Ian Greene and Carl Baar

A paper prepared for presentation to the Annual Meeting of the Canadian Political Science Association, Quebec City, May 27, 2001

This paper addresses one of the key findings of the 1990s Canadian Appeal Courts Project: that appellate judges in Canada commonly divide the decision-making process into two distinct although related steps. The first step is to decide the outcome of the appeal. The second step is to decide how best to justify or defend the decision. This second step is particularly important with regard to reserved decisions.

Although it may appear at first glance common sense that judges decide in a two-step process, a better understanding of the how the judicial two-step operates in practice helps to shed light on the unique contribution of the judiciary to the policy-making process of democratic government. From this perspective, it is surprising that there has been very little analysis of the two-step process of appellate decision-making in the literature on judicial behaviour. We argue that the policy questions addressed by appellate judges in the decision-making process are typically different from those addressed by legislators and policy analysts in government. Therefore, judges perform an essential function in the policy process that the legislative and executive branches are unlikely to duplicate. That function is to ensure that policies that are the subject of litigation meet basic standards of individual fairness and justice.

We begin by describing the source of our data: the Canadian Appeal Courts Project. We then analyze the two-step decision-making process of appellate judges, distinguishing it from the decision-making process of trial court judges. In the last section of the paper, we discuss the implications of the appellate judicial two-step for the debate about the proper extent of the policy-making role of judges in a democratic political system.

The Canadian Appeal Courts Project was conceived first as a descriptive enterprise to fill a gap in the literature on judicial behaviour. Political scientists had already looked at the work and voting patterns of justices of the Supreme Court of Canada, and criminologists had looked at the work of criminal courts at the trial level. No one had done an empirical examination of the work of the courts of appeal that resolve legal issues in every province -- in most cases without further review by the Supreme Court of Canada. The design of the CACP also allowed us to address appellate courts in a way rare in existing literature. Rather than seeing them only as articulating legal rules, or reflecting the voting patterns of individual justices, we saw courts as composed of organized processes conducted and directed by autonomous professionals. Thus we combined two sources of data: personal interviews (with 75% to 80% of all the appellate judges between 1991 and 1995 -- 101 out of 125 or so) and caseflow data from 6,000 cases to capture the flow of appellate litigation. As a result, we were able to capture the perspectives of the judges while comparing their views to the realities of caseflow. For example, although we felt that the judges sincerely strove for impartial judgment, the reality of the process was that the crown consistently fared better in criminal appeals -- winning most of its appeals while the accused lost most of theirs. The project was funded by a grant obtained from SSHRC in 1990, and it culminated with the publication of the book, Final Appeal: Decision-making in Canadian Courts of Appeal, in 1998.

The project was conducted by a research team composed of the two of us and Peter McCormick, George Szablowski and Martin Thomas. Ian Greene conducted fifty-six of the interviews, including interviews with eight Supreme Court of Canada judges, and appellate judges in Newfoundland, Nova Scotia, New Brunswick, Prince Edward Island, Ontario, Manitoba, Alberta, and the Federal Court of Appeal. Carl Baar conducted interviews in British Columbia and Ontario. Peter McCormick’s interviews were in Alberta and Saskatchewan, and George Szablowski conducted all of the interviews in Quebec.

The team approach meant that we had the opportunity to meet from time to time as a group to consider the results of our research. At one of these meetings, comparing interview notes made clear that one of the characteristics of the judicial process that was universally identified -- if not assumed -- was that appellate judges did their work in two steps. The first step was the collegial decision of the outcome of the case. This step included reviewing the written material before hearing oral argument -- something that was rarely done a generation ago, following the English practice, but is now universal in Canada. This practice of preparation prior to the appellate hearing, known as the "hot bench" approach, nearly always results in the judges individually formulating an initial view of the outcome of the appeal. In most appeal courts, though rarely in the Supreme Court of Canada, the judges discuss informally before the hearing whether there might be some merit to the appeal, and what they think the main contentious issues are. Then there is the hearing of oral argument by a panel, and then a brief post-hearing meeting either to agree on a result or to let counsel for the parties know that the decision is being reserved. (Another variation was informing counsel of the outcome and indicating that the reasons would come later. This happened in less than 2% of the cases, those where the panel felt that the litigants had an urgent need to know the disposition of the case.

The second step was justifying the decision, which included developing the legal reasons for the outcome. These reasons were either summarized orally for the record (most common in BC), written on the appellant's petition (i.e. the writ was endorsed, as in Ontario), or the decision was reserved and later a typewritten text was prepared for distribution to the parties (and to the public, in cases that attracted public attention). On average, provincial appellate judges reserve in about a quarter of the criminal cases and two-fifths of the civil cases. But the proportion of reserved decisions varies in criminal cases from 12% in Manitoba to 80% in New Brunswick, and in civil cases from 16% in Saskatchewan to 96% in PEI.

For the decisions that are not reserved, the decision-making process focuses on step one, with the content of step two being pro forma and perfunctory in most cases. It may be that the primary reason why step two is not engaged more than minimally is that the judges do not perceive any major issues of justice that are raised by the case. Conversely, there may be so little that is problematic about the applicable law in some unreserved decisions that, to the extent that the outcome is problematic, it is a result that is unfortunate but not unjust for the losing party. In courts that reserve frequently because the judges have the time to devote to writing reserved judgments, Step 2 becomes a significant factor only for what Ronald Dworkin refers to as "hard" cases.

It is important to distinguish the decision-making process in trial courts from that in appellate courts. The first of two major differences is that in the trial courts, judges are charged with determining the facts of a situation, and then applying the law to the facts. In contrast, in appellate courts, except in rare cases where the trial judge’s determination about the facts is called into question, the facts have already been established, and the issue is whether the trial judge correctly applied the law to the facts. Second, in the trial courts, judges work alone. Appellate courts, on the other hand, are collegial courts, so the judges on the panel work together to reach a result. There may be one or more dissenting opinions, and even where there is a consensus on the outcome, there might be separate concurring opinions that propose alternative justifications for the result.

In our interviews, we asked the appellate judges to set out what they considered to be the main differences between trial and appellate courts. Half of the responses referred to the trial courts as courts dealing primarily with facts, while the appellate courts dealt primarily with the interpretation of the law. Other responses referred to the collegial process in the appellate courts in contrast to the trial judge working alone, and the greater discretion possessed by appellate judges because of the greater frequency of having to decide difficult or unsettled issues of law.

We asked the judges if they could describe to us the thought process they had developed for deciding on the outcome of appeals (that is, what we call "step one" of the decision-making process in this paper). The responses to this question illustrate the profound differences in the decision-making process between the trial and appellate courts. More than half (55 per cent) said that they had developed no particular thought process for deciding cases. The rest described their thought process, but we could not discern any typologies or distinct patterns. At first we found this result surprising because McCormick and Greene had posed the same question in ninety interviews with Alberta judges -- primarily judges of the trial courts -- in an earlier study, and they found that nearly all the trial court judges had worked out a personal strategy for decision-making. These patterns could be divided into three distinct types for the trial court judges. These groups were labeled "strict formalists" (22 per cent of the judges), "pragmatic formalists" (44%), and "intuitivists" (24%). Only ten per cent could not describe a particular process they were) labeled "improvisers." The ten appellate judges who were interviewed in the earlier Alberta study would have ended up in the "improvisers" category, but because McCormick and Greene could not detect variations in decision-making patterns from the responses to this question by appellate judges, they did not pursue the analysis of the responses to this question.

In contrast, those appellate judges who sketched out a personal thought process during the interviews for the Canadian Appeal Courts Project simply described a very general approach to appellate court decision-making: identify the main issue or issues, check the precedents, form a tentative opinion, listen to counsel, and decide. A few said that the key part of the process for them was the balancing of the individual’s interest against general principles.

In retrospect, perhaps we should not have been so surprised by the appellate judges’ general inability to describe a personal decision-making strategy for appellate decision-making. Appellate judges are team players, and at key points before, during, and after the hearing they need to consult with their colleagues on the panel. The system does not allow them to develop more than the most general idiosyncratic approach.

Seventy per cent of appellate judges have prior experience as trial judges. Some trial judges who are elevated to an appellate court do not adjust well to giving up their role as star of a "one person show," and instead having to share the decision-making process with colleagues. Others miss the challenge of determining the facts. It is not surprising, therefore, that Mr. Justice Angers in the late 1990s requested and received a transfer from the New Brunswick Court of Appeal to New Brunswick’s superior court, a court from which he had been elevated earlier on. What is more surprising to us is that more appellate judges who would prefer to return to the trial court have not requested such a transfer.

We will now consider in more detail the nature of the two-step process in appellate courts.

The two-step process was considered as so natural by the judges we interviewed that it was rarely commented on in any detail; nor was it the subject of debate and contention. In contrast, the question of whether written reasons should be drafted in a preliminary way prior to oral argument in cases where the outcome seemed obvious was quite contentious. Most judges preferred that the decision about the outcome should be made prior to the decision about the content of the reasons, thus clearly separating the two steps in the process.

Step one, deciding the outcome, takes surprisingly little time. Judges commonly receive the written material two weeks before the oral hearing. The judges told us that on average, they take about 6 ½ hours to prepare to hear a typical case by reading through the materials. In the great majority of cases, the panel decides the outcome on the same day that the hearing occurs, or within 24 hours. It is important to remember that when the judges meet in conference after the hearing, this conference is normally very brief. The judges agree on the outcome and possibly the primary reasons for the outcome, and they agree on who will write the main decision. However, they never create an outline for the opinion that the opinion writer is expected to follow, or talk through the finer details of the opinion. We were told that when Madame Justice Wilson joined the Supreme Court, the discussions at the conference tended to become longer, more in-depth, and less structured. Given that Wilson may well be one of the most principled of judges on the Supreme Court, in that she took seriously the tests developed by the courts and tried to ensure their proper application, it is not surprising that Wilson may have attempted to ensure a closer connection between steps one and two through a more detailed discussion at the post-hearing conference.

It is interesting that the judges told us that on average, they felt that about a fifth of the cases that came to their court had no merit, and were simply a waste of judicial time. (In the Supreme Court of Canada, these were mostly the as-of-right appeals in criminal cases, but also included other cases where lawyers made a convincing argument regarding leave that proved not to be substantiated later on.) These cases typically resulted in an oral decision with very little time devoted to the step two analysis.

The judges told us that on average, they changed their minds about how the case should be decided during oral argument about 20% of the time. They all regarded this as a rather high proportion, because they expected, as a rule, not to be surprised during oral argument. Oral argument is an opportunity for the judges to focus on the issues they think are important, and perhaps to delve into details that may not have been explained fully enough in the written submissions. As well, it is an opportunity for counsel to present their arguments in a different way, and thus tempt the judges to doubt their original tentative conclusions.

The second step, when there is a reserved decision, sometimes occurs in close proximity with the first, but written reasons may sometimes not come for months after argument. (Unlike some states in the US, the pay of Canadian judges cannot be docked if the decision is not completed expeditiously. However, the courts have sometimes stayed proceedings in cases of excessive delay in producing written reasons, such as in the Rahey decision of 1987, where the Supreme Court declared that an 11 month delay by a judge in releasing his decision after trial was unreasonable.)

While in some cases the two steps sometimes meld into a seamless process for the hard reserved decisions, in most situations they are divisible as mental processes and also work processes. In our interviews, we asked the appellate justices to tell us how they decided among conflicting lines of precedent in hard cases. A frequent answer in the Supreme Court of Canada, and a relatively common one in the penultimate appellate courts, was that they decided according to what seemed to them to be the most just outcome. When they described the process of writing the opinion for reserved decisions (step two), however, they frequently focused on trying to persuade the parties and their lawyers that the outcome was both correct and just. (Fifty-five percent said that their primary audience in writing decisions is the litigants, while 18 per cent said it was the judges below and the bar.) In other words, step two is the means of legitimizing the judicial preferences in step one. It is the outcome that determines the reasons in most hard cases, more than the reasons that determine the outcome. Exceptions to this general rule occur, however, when during the writing of the opinion things do not "work out" as initially expected. The reasoning process that was expected to fall into place to justify the outcome turns out to be either illogical or unpersuasive, and the writer of the opinion becomes convinced that his or her initial decision about the outcome was in error. The opinion writer has then to try to persuade the other judges who initially agreed with the step one decision that this decision ought to be reversed. If the reasoning of the reversing judge is not persuasive enough, then the decision initially intended to be the unanimous or the majority decision becomes a dissent.

The judges told us that their initial opinion most often firmed up during the process of writing the decision. This sounds normal and routine as a mental and an organizational process, but its implications for legal theory and legal education, let alone for politics, are fundamental. Legal realists and cynical politicians alike say that judges simply follow their policy preferences; court supporters from Oxford's Ronald Dworkin to some provincial MLAs argue that judges are principled and follow the law. Judges may do both, but from the evidence that we gathered, they tend to do the former before they do the latter, at least for the hard cases. Thus it may be that law schools teach their students not how to win cases, but how to help the judges rationalize and justify their value preferences in legal terms.

It is clear from our interviews that when the judges meet after a hearing to decide the outcome of a case and to decide whether the decision will be reserved (and if so who will write the decision), they have only the vaguest notion of what the content of the main written decision will look like. In the Supreme Court, firm decisions have not usually been made about whether there will be one or more dissents, and it has certainly not been decided whether there will be separate concurring decisions on the majority outcome side. It is considered bad form for a judge to draft a dissenting opinion, or a separate concurring opinion, until he or she has seen the main decision. The main decision might convince a potentially dissenting judge either that the majority view is, after all, the best one, or that the issues that might be raised in dissent are relatively minor and therefore not worth the effort. Dissenting opinions are relatively rare in the provincial appeal courts (a high of 15% in Quebec and Manitoba, and a low of 4% in Ontario). In the Supreme Court of Canada, about 60% of decisions are unanimous, but separate concurring decisions are much more common there than in the provincial courts of appeal. For reserved cases, firm decisions about whether there will be dissents or separate concurring decisions, and the content of these dissents or separate concurring decisions, are not made until the draft of the main decision is circulated. There is often an exchange of views among the judges at this point that might resolve particular issues and reduce the number of opinions written. This process of opinion-writing underlines the separate nature of step two.

In the Supreme Court’s post-hearing conference, each judge will summarize very briefly what he or she thinks is the critical reason for the outcome. If the decision is reserved, the judge who ends up writing the draft opinion may be the judge whose turn it is to write a decision. On the other hand, it may be a judge who volunteers (or who is volunteered by the Chief Justice or the senior judge on the panel) because he or she seems to be taking a position that is capable of incorporating the views of most of the rest of the judges. The judge writing the opinion tries not only to justify the decision based on his or her own views, but tries also to incorporate the views of the other judges who appear to be thinking along compatible lines.

Eighty per cent of the appellate judges that we interviewed did not think that it was improper to try to promote a consensus among their colleagues during the conference that occurs after the hearing. Although it is important to respect the differing views of the other judges, most judges saw nothing wrong with a frank exchange of views designed to bring the other judges "on side."

When a judge is working on the draft of the main set of reasons, the critical question he or she asks is, "How can I persuade the litigants, counsel and the judges below that our decision makes sense?" The precedents that are cited are chosen for their persuasive value. The cited precedents are often, but not necessarily, the ones that helped the judge to make up his or her mind in the first place. As one Supreme Court of Canada judge mentioned, "Especially in cases where we're charting a new course, I want to show the litigants and the bar that the ideas that I’m presenting are not just the result of my own personal thinking process. There are other eminent thinkers out there who think along the same lines."

For the hard cases, the picture that emerges is one of some fluidity leading up to the decision about the outcome. A judge's mind begins to be made up -- and remains the same in 80% of the cases -- during the reading of materials. There is some change of mind during the oral argument, and some during the conference after the hearing. The judge drafting the decision takes into account these nuances, but in his or her own individual way. Even if the judges agree on the outcome, no two judges would sit down and draft exactly the same decision. About a quarter of the time, the judge drafting the decision will consult with other judges on the panel about the draft. Once the draft is completed, it is circulated to the other judges on the panel -- and in some provinces like Quebec to the entire court -- and the other judges may provide comments on the original draft. In most cases, the comments are not very extensive. Most appeal courts are busy -- or they make themselves busy by reserving more judgments, as in New Brunswick or PEI -- and so the judges give a certain amount of "artistic license" to themselves to draft decisions in their own way. As Peter McCormick’s research has shown, judges have their own individual preferences regarding which courts they like to cite to back up their reasoning processes. As one Manitoba Court of Appeal judge put it, "I love citing English decisions. I just find them a pleasure to read, and so I tend to cite them more often than other judges on this court."

In the provincial courts of appeal, the process of writing the decision is more informal than is the case for the Supreme Court of Canada. The architecture of the courthouse may play an important role in determining the dynamics among the judges during step two. In New Brunswick, all the judges' offices open into a central area, and so it is easy for any judge, while preparing his reasons, to walk into another judge's office with a draft paragraph and ask, "What do you think about the way I’m proposing to tackle this issue?" In the Ontario Court of Appeal and the Supreme Court of Canada, the judges' offices are not in such close proximity, and so the chats in others judges’ offices are less frequent. In the Supreme Court of Canada, these exchanges are nearly non-existent.

The Supreme Court justices discuss draft opinions with their clerks in much more detail than with their judicial colleagues. The clerks usually draft the parts of the decision that set out the facts of the case and summarize the decisions below. Sometimes they draft parts of the ratio, and one or two judges use these drafts as a working draft for their own ratio. The Supreme Court has a superb research support service, and the staff in the library busy themselves with finding materials that might be useful to the judges in justifying their decisions. It is not uncommon for Supreme Court justices to cite cases not cited by counsel, particularly cases from European and other foreign sources, because the research service is able to do a much more comprehensive search of relevant cases than most counsel are.

Just as academics only have an incomplete idea of how their books and articles are going to read when they are beginning to write them, judges have an incomplete idea about how their decisions will end up looking when they begin drafting them. There are, of course, important variations across courts. In the provincial courts of appeal, the average written decision is only five or six pages long, compared with the average of 100 pages plus in the Supreme Court of Canada. It is a little easier for a provincial court of appeal judge, therefore, to imagine what the decision will look like before writing it. And in the less-busy courts where nearly every decision is reserved, the routine decisions are more predictable than the decisions in the hard cases.

One advantage that appellate judges have over academics is that they are free to use the factums of the winning counsel and/or the trial court decision as the basis of their written reasons. Although there are some appellate justices who from time to time plagiarize the winning factums or a trial court decision as the basis of their opinions, in most cases step two is more complex than merely adopting the reasons in the factum of the winning counsel. Often, there are some lines of reasoning in the written materials that the judicial opinion-writer agrees with (or may assent to with modifications), and others that he or she takes exception to. The written materials are almost always helpful in constructing the opinion, but they are rarely definitive, so as to negate the need for a distinct reasoning process at step two.

In the hard cases, we focused on how the judges said they decided among conflicting lines of precedent. Only a third of the judges provided us with a technical legal explanation: they chose the most recent precedent from the highest authority. The others mentioned other factors important to them: their esteem, or lack of it, for the judges writing the precedent decisions, or a personal judgment about which line of precedents seemed to be the most "just." And by "the most just," the judges were not referring simply to their understanding of the entire body of law after being steeped in it for many years. They were also referring to their personal sense of justice, which went well beyond their legal experience. For example, in an interview on CBC radio, Madame Justice McLachlin mentioned how her own sense of justice had been shaped to some extent by the struggle of her fellow residents of Pincher Creek, Alberta, where she had grown up, to resolve a serious environmental problem related to emissions from sour gas wells. When we asked the judges why they had accepted a judicial appointment, prominent among their reasons was the idea that as a judge, they could make a contribution to the development of the law. It seems to us that one can see the evidence of that contribution more in step two of the decision-making process -- the justification stage -- than in the first stage -- merely deciding the outcome.

Had we known in 1991 what we now know, we would surely have included a question in our survey asking the judges to comment on what they thought were the critical factors in shaping their own personal sense of justice. Of course, judicial biographies are full of interesting insights into this issue. For example, Patrick Boyer's biography of James McRuer shows that McRuer's father was not afraid to stand up to authority figures who abused their power. We can see the same kind of outrage at abuse of power in McRuer’s report on civil rights in Ontario's justice system.

There has been much debate among political scientists and academic lawyers since 1982 about the appropriate role of the judiciary in the policy-making process. Both right-wing Charter critics, such as Ted Morton and Rainer Knopff, and left-wing critics, such as Michael Mandel, have lamented the increased involvement of unelected judges in the democratic policy-making process thanks to the Charter. From Mandel’s perspective, judges have interpreted the Charter so as to favour business and moneyed interests, but from the Knopff-Morton viewpoint, the judges have succumbed to the arguments of "court parties" that have been pushing the interests of mostly left-wing special interest groups. Richard Sigurdson has argued persuasively that both the left and right Charter critics have greatly exaggerated the negative side of the Charter. Charter optimists, such as David Beatty and Lorraine Weinrib, applaud the Charter as a vehicle for defending the interests of minority groups that are sometimes overlooked through the legislative policy-making process, but they have been disappointed whenever the courts do not adopt what they consider to be the "correct" way of applying the Charter. Christopher Manfredi has pointed out the pitfalls of the judicial process for assessing complex social policy issues, and is disappointed that the political branches of government have not been more active in using Section 33 of the Charter to suspend what he considers to be bad judicial decisions. Nevertheless, as Paul Sniderman and his colleagues have discovered, Canadians are much more enamored with judicial policy-making than with the public policy record of elected politicians. Peter Russell, a court-watcher who takes a more neutral stance about whether the Charter enhances Canadian democracy, has advocated more appropriate recruitment and appointment procedures for judges so that the judges will be more likely to be up to the challenges posed by the Charter and other justice system issues. Michael Ignatieff sees the Charter as an important manifestation of the global "rights revolution," marking a shift in emphasis from viewing the individual as a cog in the wheels of state, to the state as a vehicle for protecting and promoting individual self-worth and dignity.

In Final Appeal, we argued that democracy can usefully be viewed as government based on the principle of mutual respect. Mutual respect implies the intrinsic worth and equal importance of every individual in society. Thus, there is a necessary role in a democracy for partisan elected legislatures to create and revise the law, and also for impartial judges to interpret and apply the law so as to ensure respect for the rights of minorities and the disadvantaged. To push this argument further, judges and legislators approach public policy issues from different perspectives, and they ask different questions. When faced with making policy choices, cabinet ministers, other elected legislators, and public servants involved in the policy process, ask what will work to fix a particular problem, and whether the proposed solution is likely to have the requisite degree of public support to help get the government re-elected or avoid having the government defeated. Judges, on the other hand, ask whether a legislated policy is constitutional, and in the gray areas of constitutional jurisprudence, whether an impugned policy results in enough justice to justify its existence. Although what John Rawls refers to as "justice-making" considerations do enter into the calculations of elected politicians and government policy analysts, they are not usually front and centre as they are in the minds of judges.

The two-step decision-making process in appellate courts underlines the distinct approach of judges to policy issues, and the importance of this role in the democratic context. Step one of the appellate decision-making process, determining the outcome, is tied into both the application of the law, and issues of justice when the law allows for discretion. Step two is concerned with developing persuasive arguments about the justice of the step one outcome.

The reaction of many lawyers we have talked to regarding the two-step process in appellate courts has often been, "Thank goodness that the judges are open to considering arguments about just outcomes rather than just black-letter law." Political scientists, on the other hand, often remark, "This just confirms our worst suspicions about how judges are injecting their personal preferences into the policy process."

We tend to side with the viewpoint of our lawyer colleagues. It is ironic that the same political scientists who often appear to their students as cynical about the effectiveness of our political institutions also tend to gloss over the defects of the policy-making process in government when they criticize judicial involvement in policy-making. They claim that policy-making is properly the outcome of legislative deliberations that have no need for reminders from judges that democracy is also about striving for justice. The two-step process in appellate decision-making, which is about making just decisions and then defending them as legally legitimate, is distinct from the multi-step process of policy development in the partisan political branches of government, which is about the allocation of scarce resources among competing demands in ways that are calculated to lead to re-election. In a polity based on the principle of mutual respect, both partisan politics and judicial review play vital roles.