heed to the risks their victims face.” Unfortunately, the Chief Justice pays no heed to the role that the criminalization of for-profit smuggling plays in making it the gruesome business it undoubtedly is. I have addressed this issue here and, quite recently, here. I might return to it again. For now, it is enough that, bad as they still are, our immigration laws are a little less bad now than they used to be.

Twelve Banned Books Weeks

Once upon a time, I mused about whether Parliament could ban books as part of its regulation of election campaign spending. The specific question that interested me then was whether the exemption of “the distribution of a book, or the promotion of the sale of a book, for no less than its commercial value, if the book was planned to be made available to the public regardless of whether there was to be an election” from the definition of “election advertising” in section 319 of the Canada Elections Act (CEA) could be abolished. But now, just in time for Banned Books Week, life ― or, rather, the Public Service Alliance of Canada ― has come up with a somewhat different censorship scenario.

La Presse reports that the Alliance has complained to Elections Canada about political commentator, consultant, and activist Éric Duhaime’s giveaway of 5000 copies of his book Libérez-nous des syndicats! (Free Us from the Unions!). Mr. Duhaime is apparently giving the books away for free in order to counteract an anti-Conservative (and pro-NDP) campaign by Québec’s largest union, the FTQ, to which the Alliance is associated. In the Alliance’s view, the anti-union book falls with the definition of “election advertising” in section 319, and since it is being away for free during the election campaign, the exemption for books sold “for no less than [their] commercial value” does not apply. Since Mr. Duhaime has not registered with Elections Canada to advertise as a “third party” as section 353 of the CEA requires, he is, the Alliance says, acting illegally.

Mr. Duhaime says that he is not campaigning for or against a political party ― only against unions ― and thus is not infringing the CEA. But that’s not quite obvious. The CEA deems to be election advertising

the transmission to the public by any means during an election period of an advertising message that promotes or opposes a registered party or the election of a candidate, including one that takes a position on an issue with which a registered party or candidate is associated. (Emphasis mine)

The key issue, it seems to me, is whether Mr. Duhaime’s book (which, to be clear, I have not read) can be considered as “tak[ing] a position on an issue with which a registered party … is associated.” Is the anti-union position Mr. Duhaime expresses “associated with” the Conservatives ― as the Alliance seems to believe? Or is the pro-union position Mr. Duhaime combats “associated with” the NDP? I’m not sure, but I don’t think that the argument is an impossible one to make. As best I can tell, there is no case law interpreting s. 319 generally or the notion of “an issue with which a registered party or candidate is associated” in particular. And these terms aren’t exactly self-explanatory.

Which, in my view, is a big problem. Here we have a statutory provision that can be applied to punish speech, to impose fines on someone whose “crime” is to give away a book ― and we don’t actually know what it means. Mr. Duhaime probably enjoys the free publicity that comes with the complaint, but not everyone will feel that way. The problem of chilling effect from speech-restricting legislation that is imprecisely worded and thus difficult to interpret in advance of application is a real one.

Here’s another issue with the drafting of s. 319, while we’re at it. One of the exemptions from the general definition of “election advertising” concerns “the transmission by an individual, on a non-commercial basis on what is commonly known as the Internet, of his or her personal political views.” So suppose that Mr. Duhaime had put the text of his book on a freely-accessible website. That would pretty clearly fall within the exemption ― even if the website were only set up for the duration of the election campaign, since the statute says nothing about internet communications having to be “regardless of whether there was to be an election,” as it does for books. But now consider a somewhat different example. Suppose that, instead of just putting the text of his book on a website, Mr. Duhaime makes his book available as an ebook, say through the Kindle store ― again, for free. Does that count as an illegal “distribution of a book … for … less than its commercial value,” or as a legal “transmission by an individual, on a non-commercial basis on what is commonly known as the Internet, of his or her personal political views”? I have no idea. On the one hand, it’s not clear that an ebook ought to be treated any differently from a dead-tree book. On the other, it’s equally unclear why a text in .azw, or .mobi, or .epub format should be different, for the purposes of election law, from the same text in .html format. I guess it would be a fun question to put on a statutory interpretation exam, if you are a slightly sadistic professor.

But again, laws that restrict expression, especially expression on political issues, should not be written for the benefit of slightly sadistic professors of statutory interpretation. If expression must be restricted, as the Supreme Court believes the expression of “third parties” ― that is citizens and organizations who are not candidates or political parties ― must be restricted, at least the restrictions should be clear and narrowly defined. Citizens should not have to guess; nor should they be at the mercy of complaints by other citizens or groups who simply happen to detest their politics.

The shoe was once on the other foot. After the 2003 election campaign in Québec, another union associated with the FTQ was prosecuted by Québec’s election authorities for distributing a pamphlet criticizing a party that took an anti-union position ― a party whose leader Mr. Duhaime was then advising, as it happens. The union then challenged the constitutionality of the Québec legislation on third-party participation in election campaigns ― unsuccessfully. Now, it would seem, labour has learned to use this sort of law as a weapon against its enemies. (In fairness, however, Québec’s law was even more restrictive than the CEA. A union’s distribution of a pamphlet to its own members would not be a violation of the federal statute.) But we should, I think, be concerned that our election campaigns are in danger of becoming twelve-week-long periods for banning books.

The Uber Decision

Last week, Ontario’s Superior Court of Justice delivered a much noticed judgment rejecting Toronto’s claims that Uber could not operate there without registering and obtaining a license as a taxicab or limousine broker. Needless to say, the ruling is of great practical importance to Uber’s users, both passengers and drivers, as well as those who seek to regulate it out of existence. Legally, the decision, City of Toronto v Uber Canada Inc., 2015 ONSC 3572, is about a very narrow issue of statutory interpretation. Yet the recently-appointed Justice Dunphy’s thorough and well-written opinion provides us an opportunity to reflect on the importance of the Rule of Law and the processes of legal change.

The City of Toronto, like many others in Canada and elsewhere, has chosen to cartelize the transportation of persons by privately owned cars. All the cars used for that purpose are divided into the categories of “taxicabs” and “limousines.” The number of the former is fixed; the number of the latter is restricted indirectly, by imposing a variety of regulations on their owners and operators. In addition, the City requires “taxicab brokers” and “limousine service companies” to obtain licenses in order to operate within its limits. The City’s case against Uber was that Uber was acting as a “taxicab broker” or a “limousine service company,” without having done so. It asked the Court for both a declaration and an injunction that would have ordered Uber to stop its operations in Toronto. Uber, for its part, claimed that its operations were not covered by the City’s by-laws.

Justice Dunphy begins by determining whether Uber cars might be “taxicabs” or “limousines” within the meaning of the applicable by-law, chapter 545 of the City of Toronto Municipal Code. The definition if a “taxicab” is limited to categories defined by the various types of permits issued by the City. Since Uber cars lack such permits, they do not fall within this definition, reasons Justice Dunphy, and must be “limousines,” which include all cars “used for hire for the conveyance of passengers in the City of Toronto” other than “taxicabs.” To say that unlicensed cars used for that purpose are still “taxicabs” “would make nonsense of the definition of ‘limousine’ in the same enactment” [57] and thus cannot be the correct interpretation.

Having concluded that Uber cars are “limousines,” Justice Dunphy asks himself whether Uber ― or, more precisely, any one of the three members of the Uber group of companies actually sued by the City ― acted as a “limousine service company.” The by-law defines such a company as a “person or entity which accepts calls in any manner for booking, arranging or providing limousine transportation.” Uber, Justice Dunphy holds, does not “accept calls,” and thus is not covered by the definition. In Justice Dunphy’s view “accepting” a call or any sort of request “requires the intervention of some element of human discretion or judgment in the process and cannot be applied to a merely passive, mechanical role of receiving and relaying electronic messages.” [78] Yet that is precisely what Uber does.

Having provided prospective passengers and drivers with software that allows them to connect, often well in advance of any specific trip being envisioned by either party, it relays passengers’ requests for a ride to the nearest car available. Unlike a traditional taxi broker or limousine company, it cannot reject the request (for example if there are no cars available) or undertake to fulfill it. It is the driver who receives the request who takes the decision. Uber no more “accepts” requests for rides than does a phone company whose networks are used to transmit traditional calls for cabs, or automated services that connect a prospective rider with a broker. In Justice Dunphy’s view, it “is very likely” that “the by-law was drafted and the word ‘accepts’ was selected in lieu of the more generic ‘receives'” precisely in order “to exclude such businesses from the scope of the regulation.” [70]

Justice Dunphy also considers the meaning of the word “calls,” used in the definition of a “limousine service company” ― but not in that of a “taxicab broker” which, unlike the limousine company, can accept “requests.” This difference in wording, Justice Dunphy says, it must be given effect, so that “calls” cannot be taken to mean “requests.” Besides, the word “requests” is a recent innovation in the definition of a “taxicab broker,” and the City could have amended the definition of “limousine service company,” but has not done so. Online requests handled by Uber are not “calls” in any normal sense of the word, and this is an additional reason for concluding that it is not a “limousine service company.”

Although it might seem like excessive legalistic pedantry to some, I find Justice Dunphy’s analysis persuasive. Needless to say, it only applies to the specific legislative framework before him. Had the relevant definitions been drafted differently, his conclusions would presumably have been different too. But given the by-laws that were actually before him, I think that Justice Dunphy was quite right to distinguish the passive or mechanical functions of receiving or transmitting a communication and the (at least somewhat) discretionary function of accepting an order, as well as to give effect to the distinction between “calls” and “requests” which the City itself has created.

As I said in the beginning, beyond the narrow point about the meaning of the specific words used by Toronto’s city council to regulate its taxi industry, there is a broader one about the Rule of Law. As Justice Dunphy points out, “[t]he goal of statutory interpretation is not to start with the desired outcome that the regulator seeks in light of new developments to see what means can be found to stretch the words used to accomplish the goal,” [69] which as he says is what he would have had to do in order to rule for the City in this case. The Rule of Law requires, among other things, that legal rules be public and relatively stable. It also requires the government to be bound by the existing legal rules. A legal system where the meaning of the rules can change because the government wants it to, even though it cannot be bothered to follow the procedures available for legal change, is not one where the Rule of Law prevails.

It is often said that insisting on this “formal” sort of Rule of Law is not enough, because requirements as to the publicity and clarity of legislation and insistence on legal change following recognized procedures does not do much to constrain government. Government can still enact whatever rules it wants, so long as it goes about it the right way. But if it really were so easy for government to change the rules while following the applicable procedures, would it really be fighting so hard to avoid having to do so? As Justice Dunphy recognizes,

[t]he City finds itself caught between the Scylla of the existing regulatory system, with its numerous vested interests characterized by controlled supply and price, and the Charybdis of thousands of consumer/voters who do not wish to see the competition genie forced back into the bottle now that they have acquired a taste for it. [9]

Changing the rules, in this context, is not as easy as those who denigrate the formal understandings of the Rule of Law would have us believe. And so it matters a great whether

the City’s regulations, crafted in a different era, with different technologies in mind [have] created a flexible regulatory firewall around the taxi industry sufficient to resist the Uber challenge, or … instead [have] created the equivalent of a regulatory Maginot Line behind which it has retreated, neither confronting nor embracing the challenges of the new world of internet-enabled mobile communications. [12]

Justice Dunphy’s conclusion, of course, is that the City’s regulations have done the latter, and Uber is thus free to pursue its (charm) offensive. In theory, the regulatory troops can still be withdrawn from the useless, antiquated defences and thrown into the battle to stop the invaders. In practice, it may well be too late by the time they can be mobilized.

Justice Dunphy understands this, no doubt. Although he insists, as most judges not named Richard Posner are wont to do, that “[q]uestions of what policy choices the City should make or how the regulatory environment ought to respond to mobile communications technology changes are political ones” [13] and not for him to resolve, his awareness of, and willingness to mention, the conflict between “vested interests” and the “competition genie” suggest that he knows that his decision will influence the choices that will end up being made. Indeed, Justice Dunphy’s attention to the details of Uber’s technology and business model, as well as his awareness of the broader context in which the case before him fits, not to mention his rhetorical flourishes, have something at least vaguely Posnerian about them. The decision he has delivered is not only an Uber decision, meaning a decision about Uber. It’s also an über-decision ― one that is superior to what one usually sees.


In one of my very first posts, I wondered what the Supreme Court meant by describing a statute, or a common-law right, as “quasi-constitutional.” I concluded that this description probably did not mean anything substantial, and was little more than an indication that the Court considered the statute or right in question as very important. Its decision yesterday in Thibodeau v. Air Canada, 2014 SCC 67, is further evidence for that proposition. The main issue in Thibodeau was whether the limitation of an airline’s liability for “damages” to its passengers set out in the so-called Montreal Convention, an international treaty dealing with commercial air travel and made part of Canadian law by the Carriage by Air Act, prevented the Federal Court from making an award of damages for Air Canada’s violation of its duties under the Official Languages Act.

A large part of the majority’s decision, and the dissent, are concerned with the issue of whether the Montreal Convention applies to an award of damages made under a statute such as the Official Languages Act, rather than a more traditional claim (say for injury or lost luggage). Justice Cromwell, writing for a five-judge majority concludes that it does, based on his reading of the Convention’s text, his understanding of the Convention’s purpose, and his review of foreign decisions. Justice Abella’s dissent (with which Justice Wagner agrees) comes to the contrary conclusion. I will not deal with the interpretation issue here.

What I want to briefly focus on ― though don’t expect any deep thoughts here ― is the subsequent issue of the interplay between the Montreal Convention and its implementing legislation, and the Official Languages Act. Subsection 77(4) of the Act gives the Federal Court the power to “grant such remedy as it considers appropriate and just in the circumstances” for violations. This is obviously a very broad grant of remedial power, and it would normally include the possibility of awarding damages. So having concluded that the Convention purports to exclude such awards of damages, the majority must decide whether the “quasi-constitutional” Official Languages Act trumps this exclusion.

To answer this question, Justice Cromwell says, one must first determine whether the Montreal Convention and the Official Languages Act actually conflict. Only if they do will it be necessary to determine which is to prevail. When legal rules merely “overlap in the sense that they address aspects of the same subject, they are interpreted so as to avoid conflict wherever this is possible” (par. 89). Justice Cromwell concludes that there is no conflict here, because subs. 77(4) of the Official Languages Act and the Montreal Convention can be reconciled by not interpreting the former as requiring damages to be available in all circumstances (and, in particular, when such an interpretation would conflict with Canada’s international obligations). Justice Cromwell points out that “[c]ourts are … slow to find that broadly worded provisions were intended to be an exhaustive declaration of the applicable law where the result of that conclusion creates rather than avoids conflict” (par. 99). They are also reluctant to conclude that there exists a conflict between provisions enacted for different purposes. These considerations apply here.

The appellants and the Official Languages Commissioner, however, argued that because the Official Languages Act is quasi-constitutional, it must be taken to apply fully, allowing for no “reconciliation” in the case of an “overlap.” Justice Cromwell acknowledges the “quasi-constitutional” status of the Official Languages Act, and says that it “should be interpreted generously to achieve its purpose” (par. 112), but holds that “[t]hese factors, however, do not alter the correct approach to statutory interpretation” (par. 112) ― which is the same as for all other statutes. For Justice Cromwell, the Act, “read in its full context, demonstrates that Parliament did not intend to prevent s. 77(4) from being read harmoniously with Canada’s international obligations given effect by another federal statute.” Subsection 77(4), Justice Cromwell adds, is “broad and general” rather than “an exclusive and exhaustive statement in relation to the Federal Court’s remedial authority …  overriding all other laws and legal principles” (par. 113). Other remedies remain available against Air Canada, while the Montreal Convention does not restrict the availability of damages against anyone else. The provisions can be made to work together without either losing its meaning, so there is no conflict.

This may be a sensible outcome, though I find it difficult distinguish what Justice Cromwell does from an application of the principle that lex specialis derogat generalis ― a specific law applies in derogation of a broad one ― which is of course one way of resolving conflicts between statutes rather than of “harmonizing” them. And it is a way of resolving conflicts that is specifically excluded by  subs. 82(1) of the Official Languages Act, which provides that its Parts I-V “prevail to the extent of the inconsistency” with any other act of Parliament. But even taking Justice Cromwell’s reasons at face value, they very strongly suggest that a statute’s “quasi-constitutional” standing is in reality, quasi-meaningless.

The Discomforts of Religion

Religion gives law a lot of trouble. Most often, the difficult question is what to do about it ― what to do about prayer at municipa council meetings, what to do about religious believers asking for exemptions from general laws. But sometimes, the law must confront a more basic, and perhaps an even more difficult question: what is religion in the first place? The Supreme Court of Canada has not had to say very much on this vexed subject. The most it did say was Justice Iacobucci’s pronouncement, without much explanation, in Syndicat Northcrest v. Amselem, 2004 SCC 47, [2004] 2 SCR 551, at par. 39, that

[i]n essence, religion is about freely and deeply held personal convictions or beliefs connected to an individual’s spiritual faith and integrally linked to one’s self-definition and spiritual fulfilment, the practices of which allow individuals to foster a connection with the divine or with the subject or object of that spiritual faith.

The Supreme Court of the United Kingdom, however, had to address the question of the nature of religion in a case it decided earlier this week, R (on the application of Hodkin and another) v Registrar General of Births, Deaths and Marriages, [2013] UKSC 77. The issue it had to decide was whether a Scientologist church could qualify as a “place of meeting for religious worship” for the purposes of the English legislation that grants recognition to marriages performed in such “places.” And to figure out what “religious worship” means, it is necessary to have an idea of the concept of religion.

The exact same issue had arisen in the case of R v Registrar General, Ex p Segerdal [1970] 2 QB 697, where Lord Denning held that Scientology’s churches were not places of religious worship. Those, he explained, were “where people come together as a congregation or assembly to do reverence to God. It need not be the God which the Christians worship. It may be another  God, or an unknown God, but it must be reverence to a deity.” Lord Denning allowed that “[t]here may be exceptions. For instance, Buddhist temples are properly described as places of meeting for religious worship,” but Scientology did not, in his view, have anything like a diety, being more a philosophy than a religion, and so its churches did not qualify.

The Supreme Court in Hodkin declares itself unsatisfied with this reasoning. Lord Denning’s insistence on a deity as a distinguishing hallmark of religion, it says, is misplaced. Because it would exclude from the definition of religion certain groups generally recognized as religious, it would be “a form of religious discrimination unacceptable in today’s society” (par. 51). Indeed, Lord Denning himself was willing to make an exception for Buddhism ― yet his exception was unexplained and seemed unprincipled, suggesting that the rule he proposed was too. In addition, on his approach, courts would be called upon to decide whether a creed did, in fact, refer to a deity or supreme being, or whether the being it referred to was supreme enough. This is the stuff of theology, not of law, and courts would do best to stay out of it.

Embarking on its own search of the meaning of the expression “religious worship,” and thus of the concept of religion itself, the Court observes that

[t]here has never been a universal legal definition of religion in English law, and experience across the common law world over many years has shown the pitfalls of attempting to attach a narrowly circumscribed meaning to the word (par. 34).

Furthermore, the understanding of religion changes over time, and

It is no good considering whether the members of the legislature over 150 years ago would have considered Scientology to be a religion because it did not exist (par. 34).

After an in-depth review of a couple of judicial decisions, one from the US and one from Australia, the court ventures its own definition. Religion, for it, is

a spiritual or non-secular belief system, held by a group of adherents, which claims to explain mankind’s place in the universe and relationship with the infinite, and to teach its adherents how they are to live their lives in conformity with the spiritual understanding associated with the belief system. … Such a belief system may or may not involve belief in a supreme being, but it does involve a belief that there is more to be understood about mankind’s nature and relationship to the universe than can be gained from the senses or from science (par. 57).

An additional question is whether the statutory term “religious worship” involves the “adoration of a deity.” Armed with a definition of religion that does not involve a deity, the Court says no. Any religious rites will do. The statute permits

members of a religious congregation, who have a meeting place where they perform their religious rites, to carry out religious ceremonies of marriage there. Their authorisation to do so should not depend on fine theological or liturgical niceties as to how precisely they see and express their relationship with the infinite … Those matters … are more fitting for theologians than for the Registrar General or the courts (par. 63).

The Scientologists are allowed to perform marriages in their church.

This case illustrates a general and increasing reluctance of common law courts to scrutinize religious beliefs. This same feeling was one of the driving factors in the majority decision in Amselem, which held that courts cannot inquire into whether a religious believer was really required to do or not to do something by some correct interpretation of his or her faith. This attitude of the courts is fully justified: judges are not qualified to serve as theological arbiters, and, more importantly, as the agents of a secular state, they have no right to take on that role. That is something that those who defend various forms of prohibitions on the wearing of religious clothing or symbols by arguing that these are not truly required by the religion of those who wear them should keep in mind.