constitutional conventions

A Strategy against Lethargy?

Can courts order the government to appoint judges promptly? It’s complicated, but probably not

In my last post, I summarized and criticized the Federal Court’s decision in Hameed v Canada (Prime Minister), 2024 FC 242, which declared that the government of Canada has a duty under the constitution to just get on with judicial appointments already. Emmett Macfarlane has made many similar points on his Substack as well. But, as I noted, while the actual reasons given by Brown J were very bad, that doesn’t mean there weren’t serious arguments to be made for the outcome he reached ― though I’m not sure whether any such arguments were actually put to him by counsel. In this post, I review these arguments.

By way of reminder, the way I see it, there are two key substantive issues, on which I focus. First, is the government’s failure to ensure timely judicial appointments judicially reviewable in principle? And second, if it is, is there a judicially administrable standard by which the constitutionality of government inaction regarding judicial appointments might be assessed? As to the first, the government’s objection has to do with the involvement of constitutional conventions in judicial appointments. As to the second, the difficulty is that there is simply no standard apparent in the relevant constitutional or statutory provisions.

In reality, the first issue breaks down into several different ones. As Mark Walters points out in a very important article on “Judicial Review of Ministerial Advice to the Crown“, “it is important to distinguish between decisions made by the Governor General based on ministerial advice and decisions of the Governor in Council”. The former include the appointment of Superior Court judges pursuant to s 96 of the Constitution Act, 1867; the latter, the appointment of Federal Court judges pursuant to s 5.2 of the Federal Courts Act. Failure to draw this distinction is a mistake in the identification of the conventions of judicial appointments committed in Democracy Watch v. Canada (Attorney General), 2023 FC 31, which is carried over in Hameed as I mentioned in my last post.

Conventions are, indeed, involved in both situations, but in different ways. In the case of Governor in Council appointments,

Although it is only by virtue of convention that the legal powers conferred upon the “Governor in Council” are exercised by the Governor and present cabinet ministers, it is by virtue of law that these ministers, sitting as a committee of the Privy Council, participate with the Governor to make the decision. In these circumstances, ministers are not merely advising the Governor as to what he or she should decide; they are, as a matter of law, joining with the Governor to make the decision.

The law at issue is, so far as appointments under the Federal Courts Act (and other statutory provisions) are concerned, s 35(1) of the Interpretation Act. What this means, I think, is that there is no ministerial advice that can be reviewed ― or that could be unreviewable ― separately from the decision, or failure to make a decision, by the Governor in Council. Rather, so far as appointments to the Federal Court are concerned, if one can identify a legal duty to make such appointments in a timely manner and a judicially administrable standard against which the exercise of such a duty can be measured, the conventional aspect of the appointments powers is beside the point and there is no obstacle to reviewability.

Appointments on advice, i.e. those to the Superior Courts, are a different beast. Here, conventions do matter. But, as Professor Walters argues, this is not to say that they are not reviewable. Professor Walters’s article is a comment on the challenge brought by Aniz Alani against the then-Prime Minister’s policy, first implicit and then overt, of not making appointments to the Senate. (I blogged about it here.) As he notes, the government’s response was to argue

that the Crown has the legal authority to act and the minister’s role is only to advise, and it will be added, with emphasis, that the advice is given as a matter of constitutional convention only and so cannot be the subject of judicial review. In this way, executive power may be exercised in a legal black hole. (35)

Professor Walters suggests that there are two potential “way[s] to address this worrisome conclusion”. (35) Of these,

One … is to question the assumption that constitutional conventions are never justiciable. Perhaps the time has come to contemplate the possibility of at least declaratory judicial relief when ministerial advice flouts established conventional rules. In this way, the veil of law would be pierced to reveal where real power lies. (35)

This is the approach Brown J took in Hameed, but he did so in a very clumsy, and indeed quite untenable, manner. I will return to this possibility below, because more can be said about it. But first, let me recount the other option outlined, quite persuasively in my view, by Professor Walters.

Professor Walters argues that “[t]he idea that ministerial advice to the Crown is a matter unknown to and outside the law … is a misunderstanding that results from the confusion between law and convention”. (37) It is law ― specifically, s 11 of the Constitution Act, 1867 ― that sets up the Privy Council as a body of advisors to the Crown. Convention, of course, is both what dictates that its advisory function is exercised, from time to time, by the ministers for the time being, and that the Crown is required to follow its advice. But the giving of advice, as such, is no more and no less than the Privy Council’s legal duty. And if the advisors fail in the discharge of this duty by offering advice that is unlawful or unconstitutional in itself (e.g. because it was procured corruptly) or because it will result in the Crown breaking the law, they can be held accountable for that:

Ministerial advice to the Crown, though triggered by convention, is given in the performance of a legal duty by ministers who hold a legal office, and the common law has always regarded ministers as legally responsible for the advice they give. Ministerial advice does not operate within a legal black hole. Of course, the old authorities … contemplate either a criminal prosecution or an action in tort, with examples of fraud, corruption, malicious abuse of public office, or flagrant neglect of public duties being given. I am not suggesting that the law historically acknowledged a modern remedy of judicial review of ministerial advice based upon public law principles of rationality, legality and fairness. I do think, however, that once the domain of ministerial advice is understood to be one governed by law, the forms of law that discipline advice must be understood in the usual legal way as expanding with the incremental developments that have shaped modern public law generally. (39)

I find this compelling. And I think that the UK Supreme Court’s judgment in R (Miller) v Prime Minister [2019] UKSC 41, [2020] AC 373 (Miller II) can be understood as giving effect to this approach to ministerial accountability. This was a case where the UK Supreme Court treated as reviewable the advice given by the Prime Minister to the Queen to prorogue Parliament. As the Court noted,

the power to order the prorogation of Parliament is a prerogative power: that is to say, a power recognised by the common law and exercised by the Crown, in this instance by the sovereign in person, acting on advice, in accordance with modern constitutional practice. It is not suggested in these appeals that Her Majesty was other than obliged by constitutional convention to accept that advice. In the circumstances, we express no view on that matter. That situation does, however, place on the Prime Minister a constitutional responsibility, as the only person with power to do so, to have regard to all relevant interests, including the interests of Parliament. [32]

The appointment of Superior Court judges is not, to be sure, a prerogative power, but one set out in constitutional legislation, but I don’t think that changes the situation. If Canadian courts follow the Miller II approach, then the decisions of the persons and entities with the constitutional responsibility to have regard to the relevant interests, in this instance those of the judiciary and of the litigants who come before it, should be amenable to review, because there can be no suggestion that the Governor General is “other than obliged by constitutional convention to accept [their] advice” ― or, more to the point, prevented from acting in the absence of advice.

Now let’s consider the argument for the justiciability of constitutional conventions. For my part, I have long found the orthodox account of conventions as radically different from legal rules quite unpersuasive. This is not because conventions are judge-made rules, as Brown J claims, but rather because the distinction between rules that have a judicial origin and those that originate in politics simply does not matter very much. Statutes originate in politics too, after all. And of course common law rules aren’t always “judge-made” in the crude Benthamite sense. Many originate in the practices of other actors, commercial ones for example, and/or in the principles embedded in the legal system. Courts can identify such rules ― though, unlike Brown J, they must do so rigorously and carefully ― and apply them.

The trouble with this argument is that most people don’t agree with it. More importantly, apex courts don’t agree with it. As I am about to suggest, they no longer adhere to the orthodox rigid distinction between law and convention either, but nor are they prepared to simply accept what the Patriation Reference, Re Resolution to Amend the Constitution, [1981] 1 SCR 753, described as the crystallization of convention into law. The majority in the Patriation Reference is clear about this, and has never been directly repudiated in Canada or elsewhere. Indeed it was endorsed by the UK Supreme Court in R (Miller) v Secretary of State for Exiting the European Union [2017] UKSC 5, [2018] AC 61, and it too hasn’t been directly repudiated. I can and will go on thinking that these cases are wrong, but a judge bound by one or the other cannot act if it didn’t exist.

But that’s not the end of the matter either. While things are not especially clear, Reference re Senate Reform, 2014 SCC 32, [2014] 1 SCR 704 may suggest a workaround to the dichotomy between law and convention asserted in the Patriation Reference. In that opinion, the Supreme Court relied on the concept of a “constitutional architecture”, which has something to do with “[t]he assumptions that underlie the [constitutional] text and the manner in which the constitutional provisions are intended to interact with one another”. [26] A law that would interfere with this architecture is as unconstitutional as one that contravenes an explicit provision of the constitution. Because the Court’s definition of “architecture” is sparse if not cryptic, there are different views on what it encompasses. I have argued that “architecture” is, simply put, a codeword for conventions. Others who have written on the subject may take a somewhat different view of the matter, but at any rate I think there is a credible argument to be made that advice, or perhaps even more so lack of advice, to the Governor General is, in principle, judicially reviewable if it would result in the undermining of “assumptions that underlie the text” of the constitution, regardless of the involvement of conventions in the process.

So where does that leave us? Recall that, for now, we have only been considering the first issue in Hameed: whether the government’s failure to ensure that judicial vacancies are filled in a timely fashion is in principle amenable to judicial review. I have argued that the answer is yes so far as Governor in Council appointments to the Federal Court are concerned, in the same way as failure to perform a statutory duty by any official or entity. The case of Governor-acting-on-advice appointments to the Superior Courts is less clear, but there are strong arguments for their being reviewable too, either on the basis of legal accountability for unlawful advice expounded by Professor Walters and demonstrated in Miller II, or through the invocation of the constitutional architecture referred to in the Senate Reform Reference.

Assuming that the government’s slowness in appointing judges is reviewable in principle, the question nevertheless arises whether there is legal standard by which it can be judged. Or is the pace of judicial appointments the kind of policy or discretionary matter that can only be subject to political, but not legal, accountability? (It is worth emphasizing that political accountability is important and should be pursued regardless of whether legal accountability is (also) available!) As I wrote in my last post, Brown J’s invention of a convention of prompt appointments to serve this purpose is indefensible. Even if violations of conventions are justiciable, they have to be real conventions that exist out there in the political world, and not judicial concoctions. Is there something else?

This is not obvious, to say the least. It is entirely fair to note that both the Constitution Act, 1867 and the Federal Courts Act make the appointment of judges pursuant to their respective provisions mandatory, using the word “shall”, and that, as Brown J suggests, it would be disturbing if the executive could undermine legislative choices as to the number of judges by refusing to fill the judicial positions created by provincial legislatures and Parliament. At the same time, no one, including Brown J, actually believes that any vacancy existing for any length of time is ipso facto a constitutional or statutory violation. The mandatory language in the relevant provisions is, it would seem, subject to an implicit qualification accepting at least some degree of churn ― and executive discretion.

That said, this discretion cannot be unlimited either ― Roncarelli v Duplessis, [1959] SCR 121, tells us this much. A Roncarelli-like scenario, where the executive misused its discretion for an improper purpose ― for example to punish a recalcitrant court by refusing to make appointments ― should be easy for a court to deal with. This is not a far-fetched hypothetical: think, again, about the Alani litigation concerning the Senate, and also of the delays in making appointments to the Supreme Court after the debacle of l’Affaire Nadon. But I don’t think that there is any suggestion of improper motive in the current situation; that would be, to quote the one and only Sir Humphrey Appleby, to mistake lethargy for strategy.

Professor Walters suggests what might be a further constraint on advice-giving discretion:

at least in those cases where the reasons for advice are publicly given or can otherwise be established through evidence … the question of whether those reasons are consistent with constitutional values and structure may well be, in the appropriate circumstances, a question of law for the courts to determine. (40)

But that doesn’t help much. The whole problem in Hameed is that there is no advice, and seemingly no specific reasons for the lack of any. Again, lethargy, not strategy. Moreover, even apart from that, a government may be able to argue (though I take it that it has not in fact argued) that its slowness in appointing judges is due to the difficulties it encountering in balancing relevant values, such as diversity, with the operational needs of the judiciary. If the reviewing court is inclined to be deferential ― and, as the Supreme Court reminded us in Commission scolaire francophone des Territoires du Nord-Ouest v. Northwest Territories (Education, Culture and Employment), 2023 SCC 31, courts are supposed to be deferential to executive value-balancing, whether that makes any sense or not ― then this may well be enough to let the executive off the hook.

Lastly, there is the Miller II move: there, the UK Supreme Court said that prerogative powers are implicitly limited by constitutional principles, such that any use of the prerogative that undermines relevant principles needs to be justified by the executive as being proportional to the interference with the principles in question. There is at least an argument for treating constitutionally codified powers in the same manner as prerogative ones ― though I’m not sure that’s right. If this argument is accepted; and if, further, the Miller II reasoning is to be taken at face value; then it might work in the specific circumstances of Hameed, the government having, rather like in Miller II, not bothered to justify itself. But, for my part, I am not at all convinced that Miller II should be taken at face value: I have argued that, instead, what it really did was to enforce a putative constitutional convention governing the permissible length of prorogations, identified pursuant to an implicit application of the same sort of test I described in my last post. If that is so, then we are back to square one, since there is no plausible constitutional convention to be enforced here.

So there you have it: I think that failure to ensure that judicial appointments take place to fill the vacancies on Canadian courts is, in principle, judicially reviewable, but there is probably no standard by which timeliness of appointments, as opposed to something like an outright refusal to make any, could be assessed, which makes the question ultimately non-justiciable. That said, I am more confident about the first part of this conclusion than the second ― that is, about reviewability-in-principle than about the lack of standards. This case is complicated, and closer than I thought at first.

What isn’t close is the underlying question of what needs to be done. The government should just get on with making those judicial appointments. They may well have sound arguments for avoiding being told to do so by a judge. But this should never have come to litigation in the first place. The Alani case ultimately became moot when the government accepted its constitutional responsibility for appointing Senators. Ideally, Hameed should be concluded in the same manner.

Putting the Courts Together Again

The Federal Court orders the government to get on with appointing judges. But its judgment is very badly flawed.

Yesterday, the Federal Court delivered a decision declaring that the Canadian government has a constitutional duty to ensure timely judicial appointments and setting out specific targets to attain: Hameed v Canada (Prime Minister), 2024 FC 242. The outcome is a startling one, but it is Justice Brown’s reasons that are truly remarkable, and not in a good way. The outcome, indeed, might even be defensible, at least in part, though I doubt it. But Brown J’s reasons are pernicious.

Hameed is something of a constitutional law professor’s crazy exam hypothetical. The bottom line is simple enough. The federal government has long been slow to appoint judges to vacancies that arise, despite warnings from the legal community and indeed entreaties from various institutions and leaders, including, notably, a strongly-worded letter from the Chief Justice of Canada. Is this tardiness unconstitutional on the basis that the government has a duty to fill judicial vacancies promptly as they arise? But to answer this question, or more precisely for the federal court to answer this question, an impressive number of issues need to be addressed, from standing, to federal court jurisdiction, to evidentiary issues, to the nature of the alleged duty, to the remedy. And pretty much each of these issues falls somewhere on the spectrum from not-altogether-straightforward to really quite complex.

The substantive heart of the case, however, has to do with just two issues. First, is the government’s failure to ensure timely judicial appointments judicially reviewable in principle? And second, if it is, is there a judicially administrable standard by which the constitutionality of government inaction regarding judicial appointments might be assessed? This framing, to be clear, is mine, not Brown J’s, and heavily influenced, on the one hand, by the UK Supreme Court’s decision in R (Miller) v Prime Minister [2019] UKSC 41, [2020] AC 373, and on the other, by Mark Walters’s illuminating article on “Judicial Review of Ministerial Advice to the Crown“, neither of which is cited in Hameed despite their relevance.

In this post, I describe Brown J’s approach to these two issues and explain why it is wrongheaded. In a follow-up post, I will explain why, nonetheless, the outcome of the case on a proper approach to these issues is a close issue that deserves much more careful thought than Brown J has given it.

Whether the government’s role in ensuring timely judicial appointments is justiciable at all, let alone in the Federal Court, is a difficult issue because of the involvement of constitutional conventions. Brown J draws on the Federal Court’s discussion of these conventions in Democracy Watch v. Canada (Attorney General), 2023 FC 31, which found that

By constitutional convention, when appointing judges to provincial superior courts, the Governor General acts on the advice of the Committee of the Privy Council of Canada. Similarly, the [Governor in Council], which appoints judges to the Federal Court, the Federal Court of Appeal, and the Tax Court of Canada, is defined in the Interpretation Act … as the Governor General acting on the advice or consent of the Privy Council for Canada. The Privy Council is composed of all the ministers of the Crown, who meet in the body known as Cabinet … As such, all federal judicial appointments are made by the Governor General on the advice of Cabinet. In turn, Cabinet acts on the advice of the Minister of Justice … In the case of appointment of Chief Justices and Associate Chief Justices, it is the Prime Minister who provides the advice to Cabinet. [9]

Unfortunately, this description contains two important mistakes. The most glaring but less consequential one lies in the description of the Privy Council as “composed of all the ministers of the Crown, who meet in the body known as Cabinet”. As Professor Walters explains,

The Privy Council for Canada is established by section 11 of the Constitution Act, 1867 and consists of all present and former ministers of the Crown as well as a variety of other members. … [I]t is only by virtue of convention that the legal powers conferred upon the “Governor in Council” are exercised by the Governor and present cabinet ministers. (34; emphasis in the original)

In other words, the Cabinet is the political, conventional entity through which the legal body that is the Privy Council acts; they are not the same. The second mistake is more subtle but more consequential: it consists in conflating judicial appointments by the Governor General and those by the Governor-in-Council. I will return to it in the follow-up post, though you will know what I’ll say there if you read Professor Walters’s article in the meantime.

Back to Brown J’s reasoning. On the orthodox view of Westminster-type constitutions, constitutional conventions are, in Dicey’s famous words, “not in reality laws at all since they are not enforced by the courts”. Is it the case, then, that in asking the court to require the Cabinet to provide its conventionally-mandated advice to the Governor General, the applicant is seeking to have a convention enforced, which cannot be done?

Brown J says no. He cites the Patriation Reference (though he at times misnames it as the Repatriation Reference), Re Resolution to Amend the Constitution, [1981] 1 SCR 753, for the proposition that while courts cannot enforce a convention, they can issue a declaration stating its import. Moreover, the conventions regulating judicial appointments

form part of Canada’s federal constitutional common law in the sense they are judge-made rules which the courts are entitled and may recognize in the appropriate case through the Court’s declaratory power, notwithstanding they are not laws that may been [sic] enforced by the courts. [122]

A breach of this “federal constitutional common law” by the Cabinet is reviewable by the Federal Court, in the same way as a breach of any federal law by a federal official is reviewable under the Federal Courts Act.

This is fundamentally wrong. As a matter of positive law as well as learned orthodoxy, it is absolutely not the case that conventions are a form of “constitutional common law”. Indeed, the majority opinion on the legal question in the Patriation Reference explicitly rejects this very view. In dismissing the idea of a “crystallisation” of conventions into legal rules, it says:

The leap from convention to law is explained almost as if there was a common law of constitutional law, but originating in political practice. That is simply not so. What is desirable as a political limitation does not translate into a legal limitation, without expression in imperative constitutional text or statute. (784)

As I will explain in the follow-up post (and as longtime readers will know), I am no fan of the Patriation Reference legal question majority. But it is binding on Brown J, whether he is a fan of it or not! He does not refer to this passage, which strikes me as very poor judicial craft. Moreover, the one thing that the Patriation Reference unquestionably does get right is that conventions “originat[e] in political practice”; in no way, shape, or form are they “judge-made”. As the UK Supreme Court put it in R (Miller) v Secretary of State for Exiting the European Union [2017] UKSC 5, [2018] AC 61 “[j]udges … are neither the parents nor the guardians of political conventions”. [146] One can object to the “guardians” bit; but it undeniably true that judges are not the “parents” of constitutional conventions.

And the idea of common law rules that aren’t enforceable by courts makes no sense at all; statutory provisions can be made unenforceable (as parts of the Official Languages Act used to be), but the common law doesn’t exist apart from judicial enforcement. If conventions are just common law rules, then they are in principle enforceable, subject to justiciability concerns in specific cases, and not merely declarable.

Brown J’s treatment of the issue of the standard by which to assess the government’s inaction is, if anything, even more cavalier. He points out that “Parliament has determined what it considered an appropriate number of judges required by the Superior Courts, including the Federal Courts, …. in legislation authorizing that number of appointments” and that failure to appoint these judges not only undermines the judiciary but also amounts to “Canada’s executive government … ignor[ing] the express will of Parliament”. [126] I have no quarrel with that ― on the contrary, this is an important point for, at the very least, the government to consider, whether or not it can also give rise to judicial remedies. But what follows is not so sensible.

The key passage deserves to be set out at some length. Brown J writes that

[T]he acknowledged constitutional convention that it is the exclusive authority of the Respondents to advise in respect of vacancies necessarily implies the related constitutional convention that judicial vacancies must be filled as soon as possible after vacancies arise, except in exceptional circumstances.

In this connection, nothing suggests Democracy Watch, which affirmed the existence of the convention, is the last word on the subject. …

[T]he Court should now recognize that the relevant constitutional conventions include not only the responsibility to take steps to fill vacancies as soon as possible, but in this appalling and critical situation, to materially reduce the present backlog to what it was as recently as the Spring of 2016, that is to reduce the vacancies to the mid-40s across the federally appointed provincial Superior Courts and Federal Courts. [129-131]

This is just terrible. A court cannot simply will “conventions” into existence because they would be useful. The most charitable interpretation I can give to Brown J’s approach here is that he is getting high on his own supply, namely on the confusion between conventions and “judge-made” common law. But, pace Bentham, even those who would describe the common law as being “judge-made” (not an uncontroversial description), would not countenance such blithe creation of new common law rules for no other reason than their seeming desirability. At the very least, there are concerns about the limits of the judicial role in the development of the law that must be addressed.

If one accepts that it is appropriate for courts to address constitutional conventions, then one must adopt a rigorous approach to identifying them. Indeed, this is no different from what a court will do when dealing with common law or statutory rules ― it will carefully analyse precedents or interpret the legislature’s words, and not just say that a rule must exist to deal with this or that exigency. The methods for identifying the relevant rule vary depending the rule’s type, but not the need for a method.

There are two methods for identifying conventions. The easy one is by relying on authorities ― one can think, for example, of Andrew Heard’s magisterial book Canadian Constitutional Conventions. The harder one, given the imprimatur of the majority opinion on the conventional question in the Patriation Reference, is a test first developed by Sir W. Ivor Jennings, which asks whether there are precedents for the alleged conventional rule, whether the constitutional actors involved followed the rule out of a sense of obligation, and whether there is a constitutional reason for the rule.

Brown J doesn’t even pretend to follow either of these methods for identifying the purported convention requiring prompt judicial appointments. He points to no authority. Nor does he identify any precedents. This case, of course arises precisely because the government evidently does not follow the alleged rule and does not consider itself bound to do so. Now, that is not fatal: the Patriation Reference also arose out of the actions of a government that refused to recognize and be bound by a conventional rule. But there the Supreme Court pointed to a long course of past practice and acknowledgment of its binding character. That is entirely missing from Brown J’s reasons.

He thinks that he can simply conjure up a convention nobody’s thought of before, because past decisions don’t get “the last word on the subject”. But the question isn’t whether a case is “the last word”. It’s what conventions actually exist out there in the real world. It is bad enough when courts start “giv[ing] constitutional benediction” to new rights whose time they think has come, as the Supreme Court did in Saskatchewan Federation of Labour v Saskatchewan, 2015 SCC 4, [2015] 1 SCR 245. But awful as that was, still the words of the constitutional text, however distorted from their original meaning, provide some modest measure of constraint even on wilful interpreters. If courts can give benediction to dreamt up conventions, there really isn’t anything they cannot add to the constitution just because they feel like it.

I cannot help but think that Brown J was so completely blinded by what he perceived as the practical necessities of the situation as to cut every corner imaginable on the way to a result of whose vital importance he was convinced. He begins his judgment by asserting that “[a]t its core, this matter concerns” the Chief Justice’s letter to the Prime Minister regarding the pressing need for judicial vacancies to be filled. [1] He is determined not to “allow the current untenable and crisis number of vacancies to remain unacceptably high with the negative consequences set out in the letter”. [126]

But that’s not how this judging business works. A court case is never about a letter sent by a non-party. I hesitate to speak to the letter’s admissibility, not only because this post is long but also, and more importantly, because the law of evidence is not my forte. But let me just note my puzzlement at Brown J’s repeated reference to that letter as “expert opinion” [12] and indeed “expert evidence” [124] when the Chief Justice wasn’t a witness. Be that as it may, a court case can only ever be about the law. Brown J lost sight of that altogether, and produced one of the more staggering judgments I can think of in a long, long time. (I should note, in fairness, that, assuming his account of the parties’ arguments is complete, he may have been very poorly served by them, especially the applicant.)

The irony is that there would have been serious legal issues to discuss, and a decision based on careful legal argument would have much better served the cause that Brown J tries to advance. And it is not a bad cause. The slowness of the federal government’s judicial appointments has been discussed, so far as I remember, from its early days in office. When the Chief Justice and Brown J denounce the current state of affairs, they are not wrong. I can only hope that the government takes the underlying message Brown J is trying to send seriously. But if the Canadian judiciary is broken, or nearly broken, the way to put it back together again is not by judges taking the need to do so for a warrant to do it. “Something must be done” is not a legal argument. As I said, I will try to consider actual legal arguments in my next post.

And Again

Confidence, not head counts, is the key to responsible government

I thought I was done with dispelling government formation misconceptions, having responded in detail here and here to claims that the party winning a plurality of seats in the House of Commons was somehow entitled to form government even if it lacked an overall majority. But I return to this issue, briefly, to address the converse mistake: the claim, advanced by Patrice Dutil in a Macdonald-Laurier Institute Inside Policy essay, that “[r]esponsible government requires the support of the majority of the House of Commons”, so that “in order to form government in a parliament that is split among five parties, one of them must earn the support—at least the tacit support—from the others”. This is, at best, an oversimplification of the conventional position.

As explained in my previous posts, what responsible government requires is for the ministry to maintain the confidence of the House of Commons. Whether the ministry is doing this is assessed by its ability to win certain key votes: that on the Speech from the Throne, which outlines its legislative programme, at the beginning of a new session of Parliament, and then mainly “money votes” on taxes and spending, as well as anything that the ministry itself designates as a vote of confidence, or a motion of non-confidence brought by the opposition. This has a couple of important corollaries.

One is that at no point does the ministry actually need “the support of the majority”. It simply needs not to be outvoted. If it can win votes with less than a majority of MPs supporting it because even fewer are voting for the opposition, that’s fine. Maybe this is what Professor Dutil is alluding to this when he refers to “tacit support”, but I’m not sure that this is a fair description of all conceivable political arrangements. Tellingly, Professor Dutil’s review of political precedents runs from King’s minority governments in the 1920s to that of Joe Clark in 1979. It does not consider Stephen Harper’s governments from 2006 to 2011. These were kept in office, in part, by the opposition’s failure to vote against the Conservative government in sufficient numbers ― the opposition either abstained or made sure that too few of its members turned up to outvote the government. This did not involve any sort of lasting arrangement or even cooperation with the opposition, which was either in no position to fight an election campaign (being leaderless, penniless, or both) or afraid that an election would yield even worse results for it than the status quo. I don’t think this counts as support, even tacit, but it was enough to keep the government in office. In short, a government lacking majority support in the House of Commons can stay in office without making any sort of deal with the opposition, so long as the opposition does not bring its full numbers to bear to vote it down on a matter of confidence.

This brings me to the second corollary from the convention of responsible government outlined above. A government ― by whatever count or margin it wins the votes of confidence it must win to remain in office ― does not “require the support of the majority of the House of Commons” for anything else, constitutionally speaking. There is no impediment to a government remaining in office despite taking a loss on a legislative project or a symbolic motion, provided it has not been designated as a matter of confidence. And other than money votes and motions of no-confidence brought by the opposition, governments have the choice not to designate votes as implicating confidence. If they do so, that is a political choice for which voters ought to judge them. If a majority government whips its MPs and forces them to toe the party line on every vote, and not just on matters of confidence, that is also a political choice for which voters ought to judge it. If a government won’t allow a vote on a private member’s bill to be held, or won’t let it be treated as a free vote on which individual MPs are free to follow their conscience and/or their constituents’ wishes, that is a political choice for which voters ought to judge it.

Such judgments do not seem to be occurring in Canada, despite our governments controlling their MPs to a much greater extent than those in the UK or in New Zealand. I’m no political scientist and not in a position to speculate as to why that is the case. (One common hypothesis ― that with a smaller House of Commons than the UK, Canadian MPs are too blinded by the higher prospect of a ministerial position than their UK colleagues to show independence ― does not account for the greater prevalence of free votes in the much smaller House of Representatives in New Zealand.) But I think that misrepresenting the constitutional position, even by implication and even with the laudable intention of making it accessible to ordinary citizens, by suggesting that governments (always) need to be supported by majorities to remain in office does not help matters.

I may be wrong, of course, but I would like to think that the conventions of responsible government do not require oversimplification to be understood. A government needs to maintain confidence, which is assessed by its ability to win ― by whatever head count ― a limited set of votes in the House of Commons. All the other stuff that is being discussed these days ― pluralities, majorities, popular vote ― is largely or entirely irrelevant. It’s not that complicated.

Government Formation, Revisited

The orthodox understanding of the conventions of government formation is better than its populist rival

I am grateful to Charlie Buck for taking up my invitation to explain why the constitutional conventions of government formation in Canada have diverged from those recorded in the Cabinet Manuals of the United Kingdom and New Zealand, which I discussed here. However, I am not persuaded by his argument to the effect that, whatever may have been the case in Canada in the past, and still be the case in its fellow Commonwealth realms, our present conventions make the plurality party in the House of Commons uniquely entitled to govern.

There are several strands to Mr. Buck’s argument. First, conventions are political and flexible, and thus subject to revision from time to time as an “outgrowth of democratic impulses”. Second, the nature of Canadian politics today favours treating the plurality-winning party as endowed with a special governing mandate that any opposition coalition would lack. Third, recent (and perhaps not-so-recent) precedent favours this interpretation of the conventions of government formation. Finally, if the plurality party in the House of Commons is unable to govern in the face of a united opposition, the solution is a new election, rather than, potentially, an opposition takeover. At every step in this argument, Mr. Buck starts from truth, and quickly errs.

So far as the flexibility of conventions is concerned, there is no doubt that new conventions can add themselves to the constitution, as Mr. Buck’s example of the (arguable) convention requiring a referendum prior to any constitutional amendment illustrates. The present Prime Minister’s policy of only appointing bilingual judges to the Supreme Court is likely an attempt to create a convention, and there is nothing improper about it, though whether the attempt succeeds will depend on whether future Prime Ministers play ball.

Whether existing, established conventions can also decay or be replaced by something quite different strikes me as a more difficult question ― at least when the convention is as central to our constitutional order as those regulating government formation. At the very least, we should be wary of too quickly concluding that this has happened. This is because, if constitutional conventions are to serve as rules and as meaningful constraints on the behaviour of political actors, they cannot simply be whatever “partisans, acting on the wishes of their constituents” happen to believe. As I said in my original post, conventions must “have an existence independent of misrepresentation by either political partisans”. The issue here isn’t, contrary to what Mr. Buck suggests, whether we believe that conventions are different from law, let alone an academic’s self-serving wish to obtain “a monopoly on interpreting the nature of Canadian constitutional conventions”. It is that if the conventions are whatever a politician claims they are, then what we have is not a political constitution, but no constitution at all.

Besides, it is too easy to say that politicians “act on the wishes of their constituents”. Mr. Buck refers to no evidence of these wishes, other than the politicians’ own say-so ― and politicians just aren’t great at discerning their constituents wishes. At most, we have the unpopularity of the 2008 attempt by a coalition of opposition parties to oust the Conservative minority government. I will return to it below, but for now, suffice it to say that attributing it to an alleged belief that the plurality winner is entitled to govern is pure speculation.

What is not speculation, and needs to be taken into account, is that voters are generally ignorant of even basic constitutional rules, be they legal or conventional. This may very well result in their having wishes that are contrary to such rules. But that does not give the politicians license to break the rules. Even if a poll were to show unambiguously that Canadians would rather that a Governor General not assent to a bill passed by the House of Commons and the Senate, that would not justify the Governor General in withholding assent. Again, were it otherwise, we would not a political constitution, but no constitution at all.

I turn now to the claim that changes in the way politics are done and understood ― namely, the prominence of parties and party leaders at the expense of individual MPs ― favour abandoning the longstanding conventions of government formation in favour of a plurality-winner-take-all rule. The changes are real enough: I have discussed them here, in a post drawn from an article published in the McGill Law Journal (and based, for this part of the argument, on Bernard Manin’s historical and theoretical work and a close look at the 2011 election campaign in Canada). But they are not a compelling explanation, let alone justification, for an alleged change in the conventions of government formation in Canada.

For one thing, one has to wonder what took so long. These changes aren’t new. The emergence of political parties rather than individual parliamentarians as the dominant actors ― what Professor Manin describes as “party democracy” ― was well underway in the second half of the 19th century. The increased prominence of leaders due to their ability to speak directly to voters ― Professor Manin’s “audience democracy” takes off in the 1970s. Why would it be the case that government formation conventions only responded to these changes in the last 10, perhaps 15 years?

Perhaps even more strikingly, though: why would this only have occurred in Canada of all places? As Professor Manin explains, the tendencies to which Mr. Buck points are common to Western democracies. They certainly have not bypassed the UK and New Zealand. More than that: in the 1990s, New Zealand switched to a (mixed-member) proportional electoral system where the partisan composition of its House of Representatives depends almost entirely on the votes cast explicitly for a political party. To a much greater extent than Canada (and the UK), New Zealand has consecrated the dominance of parties over electoral politics in its law. Yet, as I have shown in my last post, New Zealand retains the traditional conventions of government formation, to whose operation the prominence of political parties and their leaders is evidently no obstacle. After the 2017 election, a Labour-led government under Jacinda Ardern took office, despite Labour winning 10 fewer seats than the National Party (in 120-member House!). There is nothing “nostalgic” about this; this is just the Westminster system operating as it always has.

Of course, Mr. Buck is right that “Westminster systems are also capable of diverging from each other”. But the whole point of the challenge I issued in my earlier post was to ask those who think like him to explain why they think the divergence has happened. What is different in Canada? Emmett Macfarlane usefully reminds us that the simple fact that plurality winners have always taken office after 1925 is not enough. Nor is Mr. Buck’s generic claim that “diversity in response to shifting environmental conditions is a strength not a weakness of the Westminster model”. What environmental conditions have changed in Canada in the way they have not in the UK and in New Zealand ― other, that is, than the self-serving claims of politicians?

This brings me to the question of precedents. Mr. Buck mentions three: the King-Byng Thing, the implosion of the 2008 opposition coalition, and the statements of major party leaders in 2015. None are conclusive.

King-Byng is inapposite, if not outright contrary to Mr. Buck’s point. If King was the good guy in that story, which I’m not particularly convinced of, it is worth recalling that he had formed a government despite Arthur Meighen’s conservatives having won more seats than his Liberals at the previous election. Not a problem then, I suppose. But more importantly, King-Byng is a very well known story throughout the Commonwealth ― I used to teach it in New Zealand, for instance ―, yet nowhere is it thought to stand for the proposition that the plurality party in the House of Commons (or its equivalent) is specially entitled to govern. It is, above all, a cautionary tale about the appearance of vice-regal partiality, and helps explain why the Cabinet Manuals of the UK and New Zealand are so insistent that the Sovereign or his representative are not to be drawn into government formation discussions, and that the politicians need to work out for themselves where the confidence of the House lies. As I pointed out in the previous post, none of that would be worth saying if the Sovereign simply needed to appoint the leader of the plurality party as Prime Minister.

As for the events of 2008, we know that the coalition was unpopular, and was unable to hold together long enough to actually oust the Prime Minister. But was that due to rejection of the principle of coalition governments without the participation of the largest party? Or to that particular coalition’s membership (notably its including the Quebec separatists)? To the personal unpopularity of its putative leader? To it being an attempt to change governments months after, rather than in the immediate aftermath of, an election? That, I don’t think we know.

The 2015 statements are a somewhat different matter, because they were seemingly explicit and on point. While I have struggled to find news stories to link to, I share Mr. Buck’s recollection of the events: “The leaders of all the major parties in 2015 [stated] that the largest party should form government”. What I do not share is his interpretation of these statements as concessions: hence the change I have made in quoting him. One needs to recall the peculiar dynamics of the 2015 election campaign, during much of which it was widely thought that all three major parties had a decent shot at ending up as plurality winners. The eventual result, an outright majority for one party, was unexpected until perhaps the final week if not the last few days of the campaign. In these circumstances, the leaders’ claims that the plurality winner ought to become Prime Minister were not concessions made against interest but demands for deference made in the hope that the would be in a position to claim this supposed mandate. As Professor Macfarlane notes, “we should be cautious about relying on statements by political leaders in the midst of election campaigns as if they reflect constitutional principle rather than political tactic”.

What the King-Byng and 2008 precedents can help with is the consideration of Mr. Buck’s final point: that, while it is legitimate for the opposition to deny a minority government the confidence of the House of Commons, the only legitimate consequence of such a denial is an election. The formation of an alternative government by the opposition is out. I want to credit Mr. Buck for making clear that a denial of confidence is legitimate. As I said in my earlier post and as I am about to explain again, I strongly suspect that the politicians who share his views on the conventions of government formation would inappropriately deny this. But I am not convinced that the only legitimate response to such a situation is a new election.

What 1926 and 2008 have in common is that they happened months after an election and after the government whose ouster was being contemplated had initially retained the confidence of the House of Commons. In such circumstances, a new election may well be the best response to a loss of confidence, though I would not yet say that this is a conventional requirement. Evidently the opposition leaders in both 1926 and 2008 did not believe that they were bound by any rule against mid-Parliament government changes, and what is politically unwise or even reckless is not necessarily unconstitutional, even in a conventional sense.

But the focus of the present debate about government formation, as I see it, is not so much on what might happen months after an election, but rather ― as in 2015 ― on an election’s immediate aftermath. If (like King’s liberals in 1925) the party of the incumbent government fails to win a plurality of seats, must it (unlike King) resign and give way to the new plurality winner? Or, if it does win a plurality (like New Zealand’s National in 2017), is it automatically entitled to continue in office despite the opposition joining forces to deny it confidence, (as National was not)?

I suspect that voters’ appetite or perhaps even tolerance for a new election in the immediate aftermath of the one that was just held would be nil, and understandably so. There would not be much reason to expect a second election to produce a different result from the first, and so provide a way out of the impasse. This is why I think that, although this is not Mr. Buck’s intention, adopting his interpretation of the conventions of government formation will inevitably lead to claims that it is illegitimate for the opposition to deny confidence to a government formed by the plurality party in the House of Commons. Note that, in 2008, the then-Prime Minister’s reaction to the coalition’s attempt to oust him was not to go to the Governor General and ask for a dissolution ― which, in light of King-Byng, may have been granted. It was to avoid facing a denial of confidence by the House of Commons and to cast aspersions on the legitimacy of such an eventuality.

By far the better view is the one captured by the UK and New Zealand cabinet manuals. When an election produces a hung Parliament ― i.e. one in which no one party has a majority in the House of Commons ― the politicians should work out where confidence lies. Ideally, this should be done by negotiation before the new House meets. Failing that, the incumbent administration, which of course remains in office throughout, is entitled to meet the new House and see if it can secure its confidence, which is tested by the vote on the Speech from the Throne. And if confidence is denied, which is perfectly legitimate, the opposition, whether or not it includes the plurality party, gets a chance to form a new government without an election needing to be held. Only if the alternative government cannot secure confidence either does an election become the only way out of the impasse and a constitutional necessity.

In short, there are no good reasons to think that the constitutional conventions of government formation in Canada have changed from what they have long been understood to be both in this country and elsewhere in the Commonwealth. There is no real evidence of the public understanding and demanding such a change; nor are the politics of an “audience democracy” any less suited to the orthodox understanding of conventions than those of earlier periods. Political precedents do not support the alleged change either. On the contrary, as explained both in my previous post and here, unsurprisingly, our political system will work better ― the legitimacy of opposition being preserved and wholly unnecessary and futile elections being averted ― with the orthodox understanding of conventions.

All that being so, both I in my previous post and Professor Macfarlane have been wondering why this whole debate is even happening. Part of the answer is excusable confusion. Part, I am afraid, is that, exactly as in 2015, some political actors see fanning this confusion as a shortcut to power. And part is that the populist spirit of the times is seeping even into good faith intellectual discussions, as Mr. Buck’s post illustrates. His appeal from the consensus opinion of “academics dictating” what the constitution ought to be to the authority of the voters “in this democratic age” is a familiar one. As with all such populist appeals, the irony is that allowing it would empower self-serving political elites at the expense of voters too poorly informed to see through it.

In the face of these populist tendencies, the role both of politicians and of academics is to remind the voters of the rules and the reasons these rules exist. Of course, academics have an easier time doing this because their self-interest is not nearly as implicated. But if politicians aren’t capable of doing what is right instead of what will win them some extra votes, what good are they? It is regrettable that Canada’s politicians failed in 2015, and may well fail again. But that only makes it more important for academics to hold the line.

Conventions from on High

A couple of weeks ago, I published a post arguing that the conventions of government formation in Canada did not give any special entitlement to the party winning a plurality of seats at a general election ― just like these same conventions in the UK and in New Zealand. That post invited responses from the proponents of the contrary view, and I am very pleased that one has taken up the challenge. His response is below. I will post a reply tomorrow.

Guest Post by Charlie Buck, PhD Student (Political Science), University of Toronto

In Canada, the party with the plurality of seats should form government. Professor Leonid Sirota mistakenly suggests Canada instead follow the rules of government formation outlined in the cabinet manuals of the United Kingdom and New Zealand.

He is hardly alone in holding a textbook understanding of government formation. Political scientists Peter Aucoin, Mark D. Jarvis, and Lori Turnbull argue the same in their award-winning monograph Democratizing the Constitution: Reforming Responsible Government. They blame the slide towards executive dominance of the legislature on the ambiguity that is at the heart of the constitutional conventions of responsible government. Prime ministers of all partisan stripes have routinely exploited the confusion that surrounds the governor general’s reserve powers over dissolution, prorogation, and summoning to run roughshod over parliamentary democracy. One need only cursorily point to Jean Chrétien’s knack for snap election calls or Stephen Harper’s use of prorogations to avoid opposition scrutiny, to see this prime ministerial abuse of conventional authority in action.

In the midst of their disillusionment with contemporary conventional practice, what most experts on law and politics appear to have forgotten is how those conventions materialize in the first place. Conventions are separate from legal rules (contra Professor Sirota’s doubtfulness that the two can be distinguished) precisely because they are politically rather than legally binding. A virtue that comes from this political nature is flexibility. As the written law is enforced through judges who are sensitive to precedent and may only modify what has already been settled through the ad hoc flow of real cases and controversies, the pace of legal change is slow, potentially glacial. Conventions’ responsiveness to the public mood tolerates a greater dynamism than rigid legal rules allow.

The enforcement of conventions takes place in the political arena; partisans, acting on the wishes of their constituents, are much more relevant to their operation than academics. That these academics are “out of touch” is beside the point. That they do not have a monopoly on interpreting the nature of Canadian constitutional conventions is not.

Other than their flexible nature, conventions hold the advantage over codified constitutional rules by being the outgrowth of democratic impulses. As has already been stated, it is not academics dictating the nature and scope of conventional rules (no matter how much they may wish otherwise), but the people acting through their own voices, the media, the ballot box, and their representatives. Academics and judges may observe what the conventions are in a given moment, but they do not get to ordain their meaning nor freeze that meaning for eternity. Professor Sirota is certainly correct that constitutional rules are rules and conventions are rules too. But he is wrong to collapse the differences between them.

The nostalgic understanding of government formation has evolved in this democratic age to mean that only the largest party can claim the legitimacy to form a government. It is not for Canada’s constitutional scholars to decide the rightness or wrongness of that transformation. Conventions are not just formulaic legal rules but normatively desirable governing principles. To quote A. V. Dicey, conventions are “a body not of laws, but of constitutional or political ethics.”

The ethics of transforming the standards of government formation makes some sense. Canadians no longer—if indeed they ever did—vote on the basis of the qualities of local candidates. Elections are now much more nationally driven, leader-centric, and partisan affairs. It is increasingly at odds with reality to think of the House of Commons as formed from 338 individuals representing their ridings rather than from leader-dominated parties.

In conjunction with the changing face of parliamentary representation, democratic accountability has risen to the forefront. From this so-called “decline of deference” has developed another convention surrounding an area of Canadian political life as vital as government formation: constitutional amendment. Many political scientists blame the closed-door process of negotiation for the ultimate failure of the Meech Lake Accord. Participants in the later Charlottetown round were cognizant of this fact and addressed it head-on by holding a national referendum, only the third held in Canadian history. Today, scholars such as Peter Russell and Richard Albert speculate this precedent has generated a new conventional rule that major constitutional overhauls require the direct consent of the Canadian people through national referendums. If the conventions governing constitutional amendment are open to modification as a result of democratic pressure, why not also the conventions that surround government formation?

With a demand for greater accountability and citizen participation in the workings of government, the confidence convention has evolved in the direction it has. The democratic sensibilities of Canadians now require, not just that governments command a confidence of the House cobbled together by backroom agreements and coalitions between party insiders, but that they do so while holding the most seats in the chamber. If the party with the most seats cannot muster the necessary majority to pass spending bills then an election must be called.

The exception proves the rule. The only time in Canadian history a governor general has not called an election at the insistence of a prime minister, in the famous King-Byng Affair, the Canadian people expressed their displeasure towards Governor General Byng by rewarding that prime minister with a plurality of seats in the subsequent election with which he was able to maintain the confidence of the House. This not only shows that the transformation of the confidence convention requiring a plurality of seats was a long time coming but also that political actors who violate the new conventional rule risk the wrath of the electorate.

It is telling on this front that politicians have followed the desires of the public, notwithstanding these many academic objections. The leaders of all the major parties in 2015 conceded that the largest party should form government, but this is a practice that goes back far longer. Every prime minister in Canadian history faced with a larger party in the legislature following an election resigned or asked for dissolution rather than try to continue on. Once again, the exception proves the rule. Canadians’ popular rejection of coalition government during the 2008-9 parliamentary crisis, the unravelling of the agreement between the Liberals, NDP, and Bloc, and the subsequent Conservative majority win in 2011, is a cautionary tale of what happens to parties finding themselves on the other side of the conventional line.

In the end, this dispute comes down to different understandings of what conventions are and should be. A discomfort with flexible, politically enforced rules appears to favour cabinet manuals like the ones in the United Kingdom and New Zealand. The antidote to ambiguity and political constitutionalism is the codification of clear-cut rules in cabinet manuals. An understanding of conventions on these terms is at cross-purposes with their virtues, namely, their dexterity and democratic enforceability.

In no way does this difference of opinion cloak a hidden partisan agenda. As Professor Sirota quite rightly says, the opposition, including one constituting a majority of the House, may properly oppose the government’s agenda. If this results in a loss of confidence then an election should be held and the resulting plurality (or majority) party given the opportunity to govern.

Far from something to be condemned, the debate over responsible government’s organic evolution into something other than originally understood by the framers of the British North American Act puts into sharper focus the remarkable ability of Westminster systems to adapt their governing arrangements to changing circumstances. Westminster systems are also capable of diverging from each other, as the historical experiences of Canada, the United Kingdom, New Zealand, and Australia make plain. This diversity in response to shifting environmental conditions is a strength not a weakness of the Westminster model. In the populist, turbulent moment we are living through, governing arrangements require an elasticity capable of adjusting to the people’s wants and desires.

Rules Matter

Why the rules of government formation are the way they are, and why they matter

For reasons quite beyond my comprehension, Canada is apparently having a “debate” about the rules applicable to government formation in the event an election does not produce a single-party majority in the House of Commons. Philippe Lagassé has covered the essential principles ― as they have long been accepted, certainly up until, roughly, the summer of 2015, and for the most part beyond that too. Contrary to the claims of some Conservative partisans ― and, if memory serves well, to those of the leaders of all three major federal parties during the 2015 election campaign ― winning a plurality of seats (i.e. being the single biggest party in the House of Commons) does not entitle a party to form government or its leader to become Prime Minister. Only an outright majority makes this automatic.

Of course, my saying that Professor Lagassé is right won’t persuade people who dismiss his views and hallucinate about him writing out of a partisan agenda. I don’t suppose they are open to persuasion at all, really. But here is something that Professor Lagassé does not discuss, which I think bears pointing out. Canada’s conventions of responsible government, including government formation, are not special or unique. On the contrary, they at the heart of what the framers of the Constitution Act, 1867 meant when they explained, in the preamble, that they were designing a “Constitution similar in Principle to that of the United Kingdom”. The UK follows similar conventions; so does New Zealand. Unlike Canada, though, both the UK and New Zealand have recorded the relevant conventions in authoritative (though not legally binding) Cabinet Manuals. One would think that, given the deliberate similarities between our constitutional systems, these statements ― made by people who have no conceivable interest in our partisan squabbles ― are worth considering.

The UK’s 2011 Cabinet Manual begins by explaining that “[t]he ability of a government to command the confidence of the elected House of Commons is central to its authority to govern.”(2.7) This is crucial point; I will return to its significance below. The Cabinet Manual adds that “[c]ommanding the confidence of the House of Commons is not the same as having a majority or
winning every vote”. (2.7) Note that there is no mention of a plurality here. The next paragraph notes that

Prime Ministers hold office unless and until they resign. If the Prime Minister resigns on behalf of the Government, the Sovereign will invite the person who appears most likely to be able to command the confidence of the House to serve as Prime Minister and to form a government. (2.8)

This has a number of important consequences, which Professor Lagassé explains, but I’ll simply reiterate the point that likelihood of commanding the confidence of the House of Commons is decisive; again, there is no mention of a plurality of seats mattering. This is confirmed in the next paragraph, which notes that, in ascertaining “who appears most likely to be able to command the confidence of the House”,

the Sovereign should not be drawn into party politics, and if there is doubt it is the responsibility of those involved in the political process, and in particular the parties represented in Parliament, to seek to determine [this] and communicate clearly to the Sovereign. (2.9)

Of course there would be no need for warnings about drawing the Sovereign into party politics if all that mattered were who won the most seats.

All this is again reiterated in paragraphs 2.12 and 2.13, under the heading “Parliaments with no overall majority in the House of Commons”. Finally, the Cabinet Manual outlines the possible outcomes:

The nature of the government formed will be dependent on discussions between political parties and any resulting agreement. Where there is no overall majority, there are essentially three broad types of government that could be formed:

- single-party, minority government, where the party may (although not necessarily) be supported by a series of ad hoc agreements based on common interests;
- formal inter-party agreement …
- formal coalition government, which generally consists of ministers from more than one political party, and typically commands a majority in the House of Commons … (2.17)

This is a bit of an oversimplification, because these options aren’t mutually exclusive (one can imagine, say, a coalition of two parties working on an ad hoc basis with a third one), but the essential points are, first, that a single-party minority is not inherently more legitimate than a coalition, and, once again, that winning a plurality of seats doesn’t enter into the discussion at all.

New Zealand’s Cabinet Manual 2023 is to the same effect. It notes that

The Prime Minister is appointed by warrant by the Governor-General. In making this appointment, constitutional convention requires the Governor-General to:

1. act on the outcome of the electoral process and subsequent discussions between political parties. These discussions ascertain which party, or group of parties, appears able to command the confidence of the House of Representatives (expressed through public statements) and therefore has a mandate to govern the country; and
2. act on the outcome of the political process by which the person who will lead the government as Prime Minister is identified. (2.2)

Once again, there is no mention of the winner of a plurality of seats in the House of Representatives having any significance. On the contrary, what is highlighted is “discussions between political parties” that are “subsequent” (my emphasis to “the electoral process”. Again this is confirmed in a later chapter, which states

The process of forming a government is political, and the decision to form a government must be arrived at by politicians. Government formation may involve one or more parties.

In a case where government formation involves multiple parties, once the political parties have reached an adequate accommodation, and it is possible to form a government, it is expected that the parties will make appropriate public statements of their intentions. …

By convention, the role of the Governor-General in the government formation process is to ascertain where the confidence of the House lies, on the basis of the parties’ public statements, so that a government can be appointed. (6.42-6.44)

As with the similar UK provisions, there would no need for any of this is the Governor-General simply had to invite the leader of the single biggest party to form the government.

So here are my questions to those who claim that the rule in Canada is different: when did Canada’s constitution cease to be “similar in principle” to that of the mother country and our Commonwealth siblings? For what reasons did that divergence happen? And what useful purposes does it serve? If anyone wants to offer a response, I’ll be happy to publish it as a guest post.

Meanwhile, let me return to the key point from the UK Cabinet Manual: “[t]he ability of a government to command the confidence of the elected House of Commons is central to its authority to govern.” Being Prime Minister means nothing if you are unable to actually govern, which requires, at a minimum, securing money from Parliament (and more specifically from the House of Commons, whence “Bills for appropriating any Part of the Public Revenue, or for imposing any Tax or Impost” are required to “originate” pursuant to s 53 of the Constitution Act, 1867) to keep the lights on. Ideally, securing the House’s support for key legislation is part of it too, though not, as the UK Cabinet Manual notes, winning every vote.

A party that has nothing more than a plurality ― especially, but not necessarily, a narrow plurality of seats in the House of Commons may fail to secure funds and support for legislation from the rest of the House. In that case, it will lack any “authority to govern”, or indeed any ability to do so. Having its leader as Prime Minister will have as much significance as being an Admiral in the Great Navy of the State of Nebraska. This is why the ability to command the confidence of the House of Commons ― that is (oversimplifying a bit), to secure its support in votes on budgetary matters and some other key issues ― is the fulcrum of the rules on government formation, as explained by Professor Lagassé and as recorded in the UK and New Zealand cabinet manuals.

Some of the people who are claiming that winning a plurality of seats in the House of Commons is all that matters are not thinking this through because they do not understand what governing in our constitutional system actually involves. But others ― and presumably those who manufactured this whole “debate” are among them ― surely do understand, and have a further agenda. They would not be content to have a Prime Minister take office without being able to do anything with it. Instead, they would argue that their man is entitled not to be opposed by a hostile majority of Parliament, and that opposition (if not Opposition) is antidemocratic and illegitimate.

This is the end game here. We will be told that the majority of the House of Commons needs to make itself scarce, figuratively at least. And, to add insult to injury, we will be told that this is in the name of democracy and the will of the people.

I don’t suppose any disclaimers of partisanship would be particularly useful at this point. My regular readers can surely tell where my antipathies lie (everywhere, really). Others wouldn’t believe me anyway. Like Professor Lagassé, I lament the difficulty, indeed the impossibility, of “good faith discussions with partisans” who “cast academic analyses as ideological disagreements, or to think there’s a scholarly plot afoot, if our views don’t align with partisan objectives”.

Maybe because I am lawyer and not a political scientist, though, unlike Professor Lagassé I am not quite content to “understand why an academic perspective might be dismissed as irrelevant or out of touch”. Constitutional rules are, well, rules, and the fact that the person reciting them is allegedly “out of touch” is beside the point. This is as true of conventions as it is of legal rules (assuming there is a difference between the two, which I doubt). Rules matter, and they have an existence independent of misrepresentation by either political partisans or Supreme Court judges. More often than not, this blog is devoted to pointing out the latter. But now there is seemingly a need to sound the alarm about the former. Consider yourselves warned.

Keeping Faith

A master class in public meaning originalism, delivered by the US Supreme Court’s Justice Elena Kagan

Earlier this week, the Supreme Court of the United States delivered its decision in Chiafalo v Washington, upholding the constitutionality of a state statute imposing fines on “faithless” presidential electors ― those who do not vote for the candidate who won their state’s popular vote. The majority judgment, given by Justice Kagan for a seven-judge majority (and indeed unanimous on some key points), should be of some interest to Canadian readers for what it says about constitutional interpretation and, in particular, about the role of conventions and practice. As others, notably Josh Blackman over at the Volokh Conspiracy, have noted, Justice Kagan’s opinion is a thoroughly, and intelligently, originalist ― which should remind skeptics of originalism inclined to dismiss it as a partisan affectation that it is not.

As Justice Kagan explains,

Every four years, millions of Americans cast a ballot for a presidential candidate. Their votes, though, actually go toward selecting members of the Electoral College, whom each State appoints based on the popular returns. Those few “electors” then choose the President. (1)

But what is it that ensures that the vote of the Electors is aligned with that of the electorate? The text of the Constitution of the United States says little on this. Article II, §1, cl 2 provides that

Each State shall appoint, in such Manner as the Legislature thereof may direct, a Number of Electors, equal to the whole Number of Senators and Representatives to which the State may be entitled in the Congress: but no Senator or Representative, or Person holding an Office of Trust or Profit under the United States, shall be appointed an Elector.

Nothing here suggests that the Electors are bound to follow the popular vote; indeed, nothing here suggests that a popular vote need be held at all. At least some of the framers of the Constitution expected the Electors to exercise their personal discernment in choosing the President. Alexander Hamilton’s vision, in Federalist No. 68, is the best known. He hoped that the President would be chosen

by men most capable of analyzing the qualities adapted to the station, and acting under circumstances favorable to deliberation, and to a judicious combination of all the reasons and inducements which were proper to govern their choice. A small number of persons, selected by their fellow-citizens from the general mass, will be most likely to possess the information and discernment requisite to such complicated investigations.

But this is not how things worked out. As Justice Kagan puts it, “[a]lmost immediately, presidential electors became trusty transmitters of other people’s decisions”. (13) This was the result of the development of political parties, not anticipated ― indeed feared ― when the Constitution was being drafted. George Washington was elected without meaningful opposition. But, once he retired, presidential elections were contested by parties. As Justice Kagan explains, initially

state legislatures mostly picked the electors, with the majority party sending a delegation of its choice to the Electoral College. By 1832, though, all States but one had introduced popular presidential elections. … At first, citizens voted for a slate of electors put forward by a political party, expecting that the winning slate would vote for its party’s presidential (and vice presidential) nominee in the Electoral College. By the early 20th century, citizens in most States voted for the presidential candidate himself; ballots increasingly did not even list the electors.

The alignment between the popular and the electoral votes (within each State, of course, there being, as we know, no necessary alignment at the national level) was thus secured by a combination of State law and partisanship ― but also by what looks, to an observer based in a Westminster-type constitutional system, an awful lot like constitutional convention. Law allowed partisans to be appointed as electors, and partisanship motivated them to vote for their party’s candidate. But so too did a sense of propriety, of moral obligation. This moral obligation, explains why those electors who, from time to time, broke with their party were called “faithless”. There is normally nothing “faithless”, except to a rabid partisan, about putting country ahead of party. But something greater than partisanship is at stake in the presidential election ― nothing less, indeed, than democratic principle itself. And “convention” is what Westminster systems call the settled practice of constitutional actors rooted in constitutional principle.

Some States, though, felt that relying on convention was not enough, and legislated to back up the electors’ moral duty with a threat of punishment. According to Mr Chiafalo, they could not do so constitutionally. After all, the Constitution’s framers meant for them to exercise their own judgment, guided but not fettered by that of the voters. And the very vote “elector” connotes the exercise of a personal choice.

Not so, says Justice Kagan. For her, “the power to appoint an elector (in any manner) includes power to condition his appointment—that is, to say what the elector must do for the appointment to take effect”. (9) A “demand that the elector actually live up to his pledge, on pain of penalty” (10) is nothing more than a condition of appointment, which nothing in the Constitution’s text prohibits. Justice Thomas, concurring (with the agreement of Justice Gorsuch), disagrees with this approach. For him, imposing such conditions is not part of the original meaning of the power of choosing the “manner” of the electors’ appointment. Instead, the States’ ability to do so comes from the structure of the Constitution, which preserves their powers unless expressly limited, and from the Tenth Amendment, which codifies the same principle. Justice Thomas makes some compelling points, but this disagreement is not so important for Canadian readers ― or, for that matter, for practical purposes.

What matters most is Justice Kagan’s firm rejection of an appeal to the purported authority of the Framers’ supposed expectation that “the Electors’ votes [would] reflect their own judgments”. (12) This rejection is firmly rooted in original public meaning originalism:

even assuming other Framers shared that outlook, it would not be enough. Whether by choice or accident, the Framers did not reduce their thoughts about electors’ discretion to the printed page. All that they put down about the electors was what we have said: that the States would appoint them, and that they would meet and cast ballots to send to the Capitol. Those sparse instructions took no position on how independent from—or how faithful to—party and popular preferences the electors’ votes should be. (12-13)

This is a great passage. For one thing, it refers to an important reason for being suspicious about the intentions and expectations of constitutional framers: they might not all have agreed with those whose views are on the record. For another, there is an allusion, which I personally find delightful, to Hamilton’s rather hubristic suggestion, in the first paragraph of the Federalist No. 1 that the U.S. Constitution would

decide the important question, whether societies of men are really capable or not of establishing good government from reflection and choice, or whether they are forever destined to depend for their political constitutions on accident and force.

Justice Kagan understands, as Hamilton did not (or at least affected not to) that choice and accident are not so easily disentangled, even in constitutional reflection. Most importantly, though, Justice Kagan drives home the point that “thoughts” “not reduce[d] … to the printed page”― or, more precisely, the enacted page ― do not bind. Justice Thomas specifically concurs with the majority on this point, explaining that “the Framers’ expectations aid our interpretive inquiry only to the extent that they provide evidence of the original public meaning of the Constitution. They cannot be used to change that meaning.” For all its reputation of being incorrigibly politically divided the Supreme Court of the United States is unanimous on this.

Justice Kagan goes on to make another argument, which is less straightforwardly originalist. She appeals to what she regards as the settled practice ― and what I have suggested we may regard as the convention ― of electors casting their ballots only to ratify the voters’ choice, rather than to make their own. “From the first”, Justice Kagan points, “States sent them to the Electoral College … to vote for preselected candidates, rather than to use their own judgment. And electors (or at any rate, almost all of them) rapidly settled into that non-discretionary role.” (14)

It is not quite clear how much weight this should carry on a proper originalist interpretation. In a post at Volokh, Keith Whittington suggests (based on an article which Justice Kagan actually cited ― for another point) that

we should think of this tradition of pledged electors as a “constitutional construction” that is consistent with the constitutional text but not required by the constitutional text. … But that by itself does not tell us whether such constructions can be leveraged to empower state legislatures to punish or replace faithless electors or whether this longstanding norm has fixed the meaning of the text in a way that cannot be altered by future changes in our shared practices. How constitutional text and tradition interact is a difficult conceptual problem, and the Court’s opinion highlights that problem without doing very much to explain how it ought to be resolved.

Indeed, I’m not sure that the argument from practice or convention has a great deal of weight for Justice Kagan: she might only be making it to turn the tables on Mr. Chiafalo, who invoked the (quite exceptional, as Justice Kagan shows) example of past “faithless electors” to argue that it proves that the Constitution protected their autonomy.

But Justice Kagan does suggest what I think is a good reason why the argument should have weight in the particular circumstances of this case: the practice, and arguably even the convention, forms part of the context to a constitutional text ― namely, the Twelfth Amendment to the Constitution. This amendment, “grew out of a pair of fiascos” (14) at the elections of 1796 and 1800. Prior to it, electors cast two votes; the candidate who received the most became president, and the next one, vice-president. In 1796 the top two candidates were “bitter rivals” (14) John Adams and Thomas Jefferson. In 1800, Jefferson, his party’s intended presidential candidate, was tied by its intended vice-president, Aaron Burr, as the electors who supported the one all supported the other. To prevent this reoccurring, the presidential and vice-presidential ballots were split. Justice Kagan points out that, in this way, “[t]he Twelfth Amendment embraced” party politics, “both acknowledging and facilitating the Electoral College’s emergence as a mechanism not for deliberation but for party-line voting”. (14)

The issue isn’t quite the same as the one that, as I argue in a recent article about which I blogged here, the Supreme Court of Canada faced in the Senate Reform Reference, 2014 SCC 32, [2014] 1 SCR 704. There, the original public meaning of Part V of the Constitution Act, 1982 had to be established by referring to conventions. In Chiafolo, conventions are not necessary to establish the original meaning of the Twelfth Amendment. But it is arguably fair to say that the Twelfth Amendment implicitly ratifies them, or takes them into account; while it might have been written as it was in the absence of conventions, the fact that is that it was written as it was because the conventions existed. As a result, Justice Kagan’s appeal to practice, or to convention, is, at least, less troubling here than it might have been in the absence of something like the Twelfth Amendment.

All in all, then, her opinion is an interesting demonstration of what good originalism looks like ― and also of the fact that it can be practiced by a judge who is nobody’s idea of a conservative or a libertarian, and with the agreement of her colleagues, including those whose ideological leanings are quite different from hers. Justice Kagan may or may not be correct: at the Originalism Blog, Michael Ramsay argues that she is not. But that does not matter so much to me. As Asher Honickman recently argued in response to another American decision, textualist ― and originalist ― interpretive methods do not promise complete legal certainty, but they are still valuable because (among other things) they narrow the scope of possible disagreements, and do provide more certainty than alternatives. Justice Kagan and her colleagues show us how to keep faith with a constitutional text. We should pay attention.

Immuring Dicey’s Ghost

Introducing a new article on the Senate Reform Reference, constitutional conventions, and originalism ― and some thoughts on publishing heterodox scholarship

The Ottawa Law Review has just published a new paper of mine, “Immuring Dicey’s Ghost: The Senate Reform Reference and Constitutional Conventions“. It’s been many years in the making ― apparently, I started working on this paper in August 2016, a prehistoric time in my own life, to say nothing of the outside world ― and I don’t think I have ever said much about this project here. So let me introduce it ― and let me also say something about its “making of”, in the hope that its complicated, but ultimately successful fate will inspire readers who may be struggling with wayward papers of their own.

Here is the article’s abstract:

Although the metaphor of “constitutional architecture” appeared in some of the Supreme Court of Canada’s previous opinions, it took on a new importance in Reference re Senate Reform, where the Court held that amendments to constitutional architecture had to comply with the requirements of Part V of the Constitution Act, 1982. However, the Court provided very little guidance as to the scope of this entrenched “architecture.” As a result, the metaphor’s meaning and implications have been the subject of considerable scholarly debate. This article contributes to this debate by arguing that “constitutional architecture” incorporates (some) constitutional conventions. It further takes the position that, instead of relying on this confusing metaphor, the Court should have candidly admitted that conventions were central to its decision by acknowledging that the text of the Canadian Constitution cannot be fully understood without reference to conventions.

Part I reviews, first, the Supreme Court’s opinions in which the notion of constitutional “architecture” has been mentioned, focusing first on this concept’s place in the Senate Reform Reference, and second, some of the scholarly commentary that has endeavoured to make sense of it. Part II sets out my own view that constitutional “architecture,” as this concept is used by the Supreme Court, is concerned primarily if not exclusively with constitutional conventions. Part III considers whether it is possible to determine just which conventions the notion of constitutional architecture encompasses, examining the conventions’ importance and their relationship to the constitutional text as possible criteria, and concluding that neither allows precise determinations. Part IV sets out what would have been a less confusing way of addressing the significance of conventions to the questions the Court was facing in the Senate Reform Reference: frankly recognizing that conventions were relevant to the interpretation of the applicable constitutional texts. Part V examines two objections to the incorporation of conventions (via “architecture” or through interpretation) into the realm of constitutional law, arguing that this incorporation is not illegitimate and that it will not stultify the Constitution’s development. Part VI concludes with an appeal for greater transparency on the part of the Supreme Court.

Actually, the article’s core idea ― that the architecture to which the Senate Reform Reference refers incorporates constitutional conventions ― was part of my initial reaction to the Supreme Court’s opinion. And of course it only develops the suggestions made by Fabien Gélinas and me in a paper we wrote before the Senate Reform Reference was argued. It is also of a piece with my other work on conventions, which argues against the theoretical validity of a sharp distinction between the conventions and the law of the constitution.

The other thing the article does, though, is a new departure. When Professor Gélinas and I wrote about the role of conventions in the then-upcoming Senate Reform Reference, we accepted that the constitution is a “living tree”, and indeed made it the basis of our argument that constitutional interpretation must incorporate conventions. But of course I no longer think that living constitutionalism is the correct approach. So the article begins the project of making sense of the reality that a very significant part of the Canadian constitution is “unwritten”, or rather extra-textual, uncodified, from an originalist perspective.

The argument, as it happens, does not change: as I explain, an originalist must also read the constitutional text in light of conventions which were ― in originalist terms ― part of the publicly available context at the time of the text’s framing. Still, it was important for me to set out this argument from an originalist, as well as a written constitutionalist perspective. It was also important to give the reader a glimpse of how this originalist argument works. To this end, the article wades into historical evidence, looking at the Confederation debates to argue that the conventions relative to the functioning of the Senate were anticipated by the framers of the Constitution Act, 1867 (in addition to being well known to those of the Constitution Act, 1982). Future work ― mine and perhaps that of others ― can build on this foundation, and on Ryan Alford’s recent book Seven Absolute Rights: Recovering the Historical Foundations of Canada’s Rule of Law, to fully integrate not only conventions (and therefore “architecture”) but also underlying principles and structural arguments into a comprehensive originalist conception of the Canadian constitution.

This brings me to the “making of” part of the post. As you might imagine, getting the originalist arguments through peer review was not an entirely straightforward proposition. I deliberately diluted them, presenting them only as alternative to the living constitutionalist approach, to which I gave equal attention and which I refrained from criticizing.

Still, at first, this was not enough. The reviewers selected by the first journal to which I submitted the paper were quite skeptical of the whole project, and the attention it devoted to history and to originalism contributed to that skepticism. I was asked to revise and resubmit in light of the reviewers’ comments, and did so, although I could not make the sorts of changes that would have assuaged their concerns without changing the nature of the whole piece. The editor referred the revised article to the same reviewers, who understandably were unimpressed with my revisions, and the article was rejected. Frankly, the revision and resubmission was a waste of my time, as well as of the reviewers’. Their initial objections were so fundamental that there was no real chance of their accepting any revisions I might plausibly have made.

So, after sulking a bit, I submitted the paper elsewhere ― namely, to the Ottawa Law Review. The reviewers there were more open-minded, though one remarked on the oddity, as he or she thought, of granting so much airtime to originalism, and suggested cutting that part of the paper. But the article was accepted, and so revisions were more at my discretion than they would have been in a revise-and-resubmit process. To me, of course, the discussion of originalism was very much part of the point of the paper, so I insisted on keeping it. (I have to say that, while many scholars will of course disagree with originalism as a normative matter, I find it hard to understand how one still can argue that it simply isn’t relevant to Canadian constitutional law; and least of all, how one can make such an argument in a discussion of the Senate Reform Reference, which very much relies on arguments about the intentions of the framers of the Constitution Act, 1982.)

To my mind, there are a few lessons here. One is that if you have an unorthodox agenda, it might be useful to go slowly, and plan to make several steps before getting to your ultimate destination. If you present your idea, not as certain truth right away, but as a possibility to be entertained, you make the pill easier to swallow while still moving the argument from being, as American scholars put it, “off the wall” to “on the wall”. I’m not sure, of course, but I think that this cautious approach helped me here.

The second lesson is that the peer review process is a bit of a crapshoot. Even if you are cautious, some reviewers will bristle and see their role as that of gatekeepers preserving scholarship from heresy. But others may see their role differently, and say that, while they disagree with the paper, it is still well argued and deserves a hearing. (Of course, you have to make their life easier and make sure that the paper is indeed well argued; the more heterodox you are, the more you need to dot your i’s and cross your t’s.) To be sure, there are limits to such tolerance: at some point, heterodoxy veers into kookiness, and even an open-minded reviewer should say so. And, of course, where heterodoxy ends, and kookiness begins is not a question that admits of easy answers. Perhaps to the original reviewers who rejected my piece I was a kook.

But this brings me to the third lesson. If at first you don’t succeed, try again. Try with a different journal, hope you get different reviewers, perhaps a more sympathetic editor. That’s easier to do when your paper is one that doesn’t need to be out right away ― I’ve given up on a comment on R v Comeau, in part because a case comment loses its relevance after a while ― whereas this article, making a less topical and more fundamental claim, could wait. And perhaps there is a further lesson here, which is that it is better to reserve heterodox ideas for articles of this sort, knowing that it might be a while before they can run the peer review gamut. But, be that as it may, the point is that, precisely because it is a crapshoot, precisely because it empowers people who enjoy being more Catholic than the Pope, the peer review process can be dispiriting ― but knowing why it is this way should remind us that it isn’t always this way.

Good luck with your heterodox articles ― and please read mine, and let me know what you think!

Immuring Dicey’s Ghost

The Senate Reform Reference and constitutional conventions

In its opinion in Reference re Senate Reform, 2014 SCC 32, [2014] 1 SCR 704, the Supreme Court notoriously relied on a metaphor that had previously popped up, but played no real role, in its jurisprudence: “constitutional architecture”. Notably, the court was of the view that moves towards an effectively elected Senate would modify the constitution’s architecture, and such modifications required formal amendment under Part V of the Constitution Act, 1982, just as much as changes to the explicit provisions of the constitution’s text. Yet the court’s explanations of just what this architecture was were short and cryptic, and haven’t been elaborated upon ― judicially ― in the intervening years.

To fill in this void, an academic cottage industry sprang up to speculate about the meaning of the architectural metaphor and about what other constitutional reforms it might block. For example, Kate Glover Berger suggested that “action taken to dismantle or undermine the administrative state could be deemed unconstitutional” because the administrative state is built into the architecture of the Canadian constitution. Lorne Neudorf invoked architecture in the service of an argument to the effect that courts can read down or indeed invalidate vague delegations of legislative power to the executive branch. Michael Pal speculated that the first-past-the-post electoral system might be entrenched as part of the constitutional architecture.

All this while, I have been working on my own contribution to this genre, called “Immuring Dicey’s Ghost: The Senate Reform Reference and Constitutional Conventions”, which is finally going to be published by the Ottawa Law Review later this year. In a nutshell, I argue that “architecture” is really just code for “conventions” ― those supposedly non-legal but fundamentally important constitutional rules, arising out of political practice and morality, which courts have long said they could not possibly enforce. And I argue, further, that the Supreme Court should have squarely addressed the fact that it was relying on conventions, instead of playing confusing rhetorical games.

A draft is now available, for your reading pleasure. Here is the abstract:

Although the metaphor of “constitutional architecture” had appeared in some previous opinions of the Supreme Court of Canada, it took on a new importance in Reference re Senate Reform, where the Court held that amendments to constitutional architecture had to comply with the requirements of Part V of the Constitution Act, 1982. However, the Court provided very little guidance as to the scope of this entrenched “architecture”. As a result, the metaphor’s meaning and implications have been the subject of considerable scholarly debate.

This article contributes to this debate by arguing that “constitutional architecture” incorporates (some) constitutional conventions. It further takes the position that, instead of relying on this confusing metaphor, the Court should have candidly admitted that conventions were central to its decision by acknowledging that the text of the Canadian constitution cannot be fully understood without reference to conventions.

Part I reviews, first, the Supreme Court’s opinions in which the notion of constitutional “architecture” has been mentioned, focusing on this concept’s place in the Senate Reform Reference, and then some of the scholarly commentary that has endeavoured to make sense of it. Part II sets out my own view that constitutional “architecture”, as this concept is used by the Supreme Court, is concerned primarily if not exclusively with constitutional conventions. Part III considers whether it is possible to determine precisely which conventions are encompassed by the notion of constitutional architecture, examining the conventions’ importance, and their relationship to the constitutional text as possible criteria, and concluding that neither allows precise determinations. Part IV sets out what would in my view have been a less confusing way of addressing the significance of conventions to the questions the Court was facing in the Senate Reform Reference: frankly recognizing that conventions were relevant to the interpretation of the applicable constitutional texts. Part V examines two objections to the incorporation of conventions (via “architecture” or through interpretation) into the realm of constitutional law, arguing that this incorporation is not illegitimate, and that it will not stultify the constitution’s development. Part VI concludes with an appeal for greater transparency on the part of the Supreme Court.

The last thing I mention here is that this paper begins the project of bringing together two subjects on which I had mostly been writing separately: constitutional conventions on the one hand, and originalism on the other. As explained here, Canadian originalism has to grapple with the fact that some of our most important constitutional rules are unwritten. This paper, although it doesn’t make a case for originalism, begins to outline what that an originalist approach to conventions will look like.

Heresy!

The UK Supreme Court’s decision in “the Case of Prorogations” and the political constitution

I wrote last week about the UK Supreme Court’s decision in R (Miller) v Prime Minister, [2019] UKSC 41 (Miller (No 2)), which unanimously held that the Prime Minister’s advice that the Queen prorogue Parliament for five weeks was unlawful, and the prorogation itself, therefore invalid. There was, however, one aspect of Miller (No 2) that I did not discuss in any detail: that of the Court’s treatment of the “political constitution”, and the distinction between those constitutional rules that are part of constitutional law and those that are not. In this post, I want to come back to this issue.

It is useful to begin with the orthodox view of the political constitution, articulated by scholars such as A.V. Dicey and courts in cases like Re: Resolution to amend the Constitution, [1981] 1 SCR 753 (Patriation Reference) and R (Miller) v Secretary of State for Exiting the European Union, [2017] UKSC 5 (Miller (No. 1)). On the orthodox view, only some constitutional rules ― statutes and common law rules, such as those circumscribing the scope of the royal prerogative and, in part, of Parliamentary privilege ― are part of constitutional law. Other rules, known as constitutional conventions, are not constitutional law and the courts will not enforce them, although they can sometimes take note of them in resolving properly legal issues.

In the Patriation Reference, the Supreme Court of Canada suggests a number of reasons for distinguishing convention and law. First, the majority opinion on the conventional question insists that “unlike common law rules, conventions are not judge-made rules. They are not based on judicial precedents but on precedents established by the institutions of government themselves.” (880) The majority opinion on the legal question makes the same point, and adds that “[t]he very nature of a convention, as political in inception and as depending on a consistent course of political recognition … is inconsistent with its legal enforcement”. (774-75) In Miller (No 1) the majority of the UK Supreme Court put it more pithily: “[j]udges”, it said, “are neither the parents nor the guardians of political conventions”. [146]

Second, and relatedly, the Patriation Reference suggests that it would be inappropriate to enforce conventions, given their questionable pedigree. “What is desirable as a political limitation ”, it says, “does not translate into a legal limitation, without expression in imperative constitutional text or statute”. (784) Third, the majority opinion on the conventional question argues that the courts lack remedies to compel compliance with conventions. Fourth and last, the same opinion notes that conventions conflict with legal rules, and courts are bound to apply the latter. Others have also argued that conventions are too shrouded in uncertainty―that both their very existence and their implications for specific situations are too doubtful―for them to function as meaningful legal rules.

Miller (No 2) doesn’t explicitly engage with any of this. But by the time the UK Lady Hale and Lord Reed are done with the case, not much of the old orthodoxy is left standing. They not only regularly advert to conventions (which courts can do on the orthodox view), but seem to assimilate the exercise of conventional and legal powers, and arguably provide a way for judicial enforcement of conventions, in disregard of the traditional distinction between conventions and law. This might be a good thing, but I am uneasy at the way it is accomplished.

The tone is set early on. At the beginning of the judgment, Lady Hale and Lord Reed explain what a prorogation is, and contrast it with a dissolution of Parliament. Following the latter, they note, “[t]he Government remains in office but there are conventional constraints on what it can do during that period”. [4] There is no particular need to mention these “conventional restraints”, even for the sake of the descriptive point the Court is making (which is itself unnecessary, although perhaps helpful, to explaining the decision in the case at bar). A more orthodox court would probably have avoided mentioning conventions here. Not this one.

More relevantly to the case, Lady Hale and Lord Reed say that they “know that in approving the prorogation, Her Majesty was acting on the advice of the Prime Minister”. [15] They go on to further explain that

the power to order the prorogation of Parliament is … exercised by the Crown, in this instance by the sovereign in person, acting on advice, in accordance with modern constitutional practice. It is not suggested in these appeals that Her Majesty was other than obliged by constitutional convention to accept that advice. [30]

The double negative allows the judgment to ostensibly “express no view on” [30] whether Her Majesty was indeed “obliged by constitutional convention” to accept the Prime Minister’s advice, but the fig leaf is quickly blown away. The Court proceeds to assess the lawfulness of the Prime Minister’s advice, which makes little sense unless one assumes that Her Majesty had to follow it. If the advice was in reality what it is in name, why would the Court be looking into it? This is further confirmed by the Court’s approach to the remedy. The applicants’ lawyers, implicitly adopting a more orthodox position, only sought “a declaration that the advice given to Her Majesty was unlawful”. [62] But the Court goes further, and says that this advice “led to the Order in Council [pursuant to which the prorogation was carried out] which, being founded on unlawful advice, was likewise unlawful, null and of no effect and should be quashed.” [69] Led to? Founded on? This, I am tempted to say, is an imitation fig leaf, not even the real thing. What Lady Hale and Lord Reed mean ― the only way they reasoning makes sense ― is that the advice required the order in council to be made; that it was legally determinative, not just factually causative.

Consider: If I write a letter to Boris Johnson with a devious master plan for executing no-deal Brexit, and he follows it to the letter, my letter, which will actually be advice, in the sense of a suggestion, will not be the subject of court proceedings. The relevant choices will still be the Prime Minister’s, and, should their legality be called into question, my intervention will be no more than a part of the factual background, if that, even though it would be fair to describe it as “leading to” the Prime Minister’s actions, which would be “founded on” it. Of course, my position vis-à-vis the Prime Minister is different, in a constitutionally significant way, from the Prime Minister’s vis-à-vis the Queen. But, on the orthodox view, this would significant as a matter only of political, not legal, constitutionalism. The Supreme Court sees things differently. To repeat, Lady Hale and Lord Reed’s approach only makes sense if the Prime Minister’s advice is binding on Her Majesty, so that there is no daylight between his ostensibly conventional role and the exercise of the Crown’s legal powers.

Perhaps one might argue that the UK Supreme Court’s treatment of conventions is orthodox because it is only a necessary step towards resolving a properly legal question as to the scope of the prerogative power of prorogation. The Court, on this view, does not do what the Diceyan dogma tells us is impossible: enforce a convention. But is that so? And if it is so in this case, what about others in which the Court’s reasoning might be applied? (As discussed last week, the Court claims that Miller (No 2) is a “one off”. That remains to be seen.)

It is crucial, I think, to Lady Hale and Lord Reed’s reasoning that they are able to confidently assert that, while “Parliament does not remain permanently in session … [i]n modern practice, Parliament is normally prorogued for only a short time”. [45] They rely, moreover, on a statement by a former Prime Minister to the effect that nothing more is necessary. And they conclude that constitutional principles (specifically, Parliamentary sovereignty and executive accountability) mean departures from modern practice would require justification. Without explicitly undertaking an analysis in terms of the Jennings test adopted by the Supreme Court of Canada in the Patriation Reference, Lady Hale and Lord Reed come close to showing that the relevant constitutional actors seem to be following a rule, that they feel bound by the rule (or at least that they have no reason not to follow it), and that there are reasons, in the shape of important constitutional principles, for this rule ― in other words, that, according to the Jennings test, there exists a convention. Only, in effect, Miller (No. 2) very nearly transmutes this “modern practice” into law. (Very nearly, because in principle it is still open to a Prime Minister to justify departure from the practice.)

And beyond what has or has not happened in this particular case, I think the reasoning deployed by Lady Hale and Lord Reed can serve as a blueprint for judicial enforcement of conventions in the future. In a nutshell, what Miller (No 2) says is that the exercise of the royal prerogative is subject to implicit limits imposed by constitutional principles, and that the location of these limits ― which can be inferred, in part at least, from “modern practice” ― is a justiciable question. So consider, for example, the convention requiring the sovereign to assent to legislation passed by the House of Commons and the House of Lords (or only the House of Commons legislating pursuant to the Parliament Act 1949). Courts couldn’t enforce that convention and either require the sovereign to assent or treat a bill passed by the Commons and the Lords as law without her assent, right? Well, they could say that the power to withhold assent is subject to implicit limits imposed by the democratic principle, such that any departures from the modern ― or, this case, centuries-old practice of not withholding assent ― must be justified, and… voilà!

As readers may know, I am a longtime skeptic of the Diceyan orthodoxy on the separation between conventions and law. I think that the courts should have regarded conventions as common law rules en devenir and enforced them if and when necessary, subject however to justiciability concerns ― for example when the conventional rule is vague and/or its application in a given case involves political judgment. So the outcome of Miller (No 2) is not all bad, from that perspective.

And its reasoning makes the arguments invoked in support of the orthodoxy that much more difficult to sustain. The emphasis that Lady Hale and Lord Reed put on the development of the common law constitutional rules in cases such as the Case of Proclamations shows that the disclaimers of the creative role of the judiciary and protestations about its inability to translate “what is desirable as a political limitation” into legal rule always proved too much. Similarly, their confident treatment of the question of the remedy and of the evidentiary issues shows that concerns about the courts’ ability to engage with conventional issues have been greatly exaggerated.

That said, I have my reservations about the approach the Miller (No 2) court takes. For one thing, I wish Lady Hale and Lord Reed had been more transparent about what they were doing. Miller (No 1), where the UK Supreme Court reiterated the orthodox view that a convention could not be judicially enforced ― even a convention enshrined in statute! ― was only decided a couple of years ago. Miller (No 2) is almost a complete U-turn from its namesake, yet we have little explanation about why the ladies and lords were for turning. Here as on other issues, the suspicion of results-oriented reasoning must weigh heavily on the Court. More substantively, Lady Hale and Lord Reed may be overconfident in the courts’ ability to dispose of the factual questions that may arise when the courts enter the realm of politics. As noted above, I think that these questions will sometimes ― though by no means always ― be difficult enough that non-justiciability is a real concern. The reasoning in Miller (No 2) does not acknowledge this, and in my view this is a mistake.

Miller (No 2) thus seems to be a very significant, albeit unacknowledged, development in the UK Supreme Court’s understanding of the nature of the constitution, and specifically of what used to be thought of as its political, non-legal component. Without saying so, the Court is, perhaps, in the process of correcting the mistake made by scholars and judges who saw a sharp separation between law and politics when, at the heart of the UK’s constitution, none existed. Views on the nature and status of conventions that were just recently said to be quite heretical now appear to have prevailed.

If anything, the pendulum may have swung too far in the other direction. The Court hasn’t thought through the implications of its reasoning. Perhaps this is just how the common law develops: case by case, without the courts fully understanding the consequences of one decision for those that will follow. In that sense, Miller (No. 2) might not be an innovation at all. The system works, perhaps, but it is not always a pretty sight.