A NETWORK ANALYSIS OF THE SINGAPORE COURT OF APPEAL'S CITATIONS TO PRECEDENT

• Author: Jerrold SOH LLB (Hons) (National University of Singapore), BSocSci (Hons) (National University of Singapore); Lecturer of Law, School of Law, Singapore Management University.
• Publication year: 2019
• Citation: (2019) 31 SAcLJ 246
• Published date: 01 December 2019
• Date: 01 December 2019

I. Introduction

1 This article examines trends in the Singapore Court of Appeal's precedent citation behaviour using network analysis, an empirical technique hitherto not applied to Singapore jurisprudence. The main objective is to introduce network analysis and establish both a theoretical framework and an empirical methodology for applying it to Singapore jurisprudence at a larger scale than presently undertaken. More substantively, this article presents findings revealed by applying network analysis to a preliminary dataset of reported Singapore Court of Appeal judgments decided 2000–2017. There are two motivations for this.

2 First, network analysis can provide a deeper empirical look at the Singapore legal system than present methods allow. As Posner notes:1

Scarcity of quantitative scholarship has been a serious shortcoming of legal research, including economic analysis of law. When hypothesis cannot be tested by means of experiments, whether contrived or

natural, and the results assessed rigorously by reference to the conventions of statistical inference, speculation is rampant and knowledge meagre.

As will be explained, network analysis is an emerging technique that precisely allows for hypotheses about the nature and use of legal precedent to be rigorously and scientifically tested. Insights relevant to how the law is practiced and/or taught may also surface.2 For this reason, network analysis is receiving close attention from legal scholars around the world.3

3 Second, a network analysis of citation practices in the Singapore courts is timely. The Singapore legal system has certainly come of age more than 50 years after the nation's independence and almost 200 years after the Second Charter of Justice.4 It is becoming less of a justification (if a justification it ever was) that not enough data on the Singapore legal system exists for meaningful analysis. More importantly, recent work on the Singapore legal system has established strong methodological foundations for empirically studying citation practices in our courts. A seminal work is Goh and Tan's comprehensive study on the development of the Singapore legal system,5 which itself was a culmination of previous empirical work.6 These studies have shed scientific light on how Singapore court judgments have over the years become longer, cited more local authorities7 and increasingly referred to

academic works across all areas of law.8 As Singapore's legal system continues to develop, and in light of its ambition to grow into the region's lex mercatoria,9 it is opportune to build on these solid foundations.

4 Furthering this article's objectives, the complete network of all 987 reported Court of Appeal judgments handed down between the years from 2000 to 2017 will be constructed. The methodology and techniques used will also be described in some detail in order to establish a paradigm which subsequent studies can follow. Four broad questions are proposed to focus the enquiry. First, which judgments are empirically the most central within Singapore's appellate jurisprudence? Second, how has this changed over time, if it has? Third, what implications exist, if any, for the legal sector? Finally, what else can network analysis reveal on the use of authority by the Court of Appeal? These questions are crafted to demonstrate the usefulness of network analysis in uncovering academically and practically interesting insights about the Singapore legal system.

5 The rest of this article proceeds as follows. Part II10 establishes the theoretical foundations for empirically studying legal citations, particularly using network analysis. Part III11 outlines the data used for this article and how it was extracted. Part IV12 presents the results, and Part V13 concludes.

II. Theoretical foundations for citations analysis

A. Theoretical foundations for studying legal citations empirically

6 The empirical analysis of legal citations is well established in legal scholarship. As early as in 1954, Merryman had already conducted an empirical study on the number and nature of judicial citations to precedent and other secondary material by each of the seven (then)

sitting judges on the California Supreme Court.14 He then conducted a follow-up study in 1974 on an expanded set of cases and found, inter alia, that concurring and dissenting judgments tended to have substantially fewer citations, including “some without any citation to authority at all”.15 In 1976, Landes and Posner developed an economic capital model to estimate how the precedential value of a judgment, as proxied for by the number of times it is cited by subsequent courts, depreciated over time in differing legal subject areas.16 Subsequently, in 1993, Landes and Posner used citations analysis to study the influence of economics on law and found evidence that the economic approach had grown, particularly throughout the 1980s, at a rate exceeding that of any other interdisciplinary approach to law.17

7 The theoretical foundation underlying these empirical studies is worth examining. To Merryman, citations analysis went beyond simply revealing “patterns of citation which may be helpful but are not startling”.18 Interpreting a compiled list of secondary material cited by the court and their respective citation frequencies by each judge, Merryman opined that:19

… [s]ome of the works listed, such as Wigmore on Evidence, Williston on Contracts, Scott on Trusts and Paul on Estate and Gift Taxation, are works of high quality prepared by men of established reputation and ability. Many of the others, however, are obvious hack jobs turned out by people nobody has ever heard of except as a name on the back of a book.

8 To Merryman, the data therefore illustrated the “unreflective and uncritical nature of the choices judges [in the California Supreme Court] make among the available works” when citing secondary material. This was a bold qualitative claim about judicial decision-making processes built upon quantitative observation of the effects of such processes. Although citations to precedent are no doubt “a conspicuous feature of most judicial opinions”,20 such a link between quantitative effect and qualitative cause is seldom self-evident, especially

in a system as complex as the law. It is thus apposite to question the premise on which citations analysis relies when making claims about the legal system. This premise is that measurable empirical patterns in legal citations reflect intangible qualitative attributes of the legal system itself. Two conditions are implied.

9 First, there must be measurable empirical patterns in legal citations to begin with. This requires that citations in judgments are not wholly random, for then there is arguably not even “a practice of citation” to study.21 This should not be controversial. Legal citations within common law systems serve a number of important legal purposes. The cited precedent may (a) be binding or persuasive for similarity of fact and/or law; (b) contain one of many competing rules under the judge's consideration; (c) be cited to lend authority to a position the judge wishes to adopt; (d) be cited as a pat on the back to the author; or (e) be cited so it may be questioned, doubted, distinguished or, in rare cases, overruled.22 A citation is therefore more likely to be the product of certain non-random, legally grounded, generative processes rather than that of judicial dice-rolling. It is unlikely that no empirical patterns can be found. In any event, this need not be established a priori; it will be clear whether empirical patterns exist once the data in Part IV has been analysed.

10 Second, even if empirical patterns exist, they must correlate sufficiently to underlying attributes of the legal system in a way that makes extracting qualitative legal insight defensible. Sceptics taking a realist view may argue that citations are motivated primarily by the personal inclinations of the authoring judges. Therefore, the argument goes, empirical patterns in citations behaviour are more informative of judges' personality than any systemic characteristics of the legal system. But this is unlikely because:23

… the extensive research and writing that lawyers, judges, and law clerks devote to discovering, marshalling, enumerating, and explaining precedents are not costless undertakings, and would not be undertaken if precedent did not enter systematically into the decision of cases.

Put another way, there is usually a good legal reason why a precedent is cited.

11 Another common objection to correlating citation statistics with legal insight is that merely counting citations ignores the context in which they were cited. Surely a lukewarm or negative citation should not go towards a precedent's citation count? The answer to this is twofold. First, as will be explained below, citations analysis has evolved beyond merely counting citations. Second, the objection, while true to some extent, does not invalidate citations analysis entirely. It depends on what point is being made. If the aim is only to establish which cases have been the most influential (as opposed to being correct in law or the most “authoritative”), then analysis of citation context ceases to be absolutely necessary. As Landes and Posner note: 24

A common criticism of citation analysis when it is used as an evaluative tool is inapplicable, or largely so, when it is used to study influence: that a critical citation should not be weighted as heavily as a favorable one and maybe should not be counted at all or even given a negative weight. When speaking of influence rather than quality, one has no call to denigrate critical citations. Scholars rarely bother to criticize work that they do not think is or is likely to become influential. They ignore it. Many favorable citations, moreover, are tokens of friendship or obeisance to colleagues, influential seniors, acolytes...