Home Economics Commitment and cooperation on high courts : a cross-country examination of institutional constraints on judges
OTHER JUDGES ON THE PANEL
A court may reach different outcomes depending on the diversity of its membership. The diversity may be in policy preferences, ideological outlook, or more personal characteristics such as gender or race. Diversity, or the lack thereof, will depend on the appointment process to the bench as a whole as well as the nature of the panel-setting process. It may also vary for a given bench depending on the area of law, such as where a court is composed of judges who are uniformly conservative on economic issues but mixed on criminal matters.
If a judge compromises on his ideal outcome despite being in the majority on a panel, it might be because the minority judge simply raises considerations that would not have occurred to him otherwise. Alternatively, the minority judge may serve as a “whistleblower” who threatens to expose the rest of the panel’s ideological extremity or disregard for legal doctrine by writing a dissenting opinion. The potential whistleblower motivates the majority to base their decision more firmly on doctrinal grounds rather than personal preference.
Similarly, a minority judge may persuade others to adopt her position if she feels more strongly about the outcome of a particular case than they do. The majority may find her conviction persuasive, or may choose to partly cooperate in hopes of unanimity (as a judge with strong preferences will be more willing to bear the costs of dissenting) or her voting their way in cases they find particularly important. The author of the majority opinion usually enjoys an agendasetting advantage, but may have to compromise her position in order to obtain a majority coalition. Conversely, judges sitting on a panel of like-minded judges may be more likely to follow their policy preferences because they don’t need to consider complex strategies.10
Differences on panels may then influence whether a judge dissents. A judge on a mixed panel may be less likely to dissent, because the majority opinion will be more likely to rest on doctrinal considerations if they fear the minority judge will act as a whistleblower.11 Moreover, as we discussed in Chapter 2, writing a dissenting opinion requires effort and potentially imposes reputational costs with colleagues, and so a judge may not dissent unless he anticipates a benefit that offsets that cost.12 He may then shirk the additional work by assenting to the majority opinion despite not agreeing entirely with its reasoning. However, he may consider the costs of writing the dissent to be justified if the differences between panel members are wide enough to preclude acceptable compromises.13
not only about winning but the size of the majority (Epstein & Jacobi, supra note 1 at 348). David W. Rohde, “Policy Goals and Opinion Coalitions in the Supreme Court” (1972) 16:2 Midwest Journal of Political Science 208 at 671-672 [Rohde, “Policy Goals”]. Rohde argued that this was the reason for the predominance of 5-4 rulings in civil liberties cases decided by the Warren Court. Alternative theories predict that cases with ideologically extreme decisions by lower courts are likelier to result in broad majorities. The rationale for this argument is that the likelihood of a particular judge joining the majority opinion increases in relation to his ideological closeness to the median justice within the majority coalition. In cases where the median of the majority aligns with the median of the court, judges with extreme preferences in either direction are more likely to dissent or concur, whereas opinions that favor a particular policy direction can gain support from judges who share that preference and those who occupy the center of the court. See Tonja Jacobi, “Competing Models of Coalition Forming and Case Outcome Determination” (2009) 1:2 Journal of Legal Analysis 411 [Jacobi, “Competing Models”]; Cliff Carrubba et al., “Who Controls the Content of Supreme Court Opinions?” (2012) 56:2 American Journal of Political Science 400 (arguing that the decision is most heavily influenced by the median justice in the majority). See Chapter 8 for a discussion of consensus.
The influence of panel diversity may vary by area of law. Judges may be less affected by their colleagues in ideologically divisive cases.14 Further, if members of the high court are strongly divided on political issues, they may be more likely to form coalitions with their ideological allies in order to prevent their “foes” from reaching a majority.15 The reasonably stable split between Democratic and Republican appointees on the uS Supreme Court may provide an illustration of this phenomenon.16 Conversely, on the less politically diverse courts in Canada and the uK, certain pairs of judges may be more likely to vote together, but these coalitions are not as consistent as they are in the united States, and do not have a significant effect on final outcomes.17
These potential impacts of panel diversity and panel size lead to the following hypothesis:
H1: Judges will be more likely to reach ideologically extreme decisions when sitting with ideologically similar judges than on more diverse panels.
e.g. Robert H. Dorff & Saul Brenner, “Conformity Voting on the united States Supreme Court” (1992) 54:3 The Journal of Politics 762 [Dorff & Brenner]; Forrest Maltzman & Paul J Wahlbeck, “Strategic Policy Considerations and Voting Fluidity on the Burger Court” (1996) 90:3 The American Political Science 581 [Maltzman & Wahlbeck, “Burger Court”]; Saul Brenner, Timothy Hagle & Harold J Spaeth, “Increasing the Size of Minimum Winning Original Coalitions on the Warren Court” (1990) 23:2 Polity 309.
Most of the literature regarding these panel effects has focused on the US Federal Courts of Appeals, which hear cases in three-judge panels. The judges have been found to reach “ideologically amplified” decisions when the rest of the panel shares their political views. Conversely, the presence of a judge with conflicting preferences is associated with an “ideological dampening effect” Sunstein et al. found that the effect of both individual preferences and panel composition varies depending on the area of law involved. Judges were influenced by their partisan commitments as well as the dispositions of other panel members in several areas, including affirmative action, sex discrimination, disability discrimination, sexual harassment, piercing the corporate veil, campaign finance, racial discrimination, contract violations, and environmental regula- tions. In these fields, Democratic judges were substantially more likely to vote for outcomes classified as “liberal” than their Republican colleagues. Moreover, both groups of judges were more likely to support liberal decisions when the other two panel members were Democrats, and correspondingly less likely to vote liberally when sitting with two Republicans. However, neither individual nor panel preferences appears to influence judicial voting in cases that involve criminal appeals, federalism, or takings claims. A third group of decisions, including abortion and capital punishment issues, was dominated by ideological voting. In these contentious areas, Democratic- and Republican-appointed judges voted in a predictable ideological pattern that was not affected by which colleagues they sat with.
Similarly, Cross and Tiller found that judges on divided panels are almost twice as likely to follow legal doctrine that conflicts with the majority’s policy preferences. In DC Circuit Court of Appeals decisions citing Chevron USA v. Natural Resources Defense Council Inc. (a key US administrative law decision), in order to reach their preferred outcome, politically unified panels did not follow the principle of deference established in Chevron two-thirds of the time. In contrast, almost the same percentage of politically divided panels followed Chevron despite a conflict with the majority’s policy goals. These results may show a strategic response to the minority judge’s potential role as a “whistleblower” who reveals the majority’s disobedience by writing a dissent.
Panel effects also appear to play a significant role in the decision-making of judges on the Ontario Court of Appeal. Applying the methodology used by Sunstein et al., Stribopoulos and Yahya found that judges’ decisions vary significantly depending on which party appointed them; however, on panels composed of both Liberal and Conservative appointees, the influence of political ideology is dampened. This was particularly evident in cases involving Charter claims or criminal appeals. Stribopoulos and Yahya suggest that this trend reflects a tendency for judges to engage in more deliberation on heterogeneous panels, and thus reach different results.
The composition of the rest of the court (not just the panel) may also matter. Kastellec found that the probability of reaching a liberal decision increases in relation to the number of Democratic judges sitting on the panel, and that the presence of a single Republican “counterjudge” had significantly more influence when the rest of the circuit was similarly aligned. In Democratic circuits, panels with two Democrats and one Republican voted liberally 52 percent of the time, whereas those composed entirely of Democrats reached liberal outcomes 62 percent of the time. In contrast, Democrats in a Republican circuit were 67 percent likely to vote liberally on an ideologically homogenous panel, but only did so 48 percent of the time when one Republican judge was on the panel. Interestingly, Republican majorities in Democratic circuits voted liberally in 35 percent of cases, regardless of whether the panel included a Democrat; however, in Republican circuits, adding a Democratic counterjudge increased the chances of a liberal vote from 27 percent to 34 percent.
Judges may be influenced not only by the political leanings of their colleagues but also by diversity in personal characteristics such as gender or race. For instance, both the ideology and race of judges on a panel appeared to influence federal appellate cases involving affirmative action programs. The difference between panels that included a black judge and those that did not was striking: panels with three non-black judges voted in favor of affirmative action in 53 percent of cases, yet adding a single black judge to the panel increased the chances to 81 percent. More surprisingly, two non-black Republican judges sitting with a black colleague were 86 percent likely to vote in favor of affirmative action—significantly higher than the 67 percent likelihood of a Republican casting a pro-affirmative action vote when sharing the panel with two non-black Democrats, even though the former judges constitute a majority and the latter does not.
Similarly, both race and ideology have been found to influence judges’ decisions in cases involving the Voting Rights Act in the United States. In fact, the racial composition of a panel appears more significant than its partisan alignment in these cases. Black judges were found to rule in plaintiffs’ favor a little over half of the time, regardless of their political party, and the presence of a black judge increased the probability of non-black panel members voting for the plaintiff by nearly 19 percent.
The presence of female judges may also have a significant effect on case outcomes. In employment discrimination cases, for example, female judges were 20 percent more likely to rule in favor of plaintiffs. Their effect on other panel members is similarly significant, as the probability of a male judge voting for the plaintiff increases by 19 percent when he sits with a female colleague. Boyd, Epstein, and Martin found female judges also influenced their colleagues in sex discrimination cases, but did not find evidence of sex-based panel effects in any other relevant areas, including sexual harassment. They argue that their findings support an informational theory of group decision-making, which presumes that female judges have knowledge about specific issues that influences how they and their male colleagues vote in gender-related issues. Panels on the Ontario Court of Appeal including male and female judges also displayed unique voting trends, especially in criminal cases involving a gender or violence-related component, sentencing appeals, and family law cases.
There is, however, inconsistency in results in the panel effects literature. Fischman recently reviewed 11 major studies and used a new empirical model to interpret their results. His analysis assumes that judges are influenced by their colleagues’ votes rather than the presence of judges with divergent preferences or characteristics. Doing so shows consistent evidence of panel effects across several areas, including several fields in which Sunstein et al. claimed judges were not influenced by their colleagues, such as criminal appeals and capital punishment cases. Fischman reports that each vote cast by a colleague increases by approximately 40 percent the likelihood that a particular judge will vote in the same direction. Most significantly, his results remain true regardless of whether panel members are distinguished on the basis of party affiliation, gender, race, or other characteristics. Based on these results, he concludes that prior studies on panel effects can be explained at least in part by a strong norm of consensus.
So far, little research has been done regarding panel effects on high courts. However, analyses from Canada, the uK, and Israel suggest that judges from these countries may be similarly influenced by their colleagues’ views. Jilani, Songer, and Johnson argue that the appointment of female justices to the Supreme Court of Canada had a significant influence on the voting patterns of male judges. Studying the court’s decisions from 1967 to 2007, they found that with the addition of each female colleague, male judges became 9 percent more likely to vote for liberal outcomes in cases involving equality and discrimination, criminal law, or civil liberties. However, non-Charter criminal appeals were an exception to this trend—with each female appointee, male judges reportedly became only 3 percent more likely to vote conservatively (in favor of the prosecution).
The law lords on the UK House of Lords also appear to be influenced by which colleagues sat on a panel with them. For example, Lord Templeman’s participation in a tax-related case correlated with a 65 percent chance that the state would prevail; yet in cases decided without him, the state’s success rate dropped to 11 percent. Further, a particular judge’s influence over the disposition of a case varies depending on the other panel members’ identities. For example, in cases heard by Lord Lowry, but not Lords Ackner or Keith, the likelihood that the weaker party will succeed is approximately 80 percent (whereas the average success rate of weaker litigants is 43 percent). Interestingly Lord Keith did not have a significant effect on case outcomes unless he was accompanied by Lord Ackner.
Similar analyses of the Supreme Court of Israel also offer some evidence that panel effects influence the justices’ behavior. For instance, the voting trends of frequently occurring panels appear to vary more significantly than those of the individual judges, suggesting that certain combinations result in amplified decision-making. Eisenberg, Fisher, and Rosen-Zyi also find weak support for the theory that female judges influence their colleagues’ voting behavior; however, excluding a single judge from these results (Justice Naor) seems to negate the influence that gender has on final outcomes.
Examining panel effects on high courts is difficult because, unlike for the US Court of Appeals, judges are not randomly assigned to panels. As we saw, some individual such as the chief justice has the power to alter the composition, meaning that we cannot tell directly if who was on the panel influenced how a judge voted, or if a judge was assigned to a panel because of how she was likely to vote. However, it is interesting to look to see if there are correlations between the characteristics of different panels of judges and outcomes.
If we look at our data for the period from 1970 to the early 2000s, there is some limited evidence that the composition of panels is related to how the panel decides. Although we could imagine that a judge would be more likely to vote in a liberal direction as the fraction of liberal appointees she is sitting with increases, such a connection is not borne out for all countries. A Supreme Court of Canada judge on a panel of all liberal appointees has a 4.5 percent higher probability of voting in a liberal direction than if the panel is composed of all conservative appointees. Similarly, for the UK House of Lords a judge on an all-liberal panel has a 7 percent higher probability of a liberal vote than if there was an all-conservative panel. However, there was no effect in the United States or Australia.
What seems more important is the average lifetime liberal voting rates of the other panelists. One way to look at the effect of how liberal are the other panelists is to look at the probability of a judge voting in a liberal direction when the other panelists have the lowest average lifetime liberal voting rates as opposed to the highest rates for that country. In Canada, a judge had just over a 50 percent probability of voting in a liberal direction if she is sitting with judges with the highest average lifetime liberal voting records, as opposed to just under 40 percent when sitting with judges with the lowest average lifetime liberal voting rate. Similarly, an Australian judge’s probability of voting in a liberal direction fell by nine percentage points. For the UK the difference was much more extreme. A judge whose other panelists had the lowest average lifetime liberal voting rate (which was about 30 percent) had only an 11 percent probability of voting in a liberal direction whereas if she was sitting other panelists having the highest average lifetime liberal voting rate (which was about 70 percent), she had an over 90% probability of voting in a liberal direction.
The composition of the panel is then related to the rate at which a judge votes in a liberal direction (again remembering that we cannot sort out causation given the data). Does it also relate to how likely a judge is to dissent? The answer seems to be yes. Figure 5-1 shows the difference in dissent rates for judges whose lifetime liberal voting rate is at the average lifetime liberal voting rate of the other panelists as compared to one whose lifetime liberal voting rate is higher or lower by 20 percent (e.g., a judge with a panel average lifetime liberal voting rate of 45 percent is compared to a judge whose lifetime liberal voting rate is either 25 percent or 65 percent). The greatest effect is for Canada and Australia. A Supreme Court of Canada judge’s probability of dissenting jumps from about 5 percent if she is similar to the average of the other panelists up to a third if she has a 20 percent higher or lower liberal voting rate than the average. For Australia the increase is from about a 10 percent probability of dissenting to about a third for a similar difference of the judge from the average of the other panelists. The increase in probabilities of dissenting was less for judges on the US Supreme Court though from a higher base rate of dissenting (from about 15 percent to about 35 percent) and much less for UK lords (from about 5 percent to 9 percent). Therefore, judges whose liberal voting tendencies are further from the average of others on the panel tend to dissent at a greater rate. The greater difference in dissent rates in Canada and Australia from a similar departure from panel averages may be related to the tighter spread of liberal voting rates in those countries.
As with most of us, judges are influenced by whom they are talking to when making decisions. There seems clear evidence of this effect at the lower court level. It would be surprising if this effect did not also play a role at the high court level, particularly as the composition of panels on many high courts can vary widely. The correlation between some measures of panel diversity and
Figure 5-1 Difference in probability of a judge dissenting if further from average lifetime liberal voting rate of other panel members. These probabilities were found by regressing whether or not a judge dissents on the difference between the judge’s lifetime liberal voting rate and the average lifetime liberal voting rates of the others on the panel, the party appointing the particular judge, the direction (liberal or conservative) of the lower court decision, and the area of law. In each case the effect of the difference from the panel average is significant at least at the .01 level except the UK, which is only significant at the .1 level. The difference between the judge and the rest of the panel is related to a greater increase in the dissent rate in Canada and Australia than the United States or the UK.
either voting or dissenting points in this direction, though more work is needed on both the measures of diversity and a means for clarifying causation. What about another aspect of how judges may relate to each other—their position on the bench?
|< Prev||CONTENTS||Next >|