When Judicial Disagreement Doesn’t Matter

What does it mean for an appellate court’s decision to be unanimous?

By Peter McCormick

To begin with the obvious: common law appellate court decisions include two major elements.  The first is the outcome – crudely, who won? – that takes the dichotomous form of“appeal allowed” or “appeal denied”.  The second is the reasons for judgment, that take the form of a legal essay, sometimes of extended length, explaining why that outcome is the appropriate one and how it is grounded in existing law.  Put more formally: an appeal court decision responds to two imperatives, the first being to provide a definitive resolution to a specific legal dispute, and the second being to provide the reasons for that outcome in such a way as to provide useful guidance to lower courts and future litigants.

From this it follows that there are two different types of judicial disagreement. A member of the panel may disagree with the outcome, saying that their colleagues got the winner wrong – we call this type of disagreement a“dissent”.  Or they can agree with the outcome but disagree, or at least not completely agree, with the reasons given to explain and justify that outcome – in Canadian usage we call this a “separate concurrence”.

Dissents have drawn a voluminous literature, both empirical and normative, to such an extent that even a preliminary list would take several pages. When the focus of discussion shifts from the consideration of a specific example to a more general level, the usual complaint is that dissent rates are too high – judges (or perhaps just some judges) are too ready to pursue their own vision of the law rather than contributing to and reinforcing a more solid institutional position.  Separate concurrences, on the other hand, are very much the forgotten poor cousin of judicial disagreement; to the best of my knowledge, there have only ever been two articles in Canadian law journals exploring the practice of separate concurrence and evaluating its contribution to the law.[1]

With respect to dissent, Jeremy Gans, in a recent piece in Inside Story and referring specifically to the High Court of Australia, has taken the highly intriguing position of flipping the “too many dissents”argument.   Quite the contrary, he complains that it is possible – and, for the current High Court, an actual achievement – to have dissent rates that are low to the point of dysfunction, so much so that it reflects badly on the Court’s performance. His “Great Assenters” title is deliberately and pointedly ironic; at a certain point, he does not think that “assent” is great at all.

This looks like a fascinating conversation that I would love to join – perhaps by suggesting a “proper” (or at least“normal”) level of judicial dissent that as a yardstick against which “too high” and “too low” can be more precisely measured, such that the reasons (commendable or otherwise) for departures from that norm can be identified.  But my enthusiasm was derailed by the second paragraph, which casually told me “All four decisions made in the High Court of Australia last month were approved by every judge who sat (even if they sometimes disagreed on the reasons).”  Our own Supreme Court has exactly the same attitude toward “disagreement on the reasons”, keeping its statistics on how many of its judgments were “unanimous as to outcome” but not seeing any necessity of taking the further step of telling us how many of those were also “univocal” (which is to say: unanimous as to reasons as well).

The clear implication of both Gans’s comment and the Supreme Court statistical reporting is that only disagreement as to the outcome really matters; differences as to the reasons are not really worth noticing – not even if they involve fundamental differences expressed at considerable length, not even if they are joined by several other judges, not even if the consequence is that there is no statement of “outcome plus reasons”statement that is supported by a majority.  To be sure, disagreeing about the outcome is much more dramatic, with greater potential for news headlines and editorial commentary aiming scathing criticism at either the majority or the minority.  It conjures visions – sometimes rebuttable but often compelling – of innocent people sent to prison or guilty people freed, of honest people victimized without remedy, of perfectly valid laws rendered null and void or bad laws upheld.  Separate concurrences are less dramatic and often harder to explain, a judicial equivalent of “insiders’ baseball.”

With all due respect to both Prof. Gans andthe Supreme Court of Canada, I think their focus on “unanimous as to outcome”is a profound mistake.  Putting the pointas starkly as possible: the outcome really matters only to the immediate parties, but the reasons matter to everybody. This is because it is the reasons, not the outcome, that constitute the precedent that constrains the immediate court and instructs the lower courts.   Since there are only two possible outcomes (allow or dismiss), how can they carry any precedential message at all?  The real point about dissent is not that the judges disagreed on the outcome but that they disagreed about the content and meaning and application of the relevant law; generally speaking, to disagree with the outcome is ipso facto to disagree with the reasoning that led to the outcome, so it is easy to conflate the two.[2] But “disagreeing on the content, meaning and application of the law” is precisely what separate concurrences are aboutas well, in ways that may be less dramatic but are often as profound and as potentially impactful as many dissents.  As Scalia once said, a judgment that gets the reasons wrong gets everything wrong that it is the function of an appeal court decision to provide;[3]it follows that minority reasons identifying that species of error are just as functional, and just as important, as minority reasons that challenge the outcome as well.

To step back for a moment: there are essentially three different kinds of separate concurrence. The first is what we might call the “just one more thing” concurrence,which expresses agreement with the majority but wants to add one additional related thought about the law that the writer could not persuade their colleagues in the majority to sign on to. The second is what we might call the “one less thing” concurrence, which expresses general agreement with the majority but specifically excludes one or more elements of the majority reasons; depending how significant those elements are, and how many other judges sign on to it, this can sometimes have real implications.  But the third kind, and as it turns out (at least in Canadian practice) the most common of the three, is the “by another route” concurrence, which opens with some variant of the apparently innocuous statement “I reach the same conclusion, but for different reasons.”  This is not innocuous at all;it is as serious as judicial disagreement gets, so much so that McLachlin J. (as she then was) once described herself as “respectfully dissenting” from the majority even though she was at the time agreeing that the immediate appeal should be dismissed (in R v Potvin [1993] 2 SCR 880).  Although she seems to have repented from this terminology, I remain convinced that she was on to something.

The distinctions I am making are highlighted by two important developments on the Supreme Court of Canada.  The first is a consistent practice dating back several decades that distinguishes between unanimous (or majority, or plurality) judgments and minority(dissenting or separately concurring) reasons;this replaced the earlier practice whereby any set of reasons delivered by a judge was referred to as a “judgment”.  The term “dissenting judgment” has become an oxymoron when applied to the current Court, although it was used by the Court itself before the late 1960s and still is appropriate for jurisdictions (such as the Ontario Court of Appeal) where the parallel labeling practice has not been adopted.  The second is a decision-delivery process that highlights the judgment (or at least the initial attempt at a judgment) by systematically framing other sets of reasons as responses (“I have read the reasons”).[4]  The joint impact of these two developments is to flag the significance of non-dissent disagreement in a very transparent way,although it is only making more visible implications that apply even in the absence of such explicit signals.

When and why and how does this matter?  To simplify the context, let us take the most dramatic position and assume a nine-judge panel that has divided 5-4 on the outcome and then 4-1 on the reasons.  We have an outcome, but what do we do about the reasons for judgment?   Is there a plurality judgment, and if so which set of reasons earns the label?  Or is there no “judgment” at all?

That depends on the nature of the disagreement between the various fragments of the majority.  If the solo judge is writing reasons of the“one more thing” variety, then we have a separate concurrence that has explicitly lined itself up with and behind the four-judge reasons in such a way as to make those reasons the judgment.  If those reasons are of the “one less thing” variety, then it may well have displaced the other reasons to become the judgment itself (because the “rule” as to which fragment of a divided majority is the judgment is not “largest fragment” but“narrowest legal grounds” – for an example, see Chaoulli v Quebec (Attorney General), 2005 SCC 35, [2005] 1 SCR 791).  But if the disagreement is of the “by another route” variety, then there may well be no “judgment” at all, which is to say that there is no majority position on the law that explains the outcome in way that clearly establishes precedent.  For the Supreme Court of Canada, this only happens about once a year, but the point is that it does happen – the most recent examples are Haaretz.com v Goldhar, 2018 SCC 28 and Centrale des Syndicats du Québec v Quebec (Attorney General), 2018 SCC 18. Or consider the even more recent case of Mikisew Cree First Nation v Canada, 2018 SCC 40, which was unanimous as to outcome but with four different sets of reasons, all of comparable length but none attracting more than three signatures on a nine-judge panel; the decision is either unanimous, or 7-2, or 5-4, depending which of the major issues attracts your attention.  Gans’s “great assenters” label hardly seems appropriate.

But my concerns apply more broadly than these dramatic and unusual developments.  More generally we might say that behind every dissent, especially one that draws multiple signatures, lurks a disagreement deep enough that it might one day grow into a dramatic explicit abandonment of the majority’s jurisprudential position – like the reversal of the 1987 Labor Trilogy (Reference re Public Service Employee Relations Act (Alberta), [1987] 1 SCR 313PSAC v Canada [1987] 1 SCR 424RWDSU v Saskatchewan [1987] 1 SCR 460) twenty years later in B.C. Health Services (Health Services and Support-Facilities Subsector Bargaining Association v British Columbia, 2007 SCC 27, [2007] 2 SCR 391) on the status of collective bargaining under the Charter’s guarantee of freedom of association, or Carter’s 2015 repudiation (Carter v Canada (Attorney General) 2015 SCC 5, [2015] 1 SCR 331) of the 1993 Rodriguez decision (Rodriguez v British Columbia (Attorney General) [1993] 3 SCR 519) on the constitutional status of the ban on assisted suicide.  By a similar logic, behind every separate concurrence (but especially those of the “different route” variety, and especially those joined by several other judges) lies the potential for a less visibly dramatic but comparably significant evolution, perhaps to the extent of having the separate concurrence gradually displace the one-time majority incitation frequency (which again is something which has happened more than once).

Differences over reasons matter because reasons are what judicial decisions are all about; the evolution of the reasons explaining outcomes is what brings about much of the incremental change in the law.  This makes it a serious mistake to assume that some judicial disagreement is necessarily less important simply because it does not involve disagreeing on the outcome as well as on the reasons that justify that outcome.   When is it that judicial disagreement doesn’t matter?  Only when we are so shortsighted as to ignore it.

[1] I admit that I wrote both of them: see Peter McCormick, “The Choral Court: Separate Concurrences on the McLachlin Court 2000-2004Ottawa Law Review, Vol. 37 (2005-6); and Peter McCormick, “Standing Apart: Separate Concurrence and the Supreme Court of Canada 1984-2006McGill Law Journal Vol. 53 (2008).

[2] That said, I think it is not impossible for judges to disagree on the outcome without disagreeing on the central legal issues and their precedential implications; my leading candidates would be R v Therens, [1985] 1 SCR 613 and Andrews v Law Society of British Columbia [1989] 1 SCR 143, the point being that for both of them the major precedential finding on the constitutional issue is made and explained in the dissenting reasons.

[3] Antonin Scalia, “The Dissenting Opinion” 1994 Journal of Supreme Court History p.33

[4] Peter McCormick, “Structures of Judgment: How the Modern SupremeCourt of Canada Organizes Its Reasons” Dalhousie Law Journal, Vol. 32 (2009)

The Economics of Unanimity

It is often thought that judicial unanimity is a valuable commodity. Chief Justices bang heads, twist arms, and break legs in order to get their courts to produce more of it, but they don’t always succeed, and unanimity remains at least somewhat scarce on the U.S. and Canadian Supreme Courts (although more on the former than on the latter, which has been unanimous in judgment in between two thirds and three quarters of its decisions rendered since 2010). The unusually high output of the unanimity production line at the US Supreme Court this year has produced much commentary. But how much do we really know about the economics of unanimity? What is it worth? More precisely, what is its purchasing power? How much does it cost? And is the cost worth what you get in return?

In the New York Times, Adam Liptak reviews some academic attempts to answer these questions (in the American context), including a recent paper by Cass Sunstein. The takeaway from this literature seems to be that unanimity is worth less than is commonly assumed. Mr. Liptak notes that people, including judges, often think that “[t]he public may be less likely to accept and follow decisions that would have gone the other way with the switch of a single vote.” Yet experiments ― and perhaps even historical experience ― do not bear out this intuition. And while another claim about the value of unanimity, that unanimous judgments are less likely to be reversed, is apparently supported by the facts, the number of overturned decisions is so small to begin with that this value is more illusory than real. Finally, although unanimous judgments might in theory make for a clearer legal landscape, they often fail to deliver on this promise too. Mr. Liptak points out that

Supreme Court opinions are the product of negotiation and compromise, which is why they can read as if written by a committee. A nine-member committee does not seem likely to produce crisper prose than a five-member one.

At the Volokh Conspiracy, Jonathan Adler chimes in, writing that

[t]he cost of broad agreement may be an opinion that speaks in generalities and pushes aside the potential points of disagreement.  Concurrences and dissents often draw clearer lines and are more analytically coherent than majority opinions. The sorts of opinions that result from efforts to achieve greater unanimity are different from those that merely seek the median vote.

At the same time, coalescing around a narrow holding allows the Court to avoid premature resolution of a potentially divisive question, perhaps leaving it to be resolved when it can be resolved in a unanimous way or even putting it off indefinitely.  This is itself a virtue of judicial minimalism, according to some.

(Paragraph break removed)

These are useful observations, so far as they go, but I think some additional clarifications are necessary for us better to assess the value (or lack thereof) of unanimity.

For one thing, we need to be clearer about what it is that we are talking about. Unanimity in judgment does not necessarily mean unanimity in reasoning, and indeed in some of the recent decisions of the U.S. Supreme Court (for example in NLRB v. Noel Canning, a case considering the constitutionality of President Obama’s “recess appointments”) unanimous judgment masks sharp disagreements about the law between a majority and a concurrence. In such cases, it seems unreasonable to expect the putative effects of unanimity, whether positive or negative, to manifest themselves.

For another, even unanimity in opinion can be of different sorts. While some unanimous decisions will indeed be the products of laboured compromise, and thus be likely to exhibit the flaws described by Mr. Liptak and prof. Adler, others are in fact the products of genuine agreement about the legal principles involved and their application. Probably most decisions of intermediate appellate courts (which have unanimity rates much higher than Supreme Courts, both in the U.S. and in Canada) are of this sort, because they are rendered in “easy” cases where the law is relatively clear. Some decisions of Supreme Courts, at least, are of this sort too. I don’t know American law well enough to give examples, but they are plenty this side of the border ― among the more notable recent cases, Reference re Securities Act, 2011 SCC 66, [2011] 3 SCR 837, comes to mind. Of course, it might not be easy for external observers to distinguish unanimity of agreement from the unanimity of compromise (and a single decision might involve both), but it seems likely that the former sort is more valuable, at least for clarifying the law (though not if one values unanimity for requiring narrow rulings!) ― but also less susceptible of deliberate manufacture by a court.

As for the value of unanimity as a means of exchange for acquiring legitimacy, I wonder whether an inquiry into the value “ordinary” people attach to it is the relevant one. The issue here does not concern unanimity alone. Rather, given well-documented and pervasive political ignorance, I wonder how much people outside the legal and political communities notice and care about judicial decisions at all, and to the extent that they do, how much their views of these decisions are influenced by what politicians (and perhaps experts) tell them. It is possible, and indeed likely, that the perceived legitimacy of the vast majority of, and perhaps of all, judicial decisions depends on the opinions of a certain class of journalists, lawyers, and politicians. If that is so, then an empirical assessment of the value of unanimity should look at the views of such people, and not of random citizens.

Finally, we might need a fuller picture of the transaction costs involved in achieving unanimity. Presumably, producing a decision that is unanimous in reasoning ― at least when unanimity of of compromise rather than of agreement is involved ― takes time and effort, which might, in theory, otherwise be expended on producing better decisions in other cases. It also, by definition, requires individual judges to sacrifice the opportunity to implement or even express their views about the law, and prevents disagreement from being aired in the open. It seems at least plausible that this will, for lack of a better term, undermine the morale of the court or at least of the more independent (or headstrong) judges. I don’t know, I’m afraid, whether this is a real problem (perhaps readers who have clerked at the Supreme Court can tell!). Judges surely know that they sometimes need to “take one for the team”, though nobody, I imagine, like to have to do that very often. In any case, the possibility is worth considering.

All that to say that, as with other commodities, unanimity doesn’t have any “true” value. How much it costs and what it can buy depends on a number of contextual factors. A quest, or demands, for unanimity that ignore these factors will likely be misguided, and perhaps pernicious.