Whether a judge dissents may be influenced not only by his policy views but by how busy he is in deciding other appeals. A number of factors affect a judge’s workload. There is of course the total number of cases heard by the court, although this is not an exogenous factor for many high courts given they can select both the number and type of cases they hear. At the most basic level, as the number of cases before the court increases, there is a greater probability that some will be easier and less divisive.

However, a judge’s workload will not only depend on the total number of cases heard by the court but on the number of cases he has to hear. It takes time for a judge to write a dissenting judgment or a concurrence, and a busy judge may be less inclined to do so.[1] Although a judge with a lower caseload may have (and use) the extra time to forge compromises and reach consensus,[2] his opportunity cost of dissenting will clearly increase as the number of cases he hears increases. The number of cases per judge will depend on the total number of judges that sit on the court, as well as whether the court hears cases en banc or in panels, and on the size of those panels.[3]

An additional factor that may impact the cost of disagreement for a judge is whether a court employs law clerks. Law clerks will decrease the judges’ workload, allowing for more time to author dissents. Of course, an opposing theory applies here as well: law clerks may give judges more time to reach a compromise and build consensus. The use of law clerks varies widely across high courts. In the United States and Canada, law clerks are given extensive responsibilities for analyzing material and helping draft opinions.[4] Justice Scalia wrote in Conroy v. Aniskoff: “I confess that I have not personally investigated the entire legislative history—or even that portion of it which relates to the four statutes listed above. The excerpts I have examined and quoted were unearthed by a hapless law clerk to whom I assigned the task.”[5] In Australia, on the other hand, “associates” are given less responsibility, particularly in drafting opinions.[6] Justice Kirby of the High Court of Australia commented “I’m far too egocentric and idiosyncratic to allow anybody to write my peerless prose”[7] For the Indian Supreme Court, clerkships are less prestigious, and the use of clerks by judges varies significantly across judges.[8] Although we will not analyze the impact of clerks on judicial decision-making, they clearly have the potential to play a significant role.[9]

Leaving aside the issue of clerks, we will examine the following hypothesis:

H4: The consensus rate will increase as the workload increases, with workload depending on the total number of cases heard by a court and on the cases heard per judge.

As with other hypotheses, empirical findings have been far from unanimous. Epstein, Landes, and Posner found that a 10 percent increase in workload was associated with a 3.3 percent decline in the dissent rate on the US Supreme Court. Similarly, for the High Court of Australia, Smyth found an increase in caseload had a negative effect on dissent rate, with a 1 percent increase in caseload per judge correlating with a 0.33 percent decrease in dissent rate.[10] Further, Narayan and Smyth conclude that increases in GDP per capita also have a negative effect on dissent rate, as this leads to more litigation and therefore a higher caseload.[11]

However, Walker, Epstein, and Dixon find no causal association between increased caseloads in the form of court filings and rates of individual expression on the uS Supreme Court, and in fact found the increase in disagreement preceded the increase in filings.[12] However, they also note that the number of decided cases declined over time, which, although they do not analyze it, may provide evidence of a decreased workload leading to greater disagreement.[13] Similarly, neither McCormick nor Songer, Szmer, and Johnson find a correlation on the Supreme Court of Canada between workload and dissents.[14]

When we look at our annual data for the period 1970 to the early 2000s, there is only limited evidence of a correlation between the caseload of a court and measures of dissent.[15] The total caseload of the court is negatively correlated with the percent of cases with at least one dissent in the uK, but the effect is very small. There is no correlation with the probability of a dissent in a case in Canada, the united States, Australia, or India. Similarly, when we look at the relationship between the total caseload for a court and the fraction of opinions that are dissents (as opposed to whether a decision has at least one dissent), there is no statistically significant relationship in Canada, the united States, or India. In India at least, the caseload may be so high that there is only very limited appetite for dissenting at any of the levels seen in this period. There is only weak evidence that judges in Australia or the uK respond to the total caseload; in both cases the effect was as expected, with a higher total number of cases correlating with a lower fraction of dissenting opinions. In both cases, however, the effects disappeared when we control for who is the chief justice.

A judge may not care about the total number of cases heard by the court so much as the average (how many cases she on average hears in a given year).

However the results using average caseload (number of cases in a year divided by the number of judges hearing cases) are largely consistent with the story, looking at the total number of cases. The principal exception is the uK. For the uK House of Lords, there is a stronger negative relationship between the fraction of dissenting opinions and the caseload per judge, and this relationship does not go away when we account for differences across chief justices.

As with policy preferences, the story of the relationship between workload and dissenting is not clear-cut. Although there is some evidence of a negative relationship, that evidence is not strong, and even when there is a relationship, the effect does not appear to be large.

  • [1] Smyth, “Variations in Dissent,” supra note 75 at 224.
  • [2] McCormick, “With Respect,” supra note 18 at 108.
  • [3] Further, the workload will vary with vacancies on the court. The workload per judge increaseswhen the court is below complement, leading to less time and fewer resources to author dissents(to the extent the vacancies are not offset by a court accepting fewer cases on appeal).
  • [4] For the United States, see Todd C. Peppers & Christopher Zorn, “Law Clerk Influence onSupreme Court Decision Making: An Empirical Assessment” (2008) 58 DePaul Law Review 51at 56. For Canada, see Lorne Sossin, “The Sounds of Silence: Law Clerks, Policy Making and theSupreme Court of Canada” (1996) 30 University of British Columbia Law Review 279 at 287-298.
  • [5] Jonathan Horton, “Justices’ Associates: Some Observations” (2002) 22:1 University ofQueensland Law Journal 114 at 119 [Horton].
  • [6] Horton, supra note 82 at 119. South Africa has in the past essentially not used law clerks,which may have contributed to a high judicial workload and reliance by judges on oral argument and legal briefs. Stacia L. Haynie & Kaitlyn L. Sill, “Experienced Advocates and LitigationOutcomes: Repeat Players in the South African Supreme Court of Appeal” (2007) 60:3 PoliticalResearch Quarterly 443 at 451.
  • [7] Horton, supra note 82 at 121.
  • [8] Abhinav Chandrachud, “From Hyderbad to Harvard: How U.S. Law Schools Make Clerkingon India’s Supreme Court Worthwhile” (2014) HLS Program on the Legal Profession ResearchPaper at 24, online: [Chandrachud, “Hyderbad”].
  • [9] The effect of law clerks on the connection between legal regimes is also an important areafor further research. See e.g. Marie Gren, “An Empirical Study of Judicial Fertilization: Theuse of Foreign Clerks by the Israeli Supreme Court” (paper delivered at the 9th WorldCongress Conference, “Constitutional Challenges: Global and Local,” Oslo, June 16-20, 2014);Chandrachud, “Hyderbad,” supra note 85.
  • [10] Smyth, “Variations in Dissent,” supra note 75 at 238; Smyth, “Role of Attitudinal,” supranote 51 at 535.
  • [11] Narayan & Smyth, “What Explains Dissent,” supra note 75 at 420.
  • [12] Walker, Epstein & Dixon, supra note 8 at 367-368. In contrast, on courts of appeals, it hasbeen found that an increased caseload makes authoring a dissent “quite unattractive,” which isconsistent with our hypothesis (see, for an early example, Burton M. Atkins & Justin J. Green,“Consensus on the United States Courts of Appeals: Illusion or Reality?” (1976) 20:4 AmericanJournal of Political 735 at 746).
  • [13] Walker, Epstein & Dixon, supra note 8 at 368, n. 4.
  • [14] McCormick, “With Respect,” supra note 18 at 108; Songer et al., “Explaining Dissent,” supranote 22 at 404 (examining influences on dissents between 1976 and 2005 including workloaddefined as the number of cases heard by the Court in that term and finding no statistically significant relationship between workload and dissents, which they attribute to a relatively stablecaseload over the period and docket control by the Court). As for Israel, Weinshall-Margelhas hypothesized that a higher caseload will diminish the impact of justices’ attitudes on theirdecisions, thereby reducing the frequency of dissent, but has not tested this hypothesis fully(Weinshall-Margel, “Attitudinal,” supra note 40).
  • [15] See Appendix 3 for regression results relating to fraction of dissenting opinions and to caseswith at least one dissent.
< Prev   CONTENTS   Source   Next >