Desktop version

Home arrow Economics arrow Commitment and cooperation on high courts : a cross-country examination of institutional constraints on judges


Norms, Leadership, and Consensus

Lord Ackner was supposed to have observed that one only dissents when one’s sense of outrage at the majority decision outweighs one’s natural indolence, so it could be said that I am recommending more indolence.

—Lord Neuberger, President of UK Supreme Court (2012)'

So far we have mainly discussed how different rules influence voting by judges, such as rules about appointment processes, jurisdiction of courts, or interveners. Norms may also be very influential. The Supreme Court of India, for example, has a norm that junior judges on panels defer to the senior judge, which is powerful given the prevalence of two-person panels.2 In the UK there may have been a norm of making “professional” as opposed to political appointments to the House of Lords, which would affect the nature of the ideological division on the court. These norms may change over time. For example, there may have been a radical shift in norms of consensus on the US Supreme Court in the early 1940s, playing a part in a sharp rise in the number of dissents.3

Norms then lie in the background of the formal rules but can have a large effect. By “norms,” we generally mean, following McAdams, “informal social regularities that individuals feel obligated to follow because of an

  • 1. Lord Neuberger, “No Judgment—No Justice” (Paper delivered at First Annual BAILII Lecture, 20 November 2012), online:
  • 2. Nick Robinson, “Structure Matters: The Impact of Court Structure on the Indian and U.S. Supreme Courts” (2013) 61:1 American Journal of Comparative Law 173 at 186.
  • 3. Cass R. Sunstein, “Unanimity and Disagreement on the Supreme Court” (2015) 100 Cornell L.R. 769 at 772 [Sunstein, “Unanimity”].

Commitment and Cooperation on High Courts. Benjamin Alarie and Andrew J. Green.

© Oxford University Press 2017. Published 2017 by Oxford University Press.

internalized sense of duty, because of a fear of external non-legal sanctions, or both.”[1] These norms then have a “normative” content—that is, a sense they should be followed—and can be enforced through internal shame, guilt, or pride or through the esteem or disesteem of others.[2] They can reinforce, contradict, or fill in the gaps in more formal rules and may change as the context changes.[3]

This chapter examines how norms may influence voting by judges. It is hard to observe norms and to identify their effects.[4] We focus on one particular norm— the norm of consensus. Different courts may have different norms about the need for the judges to agree on a particular result or a set of reasons. Although some studies have defined a norm of consensus as meaning that a court has a high level of agreement,[5] countries can have a norm that favors a high level of agreement (a norm of high consensus) or that tolerates significant disagreement (a norm of low consensus).[6]

Figure 8-1 returns to an issue we discussed briefly in Chapter 1—the difference in the percent of cases per year that involved at least one dissent. The median dissent rates varied widely over the 1970s through the early 2000s, from less than 2 percent in India to about 40 percent in Australia and over 50 percent in the uS Supreme Court. Canada and the uK had more intermediate levels of dissents. Differing norms of consensus may be part of the explanation for these varying dissent rates, though as we will see other factors are also important. The ordering of countries is similar in recent years—with

Percent of cases per year that have at least one dissent over the period from 1970 to the early 2000s

Figure 8-1 Percent of cases per year that have at least one dissent over the period from 1970 to the early 2000s. The box shows the middle 50 percent of the data for each country, and the line in the box shows the median percentage for cases with at least one dissent over the period. The lines above and below each box provide the maximum and minimum annual percentages of opinions with at least one dissent for each country in the period. The united States had by far the highest rate of dissenting, with over 50 percent of opinions having a dissent in most years. At the other extreme, most years India had less than 5 percent of opinions with at least one dissent. Canada and the uK had similar rates of opinions with dissents. Australia had the second highest level of dissent with in general over 40 percent of opinions having a dissent. note that this Figure is the same as Figure 1-3.

the US Supreme Court on the high side with at about 60 percent for 20142015,[7] Australia at about a third of the cases in 2012,[8] the Canadian Supreme Court about 30 percent in 2015,[9] and the UK Supreme Court having about 24 percent of cases with at least one dissent between 2009 and 2013.[10]

The percentage of cases with at least one dissent for each country, 1970 through the early 2000s

Figure 8-2 The percentage of cases with at least one dissent for each country, 1970 through the early 2000s. The united States had uniformly high and India uniformly low dissent rates over the whole period. Australia, Canada, and the uK started the period with about a third of cases having at least one dissenting opinion, but diverged in the mid-1970s. Australia rose to uS levels by the end of the period. Canada and the uK followed similar patterns, with the percent of cases with a dissent dipping before rising again in the 1980s.

These dissent rates may change over time. Figure 8-2 shows how the percent of cases involving at least one dissent varied across the countries between 1970 and the early 2000s. The united States had some variation in the dissent rate but was almost always over 50 percent. Interestingly prior to 1941, justices on the uS Supreme Court dissented at an even lower rate than the current uK Supreme Court with a uS Supreme Court dissent rate of under 15 percent for the period from 1900 to 1941.[11] The Indian Supreme Court has uniformly had few dissents. However, Australia, Canada, and the uK showed more variation. They all started the period with very similar dissent rates, but justices on the Supreme Court of Canada and the uK House of Lord dissented less often beginning in the mid-1970s whereas the Australian High Court judges began dissenting more often. Dissent rates in the uK and Canada began to creep up again in the early 1980s, though ending up slightly less than they started the period. The Australian High Court, on the other hand, ended the period with similar dissent rates to those in the united States.

Proportion of Written Judgments That Are Plurality/Majority Opinions, Concurrences and Dissents

Figure 8-3 Proportion of Written Judgments That Are Plurality/Majority Opinions, Concurrences and Dissents. The time periods covered are India and South Africa 1970-2000, UK, and Australia 1970-2002 and Canada and the United States 1954-2012. source: HCJD, Supreme Court of Canada Database and the US Supreme Court Database.

The dissent rate does not completely capture the level of disagreement on a court.[12] Three variables that are often studied are dissenting opinions, concurring opinions, and dissenting votes.[13] Figure 8-3 shows the percentage of written judgments on each court that was a majority or plurality judgment, a concurrence, or a dissent. The countries are ranked left to right from the highest levels of dissent to the lowest as in Figure 8-1. At the two ends of the spectrum are India with an extremely low rate of dissent (much lower than 10 percent) and the united States with a very high rate of dissent (almost 40 percent).

However, Figure 8-3 also shows that focusing on the dissent rate alone masks other forms of disagreements expressed through concurring judgments. If both dissents and concurrences are taken into account, the united States is an even greater outlier with about 60 percent of written decisions being either dissents or concurrences. Also on this measure the uK has a greater combined level of disagreement than Canada, as its lower dissent rate is offset by its higher concurrence rate. India remains at the extreme low end of disagreements, with few dissents or concurrences.

The absence of concurrences written in Australia points to a difficulty in analyzing rates of consensus across courts. Australia has been described as having “no single consensual norm,” in part because of its tradition of seriatim judgments.[14] In a court with a tradition of seriatim judgment writing, there is no organized majority or minority, thus making it difficult to identify disagree- ment.[15] Further, the factors that affect consensus may vary over time. Both the uS Supreme Court and the Supreme Court of Canada, for instance, began by disseminating their judgments seriatim, adopting a majority style only later (prior to the beginning of our study).[16] In order to allow comparison across courts and analyze core disagreement, we will focus on dissents rather than concurrences, although recognizing that it is only one form of disagreement.[17]

In this chapter we examine how norms of consensus relate to these patterns of agreement and disagreement. Before we get to the factors that affect consensus, it is important to note the underlying normative question of whether consensus is particularly desirable at all. Opinions are divided on this point. On the one hand, a united bench boosts the legitimacy of the court as a whole. In the united States, this was the view of Chief Justice Taft, who desired certainty in the law, and Chief Justice Marshall, who diverted the uS Supreme Court from its previous seriatim tradition toward a practice of writing a single “opinion of the Court”[18] The counterargument centers on the beneficial effects of a multiplicity of views. Forcing a unanimous judgment may lead to unwanted compromise (or extremely narrow “lowest common denominator” decisions), or may cause the omission of an important point that ought not to be overlooked.[19] We will not enter into this normative argument in this chapter but instead focus on what influences whatever level of consensus exists on a particular court.

As we noted earlier it is impossible to observe norms of consensus directly. We therefore need to try to separate these norms from the other factors we have discussed that may influence how judges make decisions. In particular, in Chapter 2 we focused on policy preferences, reputation, identity, compensation, and leisure. To get a better idea of how norms influence levels of consensus across courts, we need to unpack how these factors relate to agreement and disagreement on a court. Following Epstein, Landes, and Posner, we argue that a judge will not dissent, or write a concurrence, unless the benefits of the dissent or concurrence outweigh the costs.[20] We group the factors influencing these costs and benefits into three categories: policy preferences, workload, and norms.[21]

  • [1] Richard H. McAdams, “The Origin, Development and Regulations of norms” (1997) 96Michigan Law Review 338 at 340. Caldeira and Zorn argue that norms may represent an equilibrium in which “no actor, given current information and its current position, can improve itsposition on its own” and therefore may also be followed out of self-interest. Gregory Caldeira& Christopher Zorn, “Of Time and Consensual norms in the Supreme Court” (1998) 42:3American Journal of Political Science 874 at 876 [Caldeira & Zorn, “Of Time”].
  • [2] Richard H. McAdams & eric B. Rasmusen, “norms in Law and economics” (2005),online: at 12-13.
  • [3] Caldeira & Zorn, “Of Time,” supra note 4 at 876.
  • [4] Caldeira & Zorn, “Of Time,” supra note 4 at 875.
  • [5] Russell Smyth, “Historical Consensual Norms in the High Court” (2002) 37:2 AustralianJournal of Political Science 255 [Smyth, “Historical Consensual”]; Thomas G. Walker, LeeEpstein & William J. Dixon, “On the Mysterious Demise of Consensual Norms in the UnitedStates Supreme Court” (1988) 50:2 The Journal of Politics 361 [Walker, Epstein & Dixon].
  • [6] Caldeira & Zorn, “Of Time,” supra note 4 at 875; Paresh Kumar Narayan & Russell Smyth,“The Consensual Norm on the High Court of Australia: 1904-2001” (2005) 26:2 InternationalPolitical Science Review) 147 at 148.
  • [7] Kedar Bhatia, “Final Stat Pack for October Term 2014: Merits Cases by Vote Split” (June 30,2015) SCOTUSblog (blog), online: at 5.
  • [8] Andrew Lynch & George Williams, “The High Court on Constitutional Law: The 2012Statistics” (2013) 36:2 U. New South Wales LJ 514.
  • [9] “Category 4: Appeal Judgements,” Supreme Court of Canada Statistics from 2005 to 2015,online:
  • [10] Alan Paterson, “A Scarcity of Dissents?” (February 19, 2014), UKSC Blog, online:
  • [11] Sunstein, “unanimity”, supra note 3 at 776-778.
  • [12] Andrew Lynch, “Dissent: Towards a Methodology for Measuring Judicial Disagreement inthe High Court of Australia” (2002) 24 Sydney Law Review 470 at 476 [Lynch, “Dissent”]. Lynchnotes that using dissent rates alone is overly simplistic as it neglects concurrences as it is counting concurrences as the same as dissents. He argues it is better to distinguish between dissentsand concurrences.
  • [13] Stacia L. Haynie, “Leadership and Consensus on the U.S. Supreme Court” (1992) 54:4Journal of Politics 1158 at 1160 [Haynie, “Leadership”].
  • [14] Smyth, “Historical Consensual,” supra note 8 at 255, 264.
  • [15] Lynch, “Dissent,” supra note 15 at 490. Moreover, many civil law courts, such as the FrenchCourt of Cassation or the Constitutional Council, release only a single unanimous decision,prohibiting written dissents. Peter McCormick, “ ‘With Respect . . . ’— Levels of Disagreementon the Lamer Court 1990-2000” (2003) 48:1 McGill Law Journal 89 at 94 [McCormick, “WithRespect”].
  • [16] On the US Supreme Court, see Rebecca Gill, “Consensus or Ambivalence? Why CourtTraditions Matter” (8 July 2011) [unpublished] online: For theCanadian Supreme Court, see Emmett Macfarlane, “Consensus and Unanimity at the SupremeCourt of Canada” (2010) 52 Supreme Court Law Review 379 at 382 [Macfarlane, “Consensus”].
  • [17] Lynch, “Dissent,” supra note 15 also focuses on such disagreement due to the nature ofconcurrences in Australia (pp 483ff). Caldeira & Zorn, “Of Time,” supra note 4 explicitly examine the connection between dissents and concurrences through cointegration to see if there isevidence that the same factors influence both modes of disagreement.
  • [18] Smyth, “Historical Consensual,” supra note 8 at 256-257.
  • [19] Russell Smyth, “‘Some Are More Equal than Others’—An Empirical Investigation intothe Voting Behaviour of the Mason Court” (1999) 6:1-2 Canberra Law Review 193 at 201-202[Smyth, “Some Are More Equal”]. See also Donald R. Songer, John Szmer & Susan W. Johnson,“Explaining Dissent on the Supreme Court of Canada” (2011) 44:2 Canadian Journal of PoliticalScience 389 at 390-392 [Songer et al., “Explaining Dissent”] for a discussion of the benefits andcosts of unanimity.
  • [20] Lee Epstein, William M. Landes & Richard A. Posner, Behavior ofFederal Judges: A Theoreticaland Empirical Study of Rational Choice (Cambridge, MA: Harvard University Press, 2013)[Epstein, Landes & Posner, Behavior] at 256. See also Caldeira & Zorn, “Of Time,” supra note4 at 877.
  • [21] See also Songer et al., “Explaining Dissent,” supra note 22 at 393-394 examining the influence of four categories of factors on dissent: attitudinal diversity (including case salience),structural influences (panel size and workload), legal factors (whether the Court reverses andcase complexity) and leadership by the chief justice. There are other factors that may impactconsensus that we do not examine. For example, Walker, Epstein & Dixon, supra note 8 discussthe potential role of the rate of turnover of judges and inexperience in the rise of dissensus onthe uS Supreme Court in the 1940s.
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >

Related topics