Case Comment Volume 48:3

Measuring Judicial Activism on the Supreme Court of Canada: A Comment on Newfoundland (Treasury Board) v. NAPE

Table of Contents

Measuring Judicial Activism on the Supreme

Court of Canada: A Comment on

Newfoundland (Treasury Board) v. NAPE

Sujit Choudhry and Claire E. Hunter

In the recea Nefoumilud and Labador Cout of Appeal decision of
Nevfoundiid (Treasury Board) v NAPE Justice Marshall accused ti e
Supreme Coul of Clanda of “wAtD incursists .
the elected branches of government … ” He added that these ircursiona were
happeting mom frequetly and invited the Court to “revisit” its interpretation
of section I of the Canadian Charter of Rights and Freedom in R v. Oakes.

into the policy dottain of

White politically

interesting the atthors suggest that the ensuing public
activism and the role of an unelected judiciary with
debate (over judicial
respect io public policy in light of the Charter) missed an importnt point of
engaganet whether or not the enpitical claims on which Justice Marshall
based his comments ae an accurate depiction of the Court’s behaviour when
faced with the possibility of stoking down a “oajoritarian” decision, In
canvassing the quantitative research that is currenty available on th taer, it
becomes apprent that the data is incomplete as it is net tabulated with the
uances of areraalyses in mind.

The aths attempt to fill the void by building on previous stulies. and

specifically by distinguishing their handling of the data with strctee defitions
of applicable data poits. Four hypotheses, drawn fron the clams made in the
NAPEjudagement, are tested, (I) the Supteeme Court stikes down majotarian
legislation often; (2) judicial activism is increasing over time; (3) judicial
activism is largely the product of the Court’s Charter analysis under section 1h
and (4) the Charter’s legislative override under section 33 has been
deligtintisd.

Dana In d/cision Nesfontd/and (Treasury &ard) v. NAPE. renmue
rc/ment par In Cour d’appel de Tetre-Neuve et da Labradr, le juge
Marshall a aecus6 In Cour suptrte du Canada de s’itmuduire indliment das
hcs politiques publiques relevan de In branche &ae da gouvemement. Ayant
remarqu/ une augmentattio dans la fi6quencede teles incusiousiudiciaies
t
a dis Irs invit la Cour t s r
ise, son inteompttation de raticle I de laChane
canadienne des drits et li/ort/s, tetle que developpe dans farrft R. c. Oakes,
ttessant sur le plan politiqta, Ins asesu affirment quo
le d/bat public inspiad par tI critique da juge Marshall (sur ractivisme
juduiaire It In hlmi/r de In Charte) na poulant pas su aboider an point
imporant, As savoir si les 6k/ments empiriques sur lesquels le juge a fonda ses
cotmmtares relpr/sentent sae descraption juste et exacts du compotemea de
In Cone face A [a possibilit d’anuler une dftision .majritaire. Or. 6 I’eiure
aetuelle, ha rechirehe quantitatives en In matifre se c/v/lent incomp/tes. et
ne tienoent pus compte des nances fond6es s ur rtanalyse de In Chare.

Bien qu’tant a

So busant sue =w Oxides arsicures Ina auteurs tensest de combler
coe tcam en adoptuon des d/fittiona plus stactes des variables applicabtes
en leespbe. Q t e hypotses, tie/es da jugnent NAPE, soar sinsi
exanin/es: (I) [a Cour supteme annule soveant les Ikgislations issues de
d6cisions majoritaires ; (2) l’activisme judiciaire s’accroat avec le temps: (3)
l’advisme judiciare ot pnicalemcnt cxereil klrs d I’analyse par In Cour de
I’atticle I de Is Charte; et (4) La claise de dthogation expresse pt/uc h
[‘article 33 de Ia Charte a petdu sa lgitimit&

Ultimately. each of the four hypotheses is cMardicted by the data
the government wins the overwhelming majority of constitutional challenges
broughi to majorit.an dcaisoirs; judicial activism has no inereas over tim
the govetm-sent’s success mnte in tie section I analysis is highly dependant on
whetae or not an intenal limit is imposed on a protected right; and the level of
judicial activism has not increased usa response to the deligttinaization of the
section 33 override.

Whether the judiciary is “unduly” activist or not remains elusive,
however, as there we tuee significant limitations to a study based, such as this
one is, on tde analysis of govemtent “win rates”. At the core, the ambiguous
treatent of what is “apptsspiate” versus “ndue”
itetference makes it
difficult to detemine whether the Court has exceeded the constitutional sc pe
of its powers ‘ae small data pool and the possibility of a selection bias
increase the difficulty of the task at hand. The authors conclude that Justice
Marshall’s normative claims ar based on assutptions that are highly suspect
and that before the debate on the propriety of judicial activism proceeds any
farther, mot quantitative legal research should be conducted to deterine
whether or not the Court is actually activist.

ces quate hy”s

Les analyses ditontrent qu

n sor pas
coofirm6es duns les fbits: le govetnement remposte ua grande majoit des
litiges qui m n
en question des d/cisiona nssjoriiaires str e plan
s’est pas acm avec le temps: le
constitutiontel I’activisme jsliciam:re
succas da govenaement au oiveau de I’analyse de I’article I dpend largement
de rexistence ou trt de liistes internm impos es wur le odit prvd: le
judiaaire na pas augmeni en r6ponse At la perte de
ivea d’activism
tegititit6 d la clause de drogation expicsse pt/vue h ‘arficle 33.

La detemaination du crasere -excessif., de ‘activisme judiciire
den== cepemlant encore difficile As dtablir. 6ant dan/ les limites inhetentes
aux tiisia bande s str one analyse des otiaux de suc/ss, da gouvernement. Au
centre de oes limites r troWve le Iuitement anahig de cc qui est oappap it.
par rpport A oexcessif,,
rendant ainsi difficile d dtenniner si In Cot a
exc/d ou non In porke consttitiotvlle da ses pouvtirs. Le caraetre testreint
de [a base empirque et les possibles distowiona dana silecion des dantes
augmentent par ailleurs con difficaot6, Les autems en arrivent trinmains 6 [a
conuchsion que Is ohommentaires da juge Marshall deaneuent fonds stir des
hypothsts hauteent questionnables, ot que. avant de pourarivre plus loin le
debut, d’autres /tsdes quantilatives deiuicat fte effectis afit da d/uesmirr
si a Con

st effectivemem activis.

“Sujit Choudhry, Assistant Professor, Faculty of Law, University of Toronto; Claire E. Hunter, J.D.
(University of Toronto, 2003). E-mails: sujit.choudhry@utoronto.ca; claire.hunter@utoronto.ca. We
thank Lisa Vatch for invaluable assistance and an anonymous reviewer for excellent suggestions.

Sujit Choudhry & Claire Hunter 2003

To be cited as: (2003) 48 McGill LJ. 525
Mode de rfdrence : (2003) 48 R.D. McGill 525

526

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

Introduction: Normative Critique, Empirical Premises

I. Measuring Judicial Activism

A. A Quantitative Definition of Judicial Activism
B. Empirical Claims on Judicial Activism: Four Hypotheses
C. Contextualizing the Claims

II. Methodology

A. Compiling the Data Set

1. What Is a Charter Case?
2. When Is a Case Counter-Majoritarian?
3. Correcting for Overcounting and Undercounting:

Companion Cases and Multiple Provisions

B. Coding the Judgments
C. Assessing the Hypotheses in Light of the Data

Ill. Results, Analysis, and Discussion

A. Hypothesis 1: Judicial Activism Is High
B. Hypothesis 2: Judicial Activism Is Increasing over Time
C. Hypothesis 3: The Section 1 Analysis Is the Locus of

Activism
1. Overview
2. Variation by Substantive Charter Right
3. Trends over Time

. Hypothesis 4: The Override Has Been Delegitimized

Conclusion

Appendix

527

531
531
534
536

538
538
539
540

541
542
543

544
544
546

548
548
550
552
553

555

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

527

Introduction: Normative Critique, Empirical Premises

“Judicial activism has gone too far” read the headline on the front page of The
Globe and Mail the week after the Newfoundland and Labrador Court of Appeal
handed down its judgment in Newfoundland (Treasury Board) v. Newfoundland
Association of Public Employees’ last December. Justice Marshall, speaking for a
unanimous Court of Appeal, held that provincial legislation denying retroactive pay
equalization to a group of female employees violated section 15 of the Canadian
Charter of Rights and Freedoms but was saved under section 1 because it served a
legitimate deficit-reducing purpose. Notwithstanding its importance, the particular
issue decided by the Court of Appeal (the extent to which governments may depart
from equality norms for fiscal reasons) generated little public interest. Instead, media
attention focussed on a few remarkable paragraphs buried in the middle of the 169
page judgment, in which Justice Marshall accused the Supreme Court of Canada of
“undue
the public domain of the elected branches of
government … ‘”

incursions

… into

Interestingly, Justice Marshall did not rehearse the oft-repeated claim that in
constitutional adjudication, courts must not intrude on the making of public policy,
accepting that “there has always been an element of policy making in judicial
decisions … ‘” He stressed instead that courts and legislatures have different roles with
respect to public policy, with there being “no role in policy making for the judiciary
beyond that consequential to passing upon whether executive and legislative measures
achieved their intended policy through interpretations of their scope.” ‘6 To do
otherwise would contravene the “Separation of Powers Doctrine”.7 Against the
backdrop of this cluster of constitutional ideas, Justice Marshall argued that the
Supreme Court had overstepped the constitutional bounds of its limited policy making
power in its Charter jurisprudence. The focus of his attack was the Supreme Court’s
interpretation of section 1 because of what he viewed as the Court’s failure to be
sufficiently deferential to legislative decisions. Thus, although he conceded that “s. I
effectively invests [the judiciary] with responsibility to pass upon the justifiability of
policy choices behind Charter infringements, “‘ his concern was with the interpretation

Kirk Makin, “Judicial Activism Has Gone Too Far, Court Says” The Globe and Mail (12

December 2002) Al.

2 (2002), 220 Nfld. & P.E.I.R. 1, 221 D.L.R. (4th) 513, 2002 NLCA 72, leave to appeal to S.C.C.

granted, [2003] S.C.C.A. No. 45 (QL) [NAPE].

3 Part I of the Constitution Act, 1982, being Schedule B to the Canada Act 1982 (U.K.), 1982, c. 11

[Charter].

‘NAPE, supra note 2 at para. 364.
‘Ibid. at para. 349.
6Ibid. [footnotes omitted].
‘Ibid.
9Ibid. at para. 362.

528

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

given to section 1 in the Oakes test.’ In particular, the need for the means chosen by
the government to satisfy the requirement of proportionality was found to be deeply
problematic, because it “endows the judiciary with licence to stand in the shoes of the
other branches of government as ultimate arbitrator of which policy choices were in
the best interests of the governed.’ As a consequence, Justice Marshall invited the
Supreme Court to “revisit” the Oakes test,” warning of “real potential for heightening
unease”” if the Court’s activism continues unchecked.

Justice Marshall’s comments on judicial activism, and his call for the Supreme
Court to revise the Oakes test, sparked a rare and remarkable public debate in the
pages of The Globe and Mail. Newfoundland and Labrador Chief Justice Clyde Wells
took the highly unusual step of writing to the newspaper to explain that the judgment
in NAPE reflected “only the opinions and comments of one of the three judges sitting
on that appeal,”‘ notwithstanding the concurrences of the other two members of the
panel in Justice Marshall’s reasons.” Chief Justice Wells’ intervention prompted
columnist Jeffrey Simpson
that the whole episode was “bizarre”,
“extraordinary”, and “unprecedented”, and to speculate on the political motivations
of the players. Weeks later, in another dramatic twist, former federal Justice Minister
John Crosbie lodged a complaint with the Canadian Judicial Council to investigate
Chief Justice Wells’ comments.” This complaint was ultimately rejected, to public
criticism from Crosbie.’ 7

to write

But this fracas was a sideshow to the much more interesting debate on the
substance of Justice Marshall’s comments that took place through letters to the editor.
Some letters criticized Justice Marshall’s suggestion that judicial activism is a bad
thing. One expressed the view that the Supreme Court is necessarily activist because
of legislatures’ unwillingness
issues, observing that “many
legislators are probably more than happy to let the [Clourt take the heat.”” Another

to tackle difficult

‘Ibid. at paras. 354-68. See R. v. Oakes, [1986] 1 S.C.R. 103; 26 D.L.R. (4th) 200 [Oakes].
“NAPE, ibid. at para. 362.

Ibid. at para. 365.
2 Ibid. at para. 364.
Kirk Makin, “Wells Rejects Interpretation of Ruling” The Globe and Mail (13 December 2002)

Al atA9.
14 Ibid. See NAPE, supra note 2 at paras. 641-42.
” Jeffrey Simpson, “A Little Heresy at the Newfoundland Courthouse” The Globe and Mail (17
December 2002) A25. Simpson noted that Wells was Liberal premier when the pay equity payments
were rejected, and that Justice Marshall was a former Tory cabinet minister in the governments of
Frank Moores and Brian Peckford.

6John Crosbie, “Out of Line” The Globe and Mail (7 January 2003) A13.
Chantelle MacDonald, “CJC Rejects Complaint over Chief Justice’s Letter to the Globe” Lawyers
Weekly (4 April 2003) (Lexis); Christin Schmitz, “Former Justice Minister Slams CJC’s ‘Judicial
Whitewash’ Lawyers Weekly (16 May 2003) (Lexis).

” Alyn James Johnson, Letter to the Editor, The Globe and Mail (14 December 2002) A22.

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

529

echoed this view, arguing that “[r]ather than stepping back to allow governments
broader powers, our courts should continue to be proactive to ensure that governments
don’t abuse the power they already possess.”” Other letters applauded Justice Marshall
for speaking up “on the train wreck that is the increasing tendency toward judicial
activism by the Supreme Court:’ arguing that the Court is in fact “encroaching on the
prerogatives of the legislative branch.'”

As well, several letters made the link, not explicit in Justice Marshall’s reasons,
between judicial activism and the Charter’s notwithstanding clause, section 33.2″ One
contributor suggested that judicial activism is not a problem because of the
availability of section 33, since “[iut should not be forgotten that the notwithstanding
clause-the much hated section 33-of the Charter permits governments virtually
unrestricted power provided they are willing to revisit their decisions every five years
and face the electorate’:” Another contributor lamented what he viewed as the
underuse of the notwithstanding clause, writing that “concern about the Oakes test,
and judicial activism in general, would be minimized if our politicians had the will to
make use of a tool readily available to them-the notwithstanding clause. Such action
would, in many instances, put a stop to such odious judicial incursions.”

Although this debate attempted to engage Justice Marshall’s reasons, it missed an
important point of engagement. As will be discussed below, the truly interesting
feature of Justice Marshall’s reasons is that his normative critique of the Supreme
Court and its jurisprudence rely on empirical claims-for example, that the levels of
judicial activism on the Court are high and have been increasing over time–claims
that he did not substantiate with quantitative evidence. Commentators have focussed
on a determination of whether high levels of judicial activism are a good or a bad
thing, but never questioned whether high levels of judicial activism are real or
fictional phenomena. To be sure, these empirical claims were alluded to in one
editorial, which suggested that Justice Marshall was wrong to say that the Supreme
Court has been systematically activist in its interpretation of section 1, noting that the
Court “has waxed and waned, sometimes deferring to Parliament and sometimes
striking down laws.”‘2 But this editorial was the exception, not the rule.

‘9 Scott Nelson, Letter to the Editor, The Globe and Mail (13 December 2002) A24.
‘0 Jack Tate, Letter to the Editor, The Globe and Mail (13 December 2002) A24.
21 Charter, supra note 3, s. 33. Subsection 33(1) reads:

Parliament or the legislature of a province may expressly declare in an Act of
Parliament or of the legislature, as the case may be, that the Act or a provision thereof
shall operate notwithstanding a provision included in section 2 or section 7 to 15 of this
Charter.

22Nelson, supra note 19.
23Tate, supra note 20.

“Judge Marshall’s Injudicious Tirade”, Editorial, The Globe and Mail (14 December 2002) A22.

530

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

Regrettably, the use of empirical assumptions to bolster normative claims in the
absence of quantitative evidence is reflected not only in public debate, but also in the
work of prominent political scientists who write about the Supreme Court and the
Charter. For instance, there is a common, but for the most part unsubstantiated, view
among critics of the Court that judicial activism is high in absolute terms, namely that
the Court intervenes frequently in political decisions. ‘ Political scientist Christopher
Manfredi, who writes often about the Supreme Court, has made the additional claim
that the level of judicial activism on the Supreme Court has increased over time as a
consequence of the delegitimization of the legislative override,26 but has not offered
adequate evidence to substantiate his assertion. As evidence, critics of the Supreme
Court rely on anecdotes, referring to highly visible and controversial examples where
the Court has struck down legislation. Justice Marshall does just that by referring to
the highly emotive case of RJR-Macdonald v. Canada (A.G.),” where a deeply
divided Supreme Court struck down a prohibition on tobacco advertising.

It is one of the great shortcomings of Canadian legal scholarship that the legal
academy has largely failed to engage with political scientists who make empirical
claims about judicial activism. This is likely the case because of a mistaken and dated
view that political science offers little insight into the study of the courts and
constitutional adjudication. But the instances where these empirical assumptions have
been adopted by the courts, exemplified by Justice Marshall’s reasons in NAPE, now
make it urgent for the legal academy to assess the accuracy of these claims. To be
sure, the use of qualitative evidence, such as the thick narrative descriptions found in
Kent Roach’s pioneering work, The Supreme Court on Trial,” is one way to respond.
But detailed case studies should be supplemented where possible with quantitative
evidence of a systematic nature to give a more comprehensive picture of the actual
practice of Canadian constitutionalism. Anecdote will no longer do. For descriptive
statistics to be helpful, the design of a quantitative analysis of the Supreme Court’s
Charter jurisprudence must be sensitive to a doctrinal understanding of how Charter
analysis is done. Thus, rather than simply addressing the distribution of wins and
losses-the dominant concern of the political science literature-quantitative analysis
must be informed by the structure of Charter adjudication. For example, as we
explain below, the failure of previous studies to take into account that some Charter
rights contain internal limitations has led to simplistic and misleading conclusions
regarding the Court’s behaviour under section 1.

2 See e.g. FL. Morton & Rainer Knopff, eds., The Charter Revolution and the Court Party

(Peterborough, Ont.: Broadview Press, 2000) at 13.

2’6 Christopher P. Manfredi, Judicial Power and the Charter: Canada and the Paradox of Liberal

Constitutionalism (Oxford: Oxford University Press, 2001) at 5.

27 [1995] 3 S.C.R. 199, 111 D.L.R. (4th) 385 [RJR-MacDonald].
26 Kent Roach, The Supreme Court on Trial: Judicial Activism or Democratic Dialogue (Toronto:

Irwin Law, 2001).

2003] S. CHOUDHRY & C.E HUNTER- MEASURING JUDICIAL ACTIVISM

531

In this article, we take up this task. We begin, in Part I, by offering a quantitative
definition of judicial activism, by generating four hypotheses that arise from the
empirical claims made by Justice Marshall and other critics of the Supreme Court,
and by contextualizing these claims in the legal literature. These hypotheses are: (1)
the Supreme Court second-guesses “majoritarian” decisions frequently, (2) the Court
does so with increasing frequency, (3) section 1 is the central vehicle whereby the
Court exercises its counter-majoritarian power, and (4) the level of judicial activism
has risen in response to the delegitimization of the override. In Part II, we address
questions of methodology-the creation of our data set, our coding rules, and the
manner in which we assess the hypotheses in light of this data-and compare our
methods with those of previous studies. In Part 1H, we offer our results, analysis, and
discussion. We find that the data contradicts each of the four hypotheses. For instance,
the government wins the overwhelming majority of constitutional challenges brought
to “majoritarian” decisions and judicial activism has not increased over time. We also
show that although the government loss rate at section 1 is higher than its loss rate
overall, this can be explained on further analysis as a function of whether or not the
right contains an internal limit. Finally, our data demonstrates that the level of judicial
activism did not increase as a response to the delegitimization of the override. The
limitations of our study are discussed with our conclusions and our data set is
reproduced in the appendix.

I. Measuring Judicial Activism

A. A Quantitative Definition of Judicial Activism
In assessing the empirical assumptions that underlie the arguments made by
Justice Marshall and other critics of judicial activism, we must first define “judicial
activism”. Unfortunately, judicial activism is a notoriously slippery term, which
variously means the departure from well-established precedent, adjudication based on
judicial preferences, or the judicial reallocation of institutional roles between the
courts and other branches of government, depending on who is employing it and in
what context.” Nevertheless, Justice Marshall’s comments give us a starting point
from which to approach this ambiguity, at least for the limited purpose of framing a
response to his judgment. Justice Marshall suggests that we should be concerned
about “undue incursions by the judiciary into the policy domain of the elected
branches of government.”” In other words, the reallocation of institutional roles, or the
usurpation of legislative power by courts, seems to exist when judges second-guess

29 See Sujit Choudhry, Book Review of Judicial Power and the Charter: Canada and the Paradox of

Liberal Constitutionalism by Christopher P. Manfredi (2003) 1 International Journal of Constitutional
Law 379 at 386 [Choudhry, Book Review].

” NAPE, supra note 2 at para. 364.

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

democratically made decisions in a manner that is “undue”. Unfortunately, Justice
Marshall’s judgment does not yield any criteria by which we can identify when the
Supreme Court has acted in an “undue” fashion.

In this article, we rely on a quantitative definition of judicial activism. This
measure focusses on outcomes (i.e., whether a government wins or loses) and posits
that courts are more activist the more frequently they find that democratically elected
institutions have acted unconstitutionally. There is support for this approach in the
existing literature. In a study on the Supreme Court’s first one hundred Charter
judgments, F.L. Morton, Peter Russell, and Michael Withey referred to the low
government success rate in the initial phase of Charter adjudication as “the Court’s
initial activism”,” and the declining government loss rate as a “decline in judicial
activism”.” James Kelly, who updated this original study through to 1997, also
worked with a quantitative definition of activism and equated increases and decreases
in the government loss rate with parallel fluctuations in judicial activism.”

We recognize that a quantitative definition of judicial activism poses a number of
problems. First, Justice Marshall refers to “undue” incursions by the Court, which
suggests that according to a quantitative definition of judicial activism, the Supreme
Court is unduly interventionist when it passes some threshold beyond which the
government loss rate is too high. Unfortunately, Justice Marshall does not tell us what
that rate might be. Thus, determining whether the rate of government wins or losses is
too high is much like assessing whether the rates of unemployment or inflation are too
high: it is a matter of highly contextualized judgment, that may change with political
circumstances. However, the absence of an absolute threshold does not undermine our
ability to measure relative activism over time. Moreover, to the extent that critics of
the Court employ a quantitative definition of judicial activism, it is for them to explain
and justify what the threshold is.

Second, a quantitative definition of activism takes as its baseline of “non-activist”
behaviour a government win rate of one hundred percent, a situation where the
Supreme Court never second-guesses decisions made by democratically elected
institutions. The difficulty this poses is that the very idea of constitutional supremacy,
coupled with judicial review, by necessity permits, or even requires, courts to impose
some constitutional limits on the scope of government decision making by setting
constraints on both the ends and means of governance. Thus, a fuller conception of
activism would incorporate some qualitative criteria to sort out those judicial
incursions that are or are not proper, and then would assess government win and loss

” F.L. Morton, Peter H. Russell & Michael J. Withey, ‘The Supreme Court’s First One Hundred

Charter of Rights Decisions: A Statistical Analysis” (1992) 30 Osgoode Hall L.J. I at 14.

Ibid. at 10.

‘ James B. Kelly, “The Charter of Rights and Freedonms and the Rebalancing of Liberal
(1999) 37 Osgoode Hall L.J. 625 at 629 [Kelly,

Constitutionalism in Canada, 1982-1997”
“Rebalancing”l.

2003] S. CHOUDHRY & C.E HUNTER- MEASURING JUDICIAL ACTIVISM

533

rates only with respect to the latter cases to determine the degree of judicial activism.
As a result, a proper definition of the rate of activism would be as follows: the
percentage of cases
that governments lose, less the percentage of cases that
governments should lose. Justice Marshall admits as much, as it seems that he is not
concerned with all incursions into the domain of legislatures, but only with those that
are “undue”. Unfortunately, it is difficult to identify an objective set of criteria that can
tell us which incursions are proper and which are “undue”. As a result, for the limited
purposes of this article, our working definition of judicial activism is one that
encompasses all second-guessing of legislatures by courts, understanding that there
will always be some portion of this activism that is not actually “undue”.

Third, a strictly quantitative measure of outcomes treats all decisions as having
equal importance, when in fact each decision’s degree of importance may vary. As
F.L. Morton, Peter Russell, and Troy Riddell point out in their discussion of the
usefulness of empirical studies, “the reasons given to justify a decision are often more
important in the long run than a decision’s basic outcome or
‘bottom line.””
Similarly, comments are often made that are not strictly necessary in order to come to
a decision in the particular matter but that may have great effect in lower courts. The
importance of these comments cannot be measured through a crude statistical
measure. In creating a database of Supreme Court decisions, it is easy for the
importance of any one particular judgment to be overstated or understated. For
instance, an “as of right” criminal appeal may not raise any new issues, while another
case heard the same day creates a new test that redefines the application of a particular
Charter section. The two cases are clearly not of equal importance, but each becomes
merely one data point for the purposes of an empirical study.

Fourth, quantifying Charter cases as wins and losses occasionally has the effect
of emphasizing form over substance. The clearest example of this problem can be
found in comparing the coding of R. v. Butler” and R. v. Sharpe.6 Both of these cases
involved the prohibition of obscene materials by provisions of the Criminal Code that
on their face were unconstitutionally overbroad (i.e., the provisions prohibited not
only speech that was harmful but also speech that was harmless). However, despite
the fact that the provisions at issue were defective in the same way, Butler is classified
as a government win, and Sharpe as a government loss. In Butler, the Court read
down the provision prior to analyzing its constitutionality to limit its scope to harmful
speech, and thus concluded that the provision passed muster under section 1.
Conversely, in Sharpe, although the provision was read down in certain respects
(thereby reducing potential sources of unconstitutionality), in other respects it was
not. In the result, the Court found the provision to be an unjustifiable restriction on

” F.L. Morton, Peter H. Russell & Troy Riddell, “The Canadian Charter of Rights and Freedoms:

A Descriptive Analysis of the First Decade, 1982-1992” (1995) 5 N.J.C.L I at 2.

35R. v. Butler, [199211 S.C.R. 452, 89 D.L.R. (4th) 449 [Butler].
36R. v. Sharpe, [200111 S.C.R. 45, 194 D.L.R. (4th) 1 [Sharpe].

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

harmless freedom of expression. While there is no principled way to distinguish the
results, because of the form of each decision, their outcomes will appear differently in
our data set.

In short, we recognize that it is important to not fall into the reductionist trap, that
is, to privilege variables that can be measured with relative ease from the entire set of
Charter cases (e.g., wins and losses) without attention to the specific features of
constitutional adjudication under specific provisions, let alone the individual details of
each constitutional challenge that comes before the Court. Just as qualitative evidence
on its own can provide only an incomplete understanding of Canadian constitutional
practice, so too is quantitative evidence limited in its ability to shed light on the
Court’s constitutional jurisprudence. But notwithstanding the limitations of a
quantitative approach to measuring judicial activism, our method of analysis has an
important place in the study of judicial behaviour. In general, “[diescriptive statistics
provide a factual foundation on which other studies can build, qualify and elaborate.”‘
To the extent that normative claims are implicitly grounded in empirical claims,
empirical work is critical to the process of evaluating these normative claims. Without
empirical analysis, the factual support for the claims made in the judicial activism
debate cannot be verified.

B. Empirical Claims on Judicial Activism: Four Hypotheses
We have identified four hypotheses from the empirical assumptions that are either
implicit or explicit in Justice Marshall’s comments and from the political science
literature to be tested against the evidence.

Hypothesis 1: The first claim is that counter-majoritarianism is a widespread and
central part of Canadian constitutional practice. Justice Marshall alluded to this claim
in his reasons:

The seeds of this potential are already evident in the unease that has
frequently been expressed over undue incursions into the public policy field in
Charter applications. Despite protestations to the contrary,
it has to be
acknowledged there is an air of legitimacy to many of these complaints.

Simply stated, the hypothesis is that the Supreme Court strikes down legislation, and
does so frequently, making judicial activism high in absolute terms (with the caveats
we described regarding this point earlier).”

Hypothesis 2: The second claim is that judicial activism is increasing over time.
This second hypothesis is made explicit in the public debate that followed NAPE* and

37 Morton, Russell & Riddell, supra note 34 at 2.
“NAPE, supra note 2 at para. 365.
“See text accompanying notes 31-36.
,” See Tate, supra note 20.

2003] S. CHOUDHRY& CE. HUNTER- MEASURING JUDICIAL ACTIVISM

535

is implicit in Justice Marshall’s comments. It is also explicit in much of the political
science literature.”

Hypothesis 3: The third claim, which Justice Marshall made explicitly, is that
there is a particular problem of activism with respect to the Supreme Court’s use of
section 1. He argued that the Court is not sufficiently deferential to legislatures in
cases where there is a violation of a Charter right that the government attempts to
justify under section 1. This appears to encompass both absolute and relative claims.
The first part of the claim is that the Court fails to accept a sufficiently high number of
government section 1 justifications. The second part of the claim is that the Court has
become less likely to accept section 1 defences over time, particularly since its
decision in RJR-MacDonald, in which the majority ruling is singled out for special
criticism by Justice Marshall.

Hypothesis 4: The fourth claim arises out of Manfredi’s analysis of the impact of
the delegitimization of the override on Charter adjudication and requires elaboration
for our purposes here. The override permits Parliament or the provincial legislatures
to enact legislation that would otherwise contravene the Charter and applies to a
limited set of Charter rights such as fundamental freedoms, legal rights, and equality
rights.’ Declarations under the override operate for five years. When the Charter was
first enacted, defenders of the override suggested that it offered an alternative to
judicial supremacy, by allowing legislatures to respond to erroneous court judgments
without resorting to more drastic measures that might undermine the legitimacy of the
Court. One example of a measure the override was designed to avoid is the “court
packing” plan proposed by President Roosevelt in the 1930s in response to the
invalidation of New Deal legislation by the U.S. Supreme Court.” However, the
override has hardly been used. 5 Manfredi argues that the override has fallen into
disuse because it was delegitimized by its use on one very visible occasion: the
enactment of Bill 178,6 a Quebec law requiring that exterior commercial signs be
solely in French. Bill 178 generated a storm of controversy in English Canada, and
political scientists argue that since then, it has been exceedingly difficult politically for
the override to be invoked by a government.

The delegitimization of the override is a familiar story. Manfredi’s twist is his
claim that the delegitimization of the override has affected Charter adjudication, by

“See Manfredi, supra note 26.
42 NAPE, supra note 2 at paras. 275, 341-68.
43 Charter, supra note 3, s. 33.
“Sujit Choudhry, “The Lochner Era and Comparative Constitutionalism” International Journal of

Constitutional Law [forthcoming in 2004].

,5Tsvi Kahana, “The Notwithstanding Mechanisms and Public Discussion: Lessons from the

Ignored Practice of Section 33 of the Charter” (2001) 44 Canadian Public Administration 255.

, Bill 178, An Act to amend the Charter of the French Language, 2d Sess., 33d Leg., Quebec, 1988,

cl. 10 (assented to 22 December 1988), S.Q. 1988, c. 54.

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

making the Supreme Court more activist. His premise is that the Supreme Court acts
strategically because it has policy objectives, and its ability to secure those objectives
depends on the responses of legislatures to its judgments. Manfredi argues that prior
to Bill 178, the threat of override led the Supreme Court to be cautious in its
judgments
in order not to provoke a legislative reversal of its rulings. The
delegitimization of the override, accordingly, has altered the institutional balance
between courts and legislatures by removing this external restraint on judicial review.
The result, Manfredi argues, is that unconstrained by the threat of the override in the
post-Bill 178 period, courts have acted more aggressively, and judicial activism has
consequently increased. 7

Justice Marshall does not mention the override in NAPE, although it figures
prominently in the debate sparked by his reasons.”8 A connection, however, may be
drawn between that judgment and this part of Manfredi’s argument for the following
reason: Where there is a real possibility of an effective legislative reply to the
Supreme Court’s judgments, the wind comes out of the sails of those who raise
normative objections to judicial activism. Justice Marshall’s strong claims about the
potential for deep public dissatisfaction with the Court’s activism, without any
mention whatsoever of the override, seem to depend on the belief that despite the
presence of section 33 in the Charter, there is no real opportunity for legislative
response. The final hypothesis, then, is that the override was delegitimized by its use
in Bill 178 in 1988, and that its disuse has contributed to increased levels of activism
by an unconstrained Court in the period that has followed.

C. Contextualizing the Claims
These hypotheses, implicit and explicit in Justice Marshall’s judgment, are
particularly interesting because they run counter to the dominant narrative in the legal
academy. The dominant view among legal academics is that after an initial brief
adoption of an aggressive approach to Charter interpretation, the Court has not been
activist at all, but has in fact become rather deferential. Moreover, the site of deference
is section 1, the very same provision that Justice Marshall identifies as the instrument
of the Court’s activism. Kent Roach, for example, expresses this view when he writes
that “[ilt is well known that, after an initial bout of enthusiasm, the Supreme Court
has become more deferential in determining whether the government has justified
restrictions on Charter rights.” ‘ Indeed, Roach argues that deference has increased to
such an extent that “the Court will often uphold legislation if the government does a
half-way decent job of mounting a section 1 defence” ‘ The Court’s evolving

‘ Manfredi, supra note 26 at 4-5, 184-88.
49 See supra notes 22, 23 and accompanying text.

Roach, supra note 28 at 172.

50 Ibid.

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

537

approach to section 1 is a far cry from the signal sent by Oakes, which granted to
rights a presumptive importance and regarded limitations on rights as exceptional and
narrow.

The rise of deference under section 1 can be linked to a variety of doctrines that,
taken together, have significantly diluted the stringency of the Oakes test. For
example, the Court defers to governments when legislation has been enacted to
protect vulnerable persons, such as retail workers and children.” The Court also defers
to legislation that balances conflicting societal interests,” as opposed to legislation that
pits the state against the individual (so-called singular antagonist legislation).” On the
issue of what kinds of evidence are sufficient to satisfy section 1, the Court has held
that when legislation is enacted against the backdrop of conflicting social science
legislation, which does not conclusively establish that constitutionally-protected
conduct is harmful, it is enough for the government to show that it had a rational basis
for proceeding.’ And in some cases, such as those involving pornography and hate
speech, a “reasonable apprehension of harm” through the taking of judicial notice by
the Court drawing on its own experiences is sufficient.”

As well, many legal scholars believe that even a highly counter-majoritarian
Supreme Court would not be a problem because the override provides a legislative
check to judicial review. Several legal scholars have characterized the relationship
between courts and legislatures as one of dialogue whereby legislatures may respond
to Court judgments by amending legislation that has been held unconstitutional in
order to bring it in line with constitutional requirements. They have argued that the
site of this sort of conversation is in the section 1 analysis.” Moreover, if courts and
legislatures ultimately fail
to resolve differences over the constitutionality of
legislation through dialogue at the section 1 stage, the override gives the legislature
the last word. Clearly, if the override has been delegitimized, it cannot serve this
function.

Finally, some legal scholars have also considered the impact of the override on
constitutional adjudication. From the vantage point of constitutional design, the
availability of the override has been offered not as a reason for the Supreme Court to

” R. v. Edwards Books, [1986] 2 S.C.R. 713, 35 D.L.R. (4th) 1; Irwin Toy. Ltd. v. Quebec (A.G.),

[1989] 1 S.C.R. 927,58 D.L.R. (4th) 577 [Irwin Toy cited to S.C.R.].

“McKinney v. University of Guelph, [1990] 3 S.C.R. 229, 76 D.L.R. (4th) 545.
,3 This distinction was first set out in Irwin Toy, supra note 51 at 994. However, see McLachlin J’s

1

judgment in RJR-MacDonald (supra note 27 at para. 135) for skepticism about this distinction.

4 Irwin Toy, ibid.
” Thomson Newspapers Ltd. v. Canada (A.G.), [1998] 1 S.C.R. 877 at paras. 115-17; 159 D.L.R.

(4th) 385.

56 See e.g. Peter W. Hogg & Allison A. Bushell, “The Charter Dialogue between Courts and
Legislatures: (Or Perhaps the Charter of Rights Isn’t Such a Bad Thing After All)” (1997) 35
Osgoode Hall LJ. 75 at 79-82, 84-87, 92-93 ; Roach, supra note 28 at 176.

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

act cautiously, but as a reason for the Court to act aggressively. As Kent Roach
recently suggested, the idea is that eliminating judicial supremacy permits courts to
operate free of the anxiety that they have the final say.7 On this view, it also follows
that courts should act cautiously in the absence of an override. Manfredi essentially
turns these propositions on their heads by arguing that an override prompts deference,
and the absence of an override prompts activism. In other words, Manfredi (and
perhaps Justice Marshall) claims that the override has had exactly the opposite effect
on judicial review than the effect anticipated by the legal academy. Determining
whether either story is accurate is important to properly understanding Canadian
constitutional practice.

II. Methodology

A. Compiling the Data Set

In order to test the claims made in the NAPE decision, we generated a data set of
all Supreme Court decisions from 1982 to 2002 in which the constitutionality of a
“majoritarian” act was challenged on the basis that it violated a Charter right. The list
of cases (reproduced in the appendix, below) was generated from three sources. First,
for the period from 1982 to 1997, we relied on the lists generated in a study by
Morton, Russell, and Riddell that examined the period from 1982 to 1989,’8 and in a
second study by James Kelly that extended the former group’s analysis through to the
end of 1997″‘ (we will refer to these studies collectively as the “Morton/Kelly work”).
Second, we updated this list based on Supreme Court Reporters published after 1997.
Third, in order to extend the list of cases to those not yet published in the Supreme
Court Reporters at the time our data was collected, we obtained the neutral citations
for all unreported judgments from the Supreme Court Website and performed full-text
searches of those judgments in Quicklaw to ascertain in which cases there was a
Charter right at issue. ‘

We build upon and update the Morton/Kelly work, not only in generating our set
of cases, but more fundamentally, in relying on the outcomes of constitutional
challenges as a quantitative measure of activism. It is important, however, to make
clear that both the hypotheses tested and the methodology used in this study make it
distinct from previous work. Specifically, our study differs from the Morton/Kelly
work in three significant ways. First, although the Morton/Kelly work purports to be
about “Charter decisions”, its definition of Charter decisions encompasses not only
Charter rights, but also Aboriginal rights and denominational school rights. In our

Roach, ibid. at 6.

.Morton, Russell & Ridell, supra note 34 at 55-58.

Kelly, “Rebalancing”, supra note 33.

8o This set of cases extends up to 31 December 2002.

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

539

study, we include only true Charter claims, excluding other types of rights claims.
Second, the Morton/Kelly work examines both challenges to individual executive acts
(primarily decisions taken by the police) as well as challenges to legislation, both
primary and secondary. As will be explained in greater detail below, we include only
cases in which primary legislation is challenged because our definition of activism is
rooted in counter-majoritarianism. Third, in order to take into consideration the
interaction of substantive Charter rights with sections 1 and 33 in Charter challenges
to majoritarian decisions, we draw further distinctions between different categories of
Charter rights. Thus, when we assess claims about the delegitimization of the
override, we break down the data set to distinguish between overrideable and non-
overrideable Charter rights, to see if there are any differences in judicial behaviour.
We also pay close attention to the stage at which Charter claims are resolved (i.e., the
rights-infringement stage versus the section 1 analysis). To better understand this
interaction, we undertake an analysis of the patterns of wins and losses, depending on
whether the substantive Charter right has been interpreted as containing an internal
limitation.

In sum, our study can be distinguished from previous work on the basis that it is
rooted in a legal and doctrinal understanding that reflects the methodology of Charter
analysis. In particular, our study emphasizes the analytical differences between
Charter rights themselves and of the different relationships that these rights have with
sections 1 and 33. The similarities and differences between our study and the
Morton/Kelly work will be noted throughout our discussion of methodology and
results.

1. What is a Charter Case?

We included only Charter cases in our data set. This is an important way in which
our data set can be distinguished from the Morton/Kelly work, which we relied on to
generate the data points on our list prior to 1998. The two studies included Aboriginal
rights” and denominational school rights’2 in their data sets.’ We limited our data set
in this way for two reasons. First, the claim that the Court is counter-majoritarian is
premised in large part on the view that the Court is not sufficiently deferential in its
interpretation of section 1. Section 1 applies only to the rights and freedoms set out in
the Charter. It does not extend to section 35 or section 93 rights.” This begs the

6′ Constitution Act, 1982, s. 35, being Schedule B to the Canada Act 1982 (U.K.), 1982, c. 11.
62 Constitution Act, 1867 (U.K.), 30 & 31 Vict., c. 3, s. 93, reprinted in R.S.C. 1985, App. II, No. 5.
63 Morton, Russell & Riddell, supra note 34 at 55-58; Kelly, “Rebalancing”, supra note 33 at 627, n. 7.
64Charter, supra note 3, s. 1. Section I reads:

The Canadian Charter of Rights and Freedons guarantees the rights and freedoms
set out in it subject only to such reasonable limits prescribed by law as can be
demonstrably justified in a free and democratic society.

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

question, though, of whether there are proxies for section 1 that operate with respect
to Aboriginal and denominational school rights. Denominational school rights are not
subject to this sort of limitations analysis. But despite the absence of a limitations
clause for section 35, the Court’s test for Aboriginal rights incorporates a limitations
analysis that is extremely similar to the analysis applied in Charter cases.’ We
nonetheless excluded Aboriginal rights cases for a second reason: Justice Marshall’s
judgment, the debate it sparked, and the relevant academic literature all raise concerns
of judicial activism principally with respect to the Charter.

2. When is a Case Counter-Majoritarian?

Because we are measuring the extent to which the Supreme Court is counter-
majoritarian, we only included cases where a majoritarian act is challenged. Our data
set consists of three categories of cases. First, we included all challenges to federal
legislation. Second, we included all challenges to provincial legislation. Third, we
included challenges
to municipal by-laws, because they have been enacted by
democratically accountable institutions, even though they are not a form of primary
legislation.

Our definition of majoritarian decisions resulted in the exclusion of certain
categories of cases. First, we excluded Charter challenges to common law rules.
Common law rules are rules made by unelected judges, and as a result, overruling
such a rule is not counter-majoritarian.

Second, we excluded challenges to secondary legislation, that is, regulations
enacted by government departments, and by-laws or regulations enacted by
independent agencies that exercise powers delegated to them by statute. The rationale
for excluding this category of cases is that, although the enactment of regulations is
contemplated by legislation, their content has not been approved by a democratically
elected body. In some cases, the regulations are enacted by independent bodies, such
as the governing bodies of self-regulating professions. ‘ It seems clear that this
category of cases should be excluded: overruling a decision made by the governing
body of a self-regulating profession can hardly be taken to be counter-majoritarian. In
other cases, however, the bodies that enact the regulations are indirectly accountable
to the legislature. For instance, in cases where a regulation is enacted by a government

63R. v. Sparrow, [1990] 1 S.C.R. 1075, 70 D.L.R. (4th) 385.

Three data points involving challenges to a regulation of this type were excluded from the study:
Black v. Law Society of Alberta, [1989] 1 S.C.R. 591, 58 D.L.R. (4th) 317 (the challenge to Rules of
the Law Society ofAlberta, ss. 75B, 154); R. v. Amway Corp., [1989] 1 S.C.R. 21, 56 D.L.R. (4th) 309
(the challenge to Federal Court Rules, C.R.C., c. 663, s. 465(1) (1978)); Rocket v Royal College of
Dental Surgeons of Ontario, [1990] 2 S.C.R. 232, 71 D.L.R. (4th) 68 (the challenge to R.R.O. 1980,
Reg. 447, ss. 37(39), 40).

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

541

ministry, the minister is ultimately responsible to Parliament. 7 We recognize that the
exclusion of the latter category of cases is more controversial. While an argument
could be made that cases involving challenges to regulations enacted by government
bodies should be included, we have decided to exclude these cases for the additional
reason that the exclusion of both types of secondary legislation promotes replicability
and inter-coder reliability (i.e., the degree to which different observers draw the same
conclusions in the same circumstances). The exclusion of secondary legislation
distinguishes our data set from the Morton/Kelly work, which includes both
legislation and regulations in its consideration of statutes.”

Third, we excluded Charter challenges to actions by government officials,
whether elected or not. While we have eliminated these cases as a group, some
eliminations are more controversial than others. On the one hand, the elimination of
cases that challenge the constitutionality of individual decisions made by government
employees, and that are not framed as challenges to primary or secondary legislation
pursuant to which they act, should be relatively uncontroversial. For example, the
individual conduct of a police officer or an immigration officer, when not directed to
act in that way by primary legislation, is not a majoritarian action. More controversial
is our exclusion of ministerial action from the data set. The rationale for excluding
these cases, however, is the same as the rationale for excluding the cases involving
lower-level government employees: the decision of any one actor on behalf of the
government does not constitute a majoritarian action unless this actor has secured the
positive consent of the legislature.

3. Correcting for Overcounting and Undercounting: Companion Cases

and Multiple Provisions

We made two further changes to our list of cases, distinguishing it again from the
data set used by the Morton/Kelly work.’ First, we eliminated companion cases,
which we define as cases where Charter challenges are brought to the same provision
in multiple cases reaching the Supreme Court at the same time, and where the
judgments are issued together. Typically these cases are found in the criminal context
where multiple defendants in separate actions challenge the provision of the Criminal
Code’ under which they are charged. For example, in R. v. Furtney,” the defendant

6′ Two data points involving challenges to a regulation of this type were excluded from the study: R.
v. Campbell, [1997] 3 S.C.R. 3,206 A.R. 1 (the challenge to Provincial Court Judges Act: Payment to
Provincial Judges Amendment Regulation, Alta Reg. 116/94); Reference Re Renumeration of Judges
of Prince Edward Island, [ 1997) 3 S.C.R. 3, 156 Nfld. & P.E.I.R 1 (the challenge to Public Sector Pay
Reduction Act: Interpretation Regulations, EE.I. Reg. EC 1994-631).

” See Kelly, “Rebalancing”, supra note 33.
9 See Morton, Russell & Riddell, supra note 34; Kelly, ibid.
70 R.S.C. 1985, c. C-46.
” [1991] 3 S.C.R. 89,51 O.A.C. 299 [Furtney].

542

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

challenged the constitutionality of a lottery provision of the Criminal Code’ on the
basis that it violated section 11 (g) of the Charter. In a second case heard at the same
time,” the defendant challenged the same provision on the same basis. The two
judgments were handed down on the same day, and in its reasons on the issue of the
constitutionality of the provision in the companion case the Court simply referred to
the holding in Furtney. Counting both of these cases would essentially be double-
counting the Court’s holding in Furtney. As a result, the companion case was
excluded from the data set.

Just as it is an accident of litigation when the same issue arises in multiple cases,
often multiple issues arise in the same case, each of which could have given rise to a
separate constitutional challenge. Where multiple unconnected provisions of a statute,
or provisions in different statutes, are challenged in the same case, it is under-counting
to treat that case as a single constitutional challenge. In addition, in some cases the
Court arrives at a mixed result, upholding the constitutionality of some impugned
provisions while striking others down. In these cases, it is not possible to treat
constitutional challenges to multiple provisions as a single challenge. As a result,
where the Court considered multiple provisions separately in the judgment, we
counted these as separate constitutional challenges. (To note these instances, the
number of challenges counted per case is indicated in the final column of the
appendix, below). However, if multiple provisions formed a single functional unit,
and as a consequence hung or fell together, we counted them as a single constitutional
challenge.

B. Coding the Judgments

Once we compiled the list of relevant cases, we coded the cases on five variables:

1. the type of majoritarian action challenged,

2. the right at issue,

3. whether the fight challenged was subject to the section 33

override,

4. whether the Court found that a right was violated, and

5. whether the violation was saved by section 1.

Where we found that the Supreme Court did not deal with a constitutional issue
on its merits-for instance, where a constitutional challenge was dismissed for a
procedural issue such as standing or mootness-we eliminated the case from the data
set. We might have included these cases in the data set on the grounds that the Court

72 Criminal Code, supra note 70, s. 190(3).
.R. v. Jones, [1991] 3 S.C.R. 110, 8 C.R. (4th) 137.

2003] S. CHOUDHRY & C.E HUNTER- MEASURING JUDICIAL ACTIVISM

543

may have manipulated procedural rules in order to dismiss Charter applications that it
would otherwise have dismissed on substantive grounds. While the inclusion of these
cases would bolster our claim that the Court is not as activist as Justice Marshall
suggests, we have excluded them on the basis that they are not helpful to an analysis
of the Court’s approach to Charter interpretation. In general, the measures coded in
our study were sufficiently crude to produce extremely high inter-coder reliability.’

C. Assessing the Hypotheses in Light of the Data
Recall that there are four empirical claims implicit in Justice Marshall’s judgment.
First, in absolute terms, there is a high level of judicial activism. Second, judicial
activism is increasing over time. Third, the root of this activism is a failure of the
Court to be sufficiently deferential to governments in its interpretation of section 1.
Fourth, the override is of no use in combating judicial activism because it has been
delegitimized. We use the empirical data collected here to test these claims, and by
implication, the normative arguments that rely on these empirical propositions.

In order to evaluate absolute levels of activism and the increase of activism over
time (hypotheses 1 and 2), we measured the government “win rates” by year. In order
to determine win rates, we looked at the outcome in each case. Where a claimant was
in showing that a Charter right had been violated, or where the
unsuccessful
government was successful in arguing that a rights violation was justified under
section 1, the case was coded as a government “win”. Where the claimant was
successful in showing a rights violation and where the government was unsuccessful
in justifying the violation under section 1, the case was coded as a “loss”. In order to
determine win rates for any given year, the number of cases coded as wins was
divided by the total cases decided that year. If hypothesis I were true, we would
expect to see consistently low government win rates over time. If hypothesis 2 were
true, we would expect to see decreasing government win rates over time.

Hypothesis 3 required the isolation of the Supreme Court’s behaviour at the
section 1 stage of analysis to determine if that provision served as a site of activism.
In order to test whether the Court is activist at the section 1 stage of analysis, we
looked at two pieces of information. First, we isolated the cases in which a rights
violation had been found and where a justification under section 1 was considered
(whether or not the government had presented evidence under section 1).’ We

4 The authors coded all the cases used for this study. In order to test inter-coder reliability, however,
we tested twenty randomly selected cases coded by a third party and found inter-coder reliability of
one hundred percent in the sample.

” An example of a case where the government did not present arguments under section 1 is Miron v.
Trudel, [1995] 2 S.C.R. 418, 124 D.L.R. (4th) 693. The inclusion of cases in which the government
did not present evidence of a section 1 violation serves to make the Court appear more activist than if
these cases had been excluded. Although there is a case for exclusion, we decided in the interests of

544

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

calculated absolute government win rates within this group to determine the extent to
which the government was successful in justifying rights violations under section 1.
We also traced section 1 government win rates from year to year. Second, we divided
the cases in this group according to the substantive right at issue, on the basis that
section 1 might be differently applied depending on how easy or difficult it was to
establish a particular substantive rights claim. We looked specifically at three
substantive sections which cover the spectrum in terms of the difficulty of establishing
a substantive rights claim. At one end of the spectrum we looked at cases involving
section 2(b), which has been interpreted by the Court to encompass virtually all
expressive activity,’ and therefore contains little in the way of internal limits. At the
other end of the spectrum we looked at section 7, which has an explicit internal limit.’
Between these two extremes, we looked at section 15, which contains internal limits
“read-in” by the Court.” If hypothesis 3 were correct, we would expect to see low
government win rates in section 1. Furthermore, we would expect to see fairly
consistent government win rates in section 1 cases across the different substantive
rights sections.

In order to test hypothesis 4 (whether or not the section 33 override has been
delegitimized) we isolated the cases in which the right at issue was subject to the
override (i.e., cases involving section 2 or sections 7 through 15).”M We isolated these
cases from the main data set because the delegitimization of the override should have no
effect on government win rates with respect to non-overridable rights (section 33 of the
Charter has never operated as a constraint on the Supreme Court in the latter group of
cases). We calculated government win rates by year in order to test whether there had
been an increase in activism in the period after 1988, when the delegitimation of section
33 of the Charter is alleged to have occurred. If this hypothesis were correct, we would
expect to see a decrease in government win rates after 1988.

Ill. Results, Analysis, and Discussion

A. Hypothesis 1: Judicial Activism Is High

Figure 1, below, shows the possible outcomes of a Charter challenge in terms of
government wins and losses. There are three possible outcomes, two resulting in a

inter-coder reliability to include these cases because it is not always possible to determine from the
reasons for judgment whether in fact the government presented evidence on this point.

‘6 Charter, supra note 3, s. 2(b). See Irwin Toy, supra note 51.

Charter, ibid., s. 7 (“[e]veryone has the fight … except in accordance with the principles of
fundamental justice” [emphasis added]). See Reference Re Motor Vehicle Act (British Columbia)
s. 94(2), [1985] 2 S.C.R. 486, 24 D.L.R. (4th) 536 [B.C. Motor Vehicle cited to S.C.R.].

7 Charter, ibid., s. 15. See Law v. Canada, [1999] 1 S.C.R. 497, 170 D.L.R. (4th) 1. See also
Lavoie v. Canada, [2002] 1 S.C.R. 769, 210 D.LR. (4th) 193, 2002 SCC 23 (where the question of
whether section 15 contains internal limits sparked deep divisions within the Supreme Court).

“‘ See text of subsection 33(1), supra note 21.

2003] S. CHOUDHRY & CE HUNTER- MEASURING JUDICIAL ACTIVISM

545

government win and one resulting in a government loss. First, the government wins if
the claimant is unsuccessful in showing that a right has been violated (outcome A).
Second, the government wins if there has been a rights violation, but the violation is
justified under section 1 (outcome C). Finally, the government loses only where a
right has been violated and that violation is not saved under section 1 (outcome D).
Based on the table in Figure 1, the overall government win rate is (A+C)/(A+B).
When we calculated the overall government win rate for all of the cases in our data
set, we found that governments won 62.4 percent of the time. Put differently, in cases
where a majoritarian act is at issue, governments lost only 37.6 percent of the time.

Possible Outcomes of a Charter Challenge

FIGURE 1

All Charter Challenges

(A)

No Violation of Rights

(Government Win)

(B)

Violation of Rights

(C)

Violation Justified
Under Section 1
(Government Win)

(D)

Violation Not Justified

Under Section1

(Government Loss)

We imagine that this statistic could be relied on by those in the legal academy
who hold that the Court is rather deferential to governments. However, as we noted
above in Part I, it is difficult to assess whether this number should in fact be viewed as
high or low in absolute terms. If Justice Marshall believes in the idea of constitutional
supremacy, he likely recognizes that the Supreme Court is sometimes justified in its
interference with majoritarian acts. Unfortunately, it is difficult to quantify the portion
of time in which the Court is justified in interfering with majoritarian acts. As a result,
it is not possible to measure the portion of government losses that Justice Marshall
the number provides useful
would consider “undue
information for the debate over judicial activism. It is clear that the government is
successful in protecting majoritarian decisions from Charter challenges nearly two-
thirds of the time. The government loss rate of 37.6 percent provides an absolute ceiling
for the number of cases that might be characterized as activist by critics of the Court.

incursions”. However,

It is interesting to note that our overall rate is quite similar to the results obtained in
the Morton/Kelly work, detailing the Court’s activity from 1982 to 1997.’ Kelly states
that in the first 352 Charter cases, from 1982 to 1997, the government win rate was 69
percent in Charter cases involving statutes and regulations (versus our finding of 62.4

Morton, Russell & Riddell, supra note 34; Kelly, “Rebalancing”, supra note 33.

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol.48

percent).” If anything, our data shows a higher level of overall activism than the previous
work. These differences might be attributable to differences between our methodology
and that of these previous studies.’2

B. Hypothesis 2: Judicial Activism Is Increasing over Time
If hypothesis 2 were true, we would expect to see a downward trend in
government win rates across time, assuming that the types of cases heard each year
are similar. However, when we analyzed absolute government win rates by year
(Figure 2), we observed that they varied widely.

Absolute Government Win Rates, 1984-2002

FIGURE 2

Wins

Losses

Total

Win Rate

1
1

2
3
12
5
15
9
11
16
7
11
3
4
3
6
8
4
5

2
3
1

2
6

2

4
6
5
7
2

4
1
10
2
3
1
3
6

3
4
3
5
18
7
19
15
16
23
9
15
4

14
5
9
9
7
11

33.3/
25.0%
66.7%
60.0%

66.7%

71.4%
78.9%
60.0%
68.8%
69.6%
77.81/

73.3%
75.0%
28.5%
60.0/

66.7%
88.9%
57.1%
45.5%

Year

1984
1985
1986
1987
1988
1989
1990
1991
1992
1993

1994
1995
1996
1997
1998
1999

2000
2001
2002

Moreover, the graphical depiction of the trend over time (Figure 3) shows the absence
of a downward trend. Indeed, there was a significant dip in the government win rate in
1997, and had this trend continued, the government win rate should have been quite
depressed by 2000 (when, in fact, the government enjoyed its highest ever overall win

Kelly, “Rebalancing”, ibid. at 656.
82See Part I.A, above.

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

547

rate). In sum, the evidence with respect to absolute government win rates does not
support the hypothesis that judicial activism is increasing over time. These results are
bolstered by the results of Kelly’s 1997 study, which found a slight upward trend in
government win rates in the period between 1984 and 1997.11

Absolute Government Win Rates by Year, 1984-2002 (Graph)

FIGURE 3

100
80
60
S40
20

Year

The question raised by this data is whether there were other factors

that
influenced and that possibility disguised a downward trend in win rates, thereby
masking a result that would otherwise support hypothesis 2. The most plausible
is “government learning”: As time passes, Charter jurisprudence
explanation
develops and governments study the Court’s evolving interpretation of the Charter.
Over time, governments that are strategically minded learn to conform both their
legislation and their arguments in Charter cases to the Court’s jurisprudence. The
result of government learning should be an increase in the government win rate. The
absence of either an observable downward trend or an observable upward trend in
government win rates, however, could also be consistent with the government learning
story if judicial activism increased over the same period. Perhaps government win
rates were being pushed upward by government learning, at the same time as they
were being pushed downward by increasing judicial activism. The net result would be

” Kelly, “Rebalancing”, supra note 33 at 655-56. Kelly presents his data in terms of Charter
claimant success rate so he frames the trend as “downward”. However, when the trend is recast as
government win rates, it reverses (i.e., becomes “upward”). Kelly notes that for the first 100 Charter
decisions, the success rate of Charter claimants in cases involving statutes and regulations was 38
percent. This figure drops to 34 percent for the first 195 Charter decisions, and then to 31 percent for
the first 352 Charter decisions. Recall, however, the distinctions between Kelly’s data set and the data
set used here, described in Part I.A, above.

548

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

that the two forces cancel each other out, leading to the absence of a discernable trend
in the aggregate data.’

Unfortunately, it is not possible to assess the validity of the government learning
theory on the basis of the available data. An assessment of this hypothesis would
require data that assesses the behaviour and motivation of various actors, presumably
in the form of interviews. First, we would need data that analyzes whether, and to
what extent, legislators and bureaucrats take account of the possibility of potential
constitutional challenges under the Charter in framing legislation. Second, we would
need to establish, through interviews, how government litigation counsel determines
which Charter claims to defend in terms of the level of attention given to prior
jurisprudence (particularly the manner in which the counsels frame their section 1
justification cases) and the level of consideration paid to the likely success of each
case. In the absence of such data, the possibility of a counter-trend of government
learning to a trend of judicial activism is real and must be viewed as a limitation of
our study.

C. Hypothesis 3: The Section 1 Analysis Is the Locus of Activism

1. Overview

As Figure 1 shows, there are two points at which the government can win a
Charter case: at the “rights violation” stage or at the “justification” (“section 1″)
stage. Recall that Justice Marshall focussed his criticisms at the Supreme Court’s
section 1 jurisprudence. In particular, Justice Marshall seemed to claim that
governments tend to lose at section 1, which he took to be a measure of judicial
activism since the section 1 analysis requires the Court to second-guess legislative
decisions through the proportionality analysis of the Oakes test. Justice Marshall’s
judgment thus forces us to analyze the specifics of the Court’s behaviour.

The overall government win rates on their own do not permit us to draw any
conclusions about the Court’s behaviour at section 1. To understand why, consider
two radically different scenarios, suggesting two very different patterns of judicial
behaviour, which would generate the same overall government win rates. In one
scenario, if governments win, they do so at the rights violation stage. Thus, at this first
stage, we would see a mixture of government wins and losses, but very few at the
section 1 stage. In the second scenario, when governments win, they do so at the
section 1 stage. If this were the case, whether the government ultimately wins or

” Kelly alludes to this argument by suggesting that “the administrative state engaged in a process of
risk assessment of public policy to minimize the judicial nullification of statutes” (ibid. at 664). See
also James B. Kelly, “Bureaucratic Activism and the Charter of Rights and Freedoms: The
Department of Justice and its Entry into the Centre of Government” (1999) 42 Canadian Public
Administration 476.

2003] S. CHOUDHRY & C.E HUNTER- MEASURING JUDICIAL ACTIVISM

549

loses, we would expect to see most cases lead to a finding that a right had been
violated. We would also expect to see a higher rate of government wins based on a
section I justification than we would see under the first scenario.

In order to understand where the government wins and loses, it is necessary to
compare government win rates at the rights violation stage and at the justification
stage. In our data set, the government won just over 50 percent at the rights violation
stage. At the section 1 stage, the government won just over 25 percent of the time.
(Cases that reached this stage of analysis are identified as the “Section 1 Subset” in
the appendix, below.) A consequence of these differences in win rates at the different
stages of Charter analysis is that a very large proportion of government wins occured
at the rights violation stage (79.5 percent). Although Kelly did not make this exact
comparison, he noted a gap between the overall government win rate (66 percent) and
the win rate at the section 1 stage (39 percent) between 1984 and 1997, a slightly
different statistical comparison, which reflects the same underlying data.” At first,
these results seem to support Justice Marshall’s view that the Court’s activism is
located
this
interpretation of the data can be simplistic.

the section 1 stage.” Further analysis shows, however, how

in

While it is correct to observe that there is a gap between the government win rates
at the rights violation stage and at the justification stage, it is important to note that
every case that reaches the section 1 analysis has already been through an analysis at
the rights violation stage. Only the cases that are found to violate a right will proceed
to the section 1 analysis. Therefore, when we compare the government win rate at the
rights violation stage to the government win rate at the section 1 stage, the group of
cases at the section 1 stage is in fact a sample from the pool of cases tested at the
rights violation stage. Moreover, the cases where the government wins at the rights
violation stage are not tested at the section 1 stage. The potential problem that we
encounter, therefore, is sampling bias. In the cases that reach the section 1 analysis
stage, the government has already lost once. It seems prima facie plausible to suggest
that a group of cases, in which the government has already lost on one ground,
includes more cases that the government actually ought to lose than does the group of
cases that are never tested at section 1 (because the government won at the rights
violation stage).

If true, this would mean that the rights violation analysis acts as a sieve for some
constitutional challenges
that governments would have otherwise successfully
defended under section 1. We would expect that this phenomenon is occurring for
cases based on substantive rights with strong internal limits. For example, in order to
find a violation of section 7, there must be a deprivation of a protected interest (e.g.,
life, liberty, and security of the person) and that deprivation must infringe principles

s Kelly, “Rebalancing”, ibid.

This point is also made by some political scientists. See e.g. Kelly, “Rebalancing”, ibid.

550

MCGILL LAW JOURNAL / REVUE DE DROITDE MCGILL

[Vol. 48

of fundamental justice.” Once legislation is found to be in non-conformance with the
principles of fundamental justice, it seems unlikely (and the Supreme Court has said
as much) that it will be found to be “demonstrably justified in a free and democratic
society.”

In order to test this argument, we isolated section 1 justification results according
to the different substantive rights affected. If the substantive rights with internal limits
were in fact acting as a sieve, there should be a discemable difference between
government success rates at the section 1 stage, depending on whether the substantive
right at issue contains internal limits. On the other hand, if government success rates
at the justification stage were to remain constant, regardless of the presence or
absence of internal limits in the substantive right, this would bolster the claim that the
Court is activist in its use of section 1.

2. Variation by Substantive Charter Right

We observed a substantial variation in levels of judicial activism depending on the
section of the Charter at issue. As predicted, there was a monotonic relationship
between the internal restrictions of the different substantive Charter rights and the
government win rates at the section 1 stage (Figure 4). That is, the easier it was for a
violation of rights to be established, the higher the government win rate in section 1.
In cases involving section 2(b), a section with few internal restrictions, for example,
the government was successful in justifying rights violations 44.4 percent of the time.
On the other hand, in cases involving section 7, which contains strong internal limits,
the government success rate at section 1 was only 6.3 percent. In between these two
extremes are the section 15 cases, where the Court has read-in a set of internal
restrictions that are more stringent than the minimal restrictions in section 2(b), but
less stringent than section 7’s explicit restrictions. As expected, for cases involving
section 15, the government’s success rate in section 1 was 20.0 percent.

‘7 Charter, supra note 3, s. 7; B.C. Motor Vehicle, supra note 77.
“B.C. Motor Vehicle, ibid. at 518; Gosselin v. Quebec (A.G.) (2002), 221 D.L.R. (4th) 257, 100
C.R.R. (2d) 1, 2002 SCC 84 at paras. 389-90, Arbour J., dissenting; New Brunswick (Minister of
Health and Community Services) v. G.(J.), [1999] 3 S.C.R. 46 at para. 99, 177 D.L.R. (4th) 124,
Lamer CJ.C.

20031 S. CHOUDHRY& C.E HUNTER – MEASURING JUDICIAL ACTIVISM

551

Government Section 1 Win Rates by Substantive Charter Right

FIGURE 4

Substantive
Charter Right

Section 2(b)

Section 15

Section 7

Government

Wins at Section 1

(C)

Government

Losses

(D)

12

2

1

15

8

15

Section 1
Win Rate
(C/C+D)

44.4%

20.0%

6.3%

These results are consistent with our hypothesis that greater internal restrictions in
a substantive right result in lower section 1 government win rates. Internal restrictions
resulted in only the most egregious cases making it to the section 1 test. This theory is
reinforced by examining the aggregate win rate, which includes both government
wins at the section 1 stage and government wins because of a failure to establish a
rights violation, by substantive section (Figure 5). When we isolated the total
government win rate by substantive right, we found that the relationship between
internal limits and the government win rate was completely reversed. We found that
section 7, the section with the lowest section 1 government win rate, had the highest
overall government win rate. Conversely, section 2(b), which had the highest section 1
government win rate, had the lowest overall government win rate.

Aggregate Government Win Rates by Substantive Charter Right

FIGURE 5

Substantive
CharterRight

Section 2(b)

Section 15

Section 7

Government Wins

Total

(A+C)

18

16

55

Government

Losses
(D)

Overall
Win Rate

(A + C /A + B)

15

8

15

54.5%

66.7%

78.6%

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

The implications of this last result are worth noting. Not only is the tracking of section
1 government win rates a poor predictor of overall government success, it predicts in
fact precisely the opposite trend. The lower the section 1 government win rate, the
higher the overall government win rate with respect to a particular right. This
conclusion suggests that government success in section 1 is, at the very least,
unrelated to its general success in contesting Charter challenges, and as a result, a
claim of Court activism in its use of section 1 does not go very far to establishing a
general claim of judicial activism.

This last line of analysis illustrates the importance of designing quantitative
analysis around a doctrinal understanding of the structure of Charter adjudication.
Previous studies merely noted the relatively low government win rate at section 1, and
have concluded from it that the Court was particularly activist at that stage of Charter
analysis. Morton, Russell, and Riddell stated, for example, that “the Court has been
more inclined to reject governments’ section 1 defences … than to accept them …
consistent with the Court’s new judicial activism”‘ What these studies have failed to
consider is the potential interrelationship between the section 1 win rate and the
relative ease or difficulty of establishing a rights violation for particular Charter
provisions. Demonstrating that such a relationship exists not only makes it more
complicated to make charges of judicial activism at the justification stage, but also
suggests that a low section 1 win rate may predict government success for challenges
brought on the basis of rights with an internal limit.

3. Trends over Time

Another part of Justice Marshall’s claim about judicial activism in its use of
section 1 is that it has been increasing since the Court’s 1995 judgment in RJR-
Macdonald. Figure 6 shows the government’s section 1 win rates over time. There
was no clear downward trend in government win rates after 1995. In fact, in 1996, the
year immediately following the decision in RJR-Macdonald, the government had its
highest ever section 1 win rate of 66.7 percent. The absence of a discernable trend in
section 1 win rates was also noted in Kelly’s 1997 study, where he found that after an
initial reluctance to accept section 1 arguments, the Court had been inconsistent in its
level of deference under its section t analyses.”

‘9 Morton, Russell & Riddell, supra note 34 at 30.
‘0 Kelly, “Rebalancing”, supra note 33 at 663. Kelly finds that from 1984 to 1987, “only 20
[percent] of section I defences … were accepted by the Supreme Court. … In the three-year period
between 1988 and 1990, this number increased to 55 [percent] … However, between 1984 and 1992,
the Court accepted 39 [percent] … of the section I defences presented by government” (ibid.). This
rate was the same for the period between 1993 and 1997.

2003] S. CHOUDHRY & C.E HUNTER- MEASURING JUDICIAL ACTIVISM

553

Government Section 1 Win Rates, 1984-2002 (Graph)

FIGURE 6

100
90
80
70
o60
50
40
30
20
10
0


E

Year

These results support the view of those in the legal academy who discount the
importance of the section 1 analysis in RJR-Macdonald. Kent
precedent-setting
Roach, for example, remarks that “the majority’s decision striking down the tobacco
advertising restrictions did not mark the abandonment of the more deferential
approach.”” Our explanation for the lack of any discernable impact of RJR-
MacDonald on the Supreme Court’s approach to section 1 is that the judgment can be
explained by some unique features of the case. In particular, the Attorney General of
Canada took the astonishing position of refusing to disclose to the Court its own
studies on the effectiveness of a ban on the advertising and marketing of tobacco in
reducing tobacco consumption. In his judgment, Justice lacobucci (with whom Chief
Justice Lamer concurred) expressed concern that “the Attorney General of Canada
chose to withhold from the factual record evidence related to the options it has
considered as alternatives to the total ban it chose to put in place”‘ ” and arguably
resolved the appeal against the government for that reason. Our analysis of section 1
government win rates supports this view that RJR-Macdonald did not have a
significant impact on levels of judicial activism in section 1.

D. Hypothesis 4: The Override Has Been Delegitimized
The final hypothesis is that the override has been delegitimized through its
unpopular use by Quebec in 1988, and that as a result, the Court, emboldened by the
lack of constraint, has become increasingly activist in the following years. This is a
question that was not examined by previous studies. In order to test this hypothesis,

9, Roach, supra note 28 at 162.
92 RJR-Macdonald, supra note 27 at para. 186.

554

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

we isolated from the data set only the cases that involved overrideable rights (i.e.,
cases involving sections 2 or 7 through 15 of the Charter), on the basis that only these
types of cases would be subject to the impact of a new activism brought about by the
delegitimization of the override. (These cases are identified as the “Override Subset”
in the appendix, below,) If the delegitimization of the override by Bill 178 has fueled
an increase in judicial activism, there ought to be some difference between the
observed level of activism in the judgments prior to 1988 and in the period that
followed the supposed delegitimization.

Government Win Rates in Overrideable Rights Cases, 1984-2002

FIGURE 7

100
90
8o
7060
rr 50
40-
30-
20
10
0*

Year

Figure 7 shows government win rates for cases involving overrideable rights from
1984 to the present. There is no discemable trend with respect to this rate, and
certainly no obvious drop-off in 1988 or 1989. Moreover, if we take overall measures
of win rates in the periods before and after the use of the override to immunize Bill
178, we see that there is little if any difference in win rates. In the period between
1984 and 1988, governments won cases involving overrideable rights sixty-five
percent of the time. From 1989 to 2002, they won sixty-eight percent of the time. If
we move the cut-off date back one year to 1989, to allow time for the supposed
delegitimization to have taken effect, the story does not change much. In the period
from 1984 to 1989, governments won sixty-eight percent of the time and from 1990
to 2002, governments won sixty-five percent of the time.

It is important to be clear about the significance of this finding. This result
undermines the allegation of a pattern of judicial behaviour that is part of Manfredi’s
account of the impact of the delegitmization of the override on Charter adjudication
(i.e., that the levels of judicial activism have risen since the delegitimization of the
override through its use in Bill 178). However, the absence of such a trend does not
prove that the override has not been delegitimized. In fact, this result is consistent with

20031 S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

555

a number of different hypotheses. Perhaps, and contrary to Manfredi’s assertions, the
override has not been delegitimized. It is interesting, for instance, to note that the
override was used by the Alberta legislature in March 2000 to shield the heterosexual
definition of marriage from constitutional challenge? On the other hand, perhaps the
override has been delegitimized and replaced by other constraints to judicial activism
in the Canadian constitutional system, such as public opinion, to which the Court is
responding strategically. A poll taken in May 2002 reported that fifty-four percent of
respondents opposed the very idea of the override? Finally, perhaps the override was
and remains irrelevant to the adjudication of constitutional cases. Regardless of which
story is ultimately true, we did not observe anything that .suggested definitively the
deligitimization of the override on Charter adjudication.

Conclusion

There are three significant limitations of this study that prevent us from providing
a stronger positive account of judicial activism or response to Justice Marshall’s
remarks in NAPE. The first limitation is the inability of our study to differentiate
between the Supreme Court’s legitimate interference with legislative decisions and
“undue” interference. While our working definition of counter-majoritarian judicial
activism captures all instances in which the Court interfered with a majoritarian
decision, even critics of the Court’s activism, who are nonetheless committed to a
regime of constitutional supremacy enforced by judicial review, must admit that there
are some instances that call for the Court’s interference. Because we are not able to
distinguish appropriate from inappropriate interference, we are not able to ascertain
whether the levels of appropriate interference stay constant over time, or whether
there is significant variation in the levels of appropriate interference. In analyzing our
data, we have assumed that there was some constant but unknown portion of
government losses that would be considered legitimate, which allowed us to draw
inferences about the meaning of trends in win rates for judicial activism. If our
assumption is wrong and this number is changing over time, our inferences may be
faulty.

The second limitation is the small number of counter-majoritarian Charter cases
at the Supreme Court level in some years. For example, in the years of 1984 to 1987,
1996, and 1998, there are five or fewer Charter cases per year in our data set. As we
divide the data set further to isolate section 1 or overrideable cases, this problem of a
small number of cases is exacerbated.

9’ Bill 202, Marriage Amendment Act, 4th Sess., 24th Leg., Alberta, 2000, cl. 5 (assented to 23
March 2000), S.A. 2000, c. 3. The statute will probably not have this effect, because it is likely
unconstitutional on federalism grounds.

‘ Choudhry, Book Review, supra note 29 at 387.

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

Third, there is the potential problem of the Supreme Court’s selection bias. Most
of the Court’s docket (with the exception of as of right criminal appeals) is controlled
by the Court itself, which grants parties leave to appeal. As a consequence, it is open
to the Court to determine the mix of Charter challenges that it hears. Although the
Supreme Court Ace provides that the Court grants leave to appeal in cases that raise
questions of “public importance”,” the Court grants leave to appeal without issuing
reasons (save for exceptional circumstances), making it difficult to ascertain whether
other criteria enter into its decisions in leave applications. It is therefore possible that
the Court could use the leave to appeal process to respond to the activism debate. For
example, in an effort to appear less activist, the Court may intentionally grant leave to
appeal in a few cases each year where governments will win, in order to enhance the
perception that the Court is deferential to governments. On the other hand, it is
entirely plausible that the Court may disproportionately take cases where it will find a
majoritarian decision to be unconstitutional, reducing the number of cases on its
docket where it would have deferred. Another possibility is that the Court simply
hears cases that raise interesting points of law, without regard to the potential
unconstitutionality of a majoritarian decision (a neutral factor in the Court’s calculus)
or to the public debate raging over judicial activism. The issue of selection bias is
evidently an important, confounding variable, and as a consequence, must be viewed
as a serious limitation to our study.

The empirical evidence that we have collected does not appear to be consistent
with any one of the four hypotheses that we derived from the empirical assumptions
underlying Justice Marshall’s claims about judicial activism. While the Court
interferes with majoritarian decisions in 37.6 percent of cases, given that at least some
interventions would be perceived even by critics to be legitimate, this number does
not seem as high as some would lead us to believe. When we look at the government’s
success rate over time, there is little support for the view that judicial activism is
systematically increasing. We paid particular attention to section 1 cases in order to
assess Justice Marshall’s concerns about a lack of deference at the section 1 stage of
analysis. We found that deference at this stage is intimately linked with the nature of
the substantive right at issue, and that government success in the rights violation
analysis was much more
than was
government success with section I justifications. Finally, we tested the hypothesis that
the delegitimization of the section 33 override has led to widespread activism in the
post-Bill 178 period, and concluded that there was no evidence to support such a
hypothesis.

important to overall government success

The data collected in this study is not sufficiently detailed for us to offer a positive
account that explains both the scope and causes of judicial activism, and the extent to
which its levels are changing over time. The data presented here, however, does raise

R.S.C. 1985, c. S-26.
Ibid., s. 40(l).

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

557

some serious questions about the empirical assumptions made by critics of the
Supreme Court. Justice Marshall’s normative claims in NAPE rely heavily on a
number of assumptions that we have shown here to be highly suspect. While
constitutional scholars have long engaged critics of the Court on a normative level,
Justice Marshall’s judgment in NAPE makes it imperative to engage the Court’s
detractors in a logically prior conversation as well. The question should no longer be
merely whether judicial activism is good, but also whether judicial activism-is real.

558

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

Appendix

Counter-Majoritarian Cases

Counter-Majoritarlan Cases

Section 1 Over
Subset
Sub

4

Andrews v. Law Society of British Columbia, [1989] 1 S.C.R. 143
B(R.) v. Children’s Aid Society of Metropolitan Toronto, [1995] 1 SC.R. 315
British Columbia Government Employees’ Union v. British Columbia (A. G.), [1 988]
2 S.C.R. 214
British Columbia Securities Commission v. Branch, [1995] 2 S.C.R, 3
Baron v. Canada, [1993] 1 S.C.R. 416
Banner v. Canada (Secretary of State), [1997] 1 SC.R. 358
Canada (Human Rights Commission) v. Taylor, [1990] 3 SC.R. 892
Canadian Broadcasting Corp. v. New Brunswick (A.G.), [199613 S.C.R. 480
Canadian Egg Marketing Agency v. Richardson, [1998] 3 S.C.R .157
Canadian Newspapers v. Canada (A.G.), [1988] 2 S.C.R. 122
Chiarelli v. Canada (Minister of Employment and Immigration), [1992]1 S.C.R. 711
Comitdparitaire de findustrie de Ia chemise v. Potash, [19941 2 S.C.R. 406
Corbiere v. Canada (Minister of Indian and Northern Affairs), [1999] 2 S.C.R. 203
Corporaon professionelle des mdcins du Quebec v. Thibault, [1988]
1 S.C.R. 1033
Cunningham v. Canada, (19931 2 S.C.R. 143
DelZotto v. Canada, [1999] 1 S.C.R. 3
Delisle v Canada (A.G.), [1999] 2 S.C.R. 989
Devine v. Quebec (A.G.), [1988] 2 S.C.R. 790

Dunmore v. Ontario (A.G.), [2001] 3 S.C.R. 1016
Dywidag Systems International Ltd. V. Zutphen Brothers Construction Ltd., [1990]
1 S.C.R. 705
Edmonton Joumal v. Alberta (AG.), [1989) 2 SC.R. 1326
Egan v. Canada, [1995] 2 S.CR, 513

Ford v. Quebec (A.G.), [198812 S.CR. 712

Godbout v. Longueuil (City of), [1997] 3 S.C.R. 844
Granovsky v Canada (Minister of Employment and Immigration), [2000] 1 S.C.R. 703
Haig v. Canada (Chief Electoral Officer), [1993] 2 S.C.R. 995
Harvey v. New Brunswick (AG), [1996] 2 S.C.R. 876
Hunter v. Southam Inc., [1984] 2 S.C.R. 145
/dziak v Canada (Minister of Justice), [19921 3 S.C.R. 631
Irwin Toy Ltd. v. Quebec (A.G.), [1989] 1 S.CR. 927

2003] S. CHOUDHRY & C.E. HUNTER- MEASURING JUDICIAL ACTIVISM

559

Counter-Majoritarlan Cases

Section 1 Override
Subset
Subset

Number of
Challenges

Coded

Kindler v. Canada (Minister of Justice), [1991] 2 S.C.R. 779
Lavallee, Rackel & Heintz v, Canada (A.G), White, Ottenheimer & Baker v.
Canada (A.G.); R. v. Fink, 2002 SCC 61
Lavigne v. Ontario’s Public Senvice Employees Union, [1991] 2 S.C.R. 211
Lavoie v. Canada, 2002 SCC 23
Law v. Canada (Minister of Employment and Immigration), [199911 S.C.R. 497
Law Society of Upper Canada v. Skapinker, [198411 S.C.R. 357
Libman v. Quebec (A.G.), [1997] 3 S.C.R. 569
Little Sisters Book and Art Emporium v. Canada (Minister of Justice), [2000]
2 S.C.R. 1120
M. v. H., [199912 S.C.R. 3
Mackin v. New Brunswick (Minister of Finance, Rice v. New Brunswick, [2002]
1 S.C.R. 405
Mahe v. Alberta, [19901 1 S.C.R. 342
Miron v. Trudel, [1995] 2 S.C.R. 418
Nova Scotia (A.G.) v. Walsh, 2002 SCC 83
Ontario v. Canadian Pacific Ltd. (1995] 2 S.C.R. 1031
Osbome v Canada (Treasury Board), [1991] 2 S.C.R. 69
P.(D.) v, S.(C.), [199314 S.C.R. 141
Pearlman v Manitoba Law Society Judicial Committee, [1991] 2 S.C.R. 869
Professional Institute of Public Service of Canada v. Northwest Territories
(Commissioner), [1990] 2 S.C.R. 367
Quebec Association Of Protestant School Boards v. Quebec (A.G.), [1984]
2 S.C.R. 66
R. v. Advance Cutting & Coring Ltd., [2001) 3 S.C.R. 209
R. v. Bain, [199211 S.C.R. 91
R. v Beare, [1988] 2 S.C.R. 387
R. v. Bernard, [19881 2 SC.R. 833
R. v Big M Drug Mart Ltd., [1985] 1 S.C.R. 295
R. v. Brown, [19941 3 SC.R. 749
R. v. Butler, [1992] 1 S.C.R. 452
R. v. Chaulk, [1990] 3 S.C.R. 1303
R. v. Colarusso, [19941 1 S.C.R. 20
R. v Corbett, [198811 S.C.R. 670
. . Creighton, [19931 3 S.C.R. 3
R. v Darrach, [2000] 2 S.CR. 443

*

*

*

*

*

*

*

*

*

,

1
1

1

1
1
1 1
1
1
2

1
1

2
1
1
1
1
1
1
1

1

1
1
1
1
1
1
1

1
1
1
1
3

560

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

Counter-Majorltarlan Cases

Secton 1 Override
Subset
Subset

Number of
Cuaenges
Clen
Coded

F. v. OeSousa, [1992] 2 S.C.R. 944
R. v. Duarte, [1990] 1 S.C.R. 30
R. v. Edwards Books andArt Ltd., [198612 S.C.R. 713
R. v. Finlay, [1993] 3 S.C.R. 103
F. v. Finta, [1994] 1 S.CR. 701

R. v. Fitzpatrick, [1995] 4 S.C.R. 154

R. v. Furtney, [1991] 3 S.C.R. 89

R. v. Gdndreux, [1992] 1 S.CR. 259
R. v. Goltz, [1991] 3 S.C.R. 485
F?. v. Grant, [1993] 3 S.CR. 223

R. v. Guignard, [2002] 1 SCR, 472
R. v. Hall, 2002 SCC 64
R. v. Hess; F. v Nguyen, [1990] 2 S.C.R. 90

R. v. Heywood [1994] 3 S,C,R. 761

R, v. Holmes, [1988] 1 S.C,R. 914

R. v. Hufsky, [1988] 1 S.C.R. 621

R. v. Jobin, [1995] 2 S.C.R. 78

R. v Jones, [1986] 2 S.C.R, 284
R. v. Jones, [1994] 2 S.C.R. 229
R. v. Keegstra, [1990 3 S.C.R. 697

R. v. L (D.O.), [1993] 4 S.C.R. 419

F. v Laba, [199413 S.C.R. 965
F?. v. Ladouceur, [1990] 1 SC.R. 1257
F?. v. Landry, [1991] 1 S.C.R. 99
F?. v. Latimer, [2001] 1 S.C.R. 3

R. v Lee, [1989] 2 S.C.R. 1384

F?. v Levogiannis, [1993] 4 S.C.R. 475
R. v. L ppd, [1991] 2 S.C.R. 114

R v. Logan, [1990] 2 S.C.R. 731

R. v. Lucas, [1998] 1 S.C.R. 439

F?. v. Lyons, [1987] 2 S.C.R. 309

R. v Martineau, [1990] 2 S.C.R. 633
R. v. McKinlay Transport Ltd., [1990] 1 S.C.R. 627
R. v. Mills, [1999] 3 SC.R. 668

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

1

1

1

1

2

1

1

1

1

2

1

1
1

1

1

1

1

1

1

1

1

1
1
1
1

1

1

1

1

1

1

1

1

1

2003] S. CHOUDHRY & C.E HUNTER- MEASURING JUDICIAL ACTIVISM

561

Counter-Majoritarlan Cases

Section 1 Override
Subset
Subset

Number of
Challenges

Coded

R. v. Milne, [198712 S.C.R. 512
R. v. Morales, [1992] 3 S.C.R. 711
R. v. Morgentaler, [1988) 1 S.C.R. 30
R v. Morisey, [2000] 2 S.C.R. 90
R. v. Nova Scotia Pharmaceutical Society, [1992] 2 S. C.R. 606
R. v. Oakes, [1986] 1 S.C.R. 103
R. v. Osoin, [19931 4 S.C.R. 595
R. v Pan, R. v. Sawyer, [2001) 2 S.C.R. 344
R. v Pearson, [1992] 3 S.C.R. 665
R. v. Penno, [1990] 2 S.C.R. 865
R. v Pontes, [1995] 3 S.C.R. 44
F v. Potvin, [1989] 1 S.C.R. 525
R. v. Ratti, [1991]1 S.C.R. 68
R. v. Richard, [1996] 3 S.C.R. 525
R. v. Rose, [1998] 3 S.C.R. 262
R. v. Rube, [1992] 3 S.C.R. 159
R. v. Ruzic, [2001] 1 S.C.R. 687

R, v. S(R.J.), [1995) 1 S.C.R. 451
R. v. Sawyer, [1992] 3 S.C.R. 809
R. v. Schwartz, [198812 S.C.R. 443
R. v. Seaboyer, R. v. Gayme, [1991] 2 S.C.R. 577
R. v. Sharpe, [2001] 1 S.C.R. 45
R. v Simmons, [1988] 2 S.C.R. 495
R. v. Sit, [1991] 3 S.C.R. 124

R. v. Skalbania, [199713 S.C.R. 995

R. v. Skinner, [1990] 1 S.C.R. 1235
R. v. Smith, [198711 S.C.R. 1045
R. v Swain, [1991] 1 S.C.R. 933
R. v. Thomsen, [1988] 1 S.C.R. 640
R. v. Turpin, [1989) 1 S.C.R. 1296

. v. Vaillancourt [198712 S.C.R. 636
R. v. Wholesale Travel Group, [1991] 3 S.C.R. 154
R. v. Whyte, [1988 2 SC.R.3
R. v. Yorke, [1993] 3 S.C.R. 647
R v Zundel, [1992] 2 S.C.R. 731

0
0

,

*

*

*

*

*

*

*

9

*

*

*

*

*

*

*

*

9

*

9

1
3
1
1
1
1
1
1
1
1

1
1
1
1
1
1
1

1
1
1

1
1
1
1

1

1

1

1
1
1

1
1
1
1

1

MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL

[Vol. 48

Counter-Majoritarlan Cases

Section 1 Override
Subset
Subset

Number of
Challenges

Coded

Ramsden v. Peterborough (City of), [199312 S.C.R. 1084
Re Therrien, [2001] 2 S.C.R. 3

Reference Re Manitoba Language Rights, [1985] 1 S.C.R. 721
Reference Re Motor Vehicle Act (British Columbia), s. 94(2), [1985] 2 S.C.R. 486
Reference Re Provincial Electoral Boundaries (Saskatchewan), [1991] 2 S.CR. 158

Reference Re Public Schools Act (Manitoba), ss 79(3), (4) and (7), [1993]
1 S.C.R. 839
Reference Re Remuneration of Judges of the Provincial Court of Prince Edward
Island, [1997] 3 S.C.R. 3
Reference Re ss. 193 and 195. l(1)(c) of the Criminal Code (Manitoba), [1990]
1 S.C.R. 1123
Reference Re Workers’ Compensation Act, 1983 (Newfoundland), ss. 32 & 34,
[1989] 1 S.CR. 922
RJR-MacDonald v. Canada (A.G.), [1995] 3 S.C.R. 199

Rodriguez v. British Columbia (A.G.), [1993] 3 S.C.R. 519

Ruby v. Canada (Solicitor General), 2002 SCC 75
Rudolph Wolff & Co. v. Canada, [1990] 1 S.C.R. 695

Ruffo v. Conseil de la magistrature, [1995] 4 S.C.R. 267
Sauv6 v. Canada (A.G., Belczowski v. Canada, [1993] 2 S.C.R, 438

Sauv6 v. Canada (Chief Electoral Officer), 2002 SCC 68

Schachter v. Canada, [1992] 2 S.C.R. 679

Suresh v. Canada (Minister of Citizenship and Immigration), [2002] 1 S.C.R. 3

Symes v. Canada, [1993] 4 S.C.R. 695
T6trault-Gadouty v. Canada (Employment & Immigration Commission), [1991]
2 S.C.R. 22
Thibaudeau v. Canada, [1995] 2 S.C.R. 627
Thomson Newspapers Ltd. v. Canada (Director of Investigation and Research,
Restrictive Trade Practices Commission), [1990] 1 S.C.R. 425
Thomson Newspapers Ltd. v. Canada (A.G.), [1998] 1 S.C.R. 877
United Food and Commercial Workers, Local 518 v. KMart Canada Ltd., [1999]
2 S.C.R. 1083
Vancouver Society of Immigrant and Visible Minority Women v. M.N.R., [1999]
1 S.C.R. 10

Vriend v. Alberta, [1998] 1 S.C.R. 493
Walker v. Prince Edward Island, [1995] 2 S.C.R. 407
Winko v. British Columbia (Forensic Psychiatric Institute), [1999] 2 S.C.R. 625
Winnipeg Child and Family Services v. K.L.W, [2000] 2 S.C.R. 519

Young v. Young, [1993] 4 S.C.R. 3

*

*

*

*

*

*

*

9

*

*

*

*

*

*

*

*

*

*

*

*

1

1

1

1

1

1

10

2

1

3

1

2

1

1

1

1 1

1

2

1
1

1

1

1

1

1

1

1
1
1

1

La titrisation de la propriété intellectuelle au Canada in this issue

related content