R. v. Bryan, 2007 SCC 12, [2007] 1 SCR 527

                                                 SUPREME COURT OF CANADA

Between:

Paul Charles Bryan

Appellant

and

Her Majesty the Queen and Attorney General of Canada

Respondents

‑ and ‑

Canadian Broadcasting Corporation, CTV Inc., TVA Group Inc.,

Rogers Broadcasting Limited, CHUM Limited, Sun Media

Corporation, Sun Media (Toronto) Corporation, Canadian

Press, Globe and Mail, CanWest MediaWorks Inc., CanWest

MediaWorks Publications Inc., Canoe Inc. and Canadian Civil

Liberties Association

Interveners

Coram: McLachlin C.J. and Bastarache, Binnie, LeBel, Deschamps, Fish, Abella, Charron and Rothstein JJ.

______________________________

R. v. Bryan, [2007] 1 S.C.R. 527, 2007 SCC 12

Paul Charles Bryan                                                                                          Appellant

v.

Her Majesty The Queen and Attorney General of Canada                      Respondents

and

Canadian Broadcasting Corporation, CTV Inc., TVA Group Inc.,

Rogers Broadcasting Limited, CHUM Limited, Sun Media

Corporation, Sun Media (Toronto) Corporation, Canadian

Press, Globe and Mail, CanWest MediaWorks Inc., CanWest

MediaWorks Publications Inc., Canoe Inc. and Canadian Civil

Liberties Association                                                                                     Interveners

Indexed as:  R. v. Bryan

Neutral citation:  2007 SCC 12.

File No.:  31052.

2006:  October 16; 2007:  March 15.

Present:  McLachlin C.J. and Bastarache, Binnie, LeBel, Deschamps, Fish, Abella, Charron and Rothstein JJ.

on appeal from the court of appeal for british columbia

Constitutional law — Charter of Rights  — Freedom of expression — Federal elections — Premature transmission of results — Federal elections legislation prohibiting dissemination of election results from one electoral district to another if polling stations in that other district still open — Whether temporary ban on publication of election results infringes freedom of expression — If so, whether infringement justifiable — Canadian Charter of Rights and Freedoms, ss. 1 , 2 (b) — Canada Elections Act, S.C. 2000, c. 9, s. 329 .

Constitutional law — Charter of Rights  — Reasonable limit — Sufficiency of evidence adduced by government to justify infringement of constitutional right — Canadian Charter of Rights and Freedoms, s. 1 .

During the 2000 federal election, B transmitted the election results from 32 ridings in Atlantic Canada while polling stations remained open elsewhere in Canada, by posting the information on a website.  He was charged with contravening s. 329  of the Canada Elections Act , which prohibits the transmission of election results in one electoral district to another electoral district before the close of all polling stations in that other district.  B’s application for a declaration that s. 329  was unconstitutional for unjustifiably infringing his freedom of expression guaranteed by s. 2 (b) of the Canadian Charter of Rights and Freedoms  was dismissed, and B was convicted of the offence.  The summary conviction appeal judge declared the provision unconstitutional on the ground that it infringed the Charter  right and was not saved by s. 1 , and overturned B’s conviction.  The Court of Appeal held that s. 329  was a justified limit on freedom of expression and restored the conviction.

Held (McLachlin C.J. and Binnie, LeBel and Abella JJ. dissenting):  The appeal should be dismissed.

Per Bastarache J.:  Although s. 329 of the Canada Elections Act  infringes freedom of expression, this infringement is justified under s. 1  of the Charter .  The true objective of the Canada Elections Act  in the context of the provisions under scrutiny is to ensure informational equality by adopting reasonable measures to deal with the perception of unfairness created when some voters have general access to information that is denied to others, and the further possibility that access to that information will affect voter participation or choices.  In determining the nature and sufficiency of the evidence required to justify an infringement of s. 2 (b) of the Charter , s. 329  must be viewed in its context.  Here, given that the harm associated with the loss of public confidence in the electoral process or with a breach of the principle of informational equality is difficult to measure, logic and reason assisted by some social science evidence could constitute sufficient proof of the harm.  Furthermore, the subjective perceptions of Canadian voters that the electoral system is fair is a vital element in the integrity of the electoral system.  Since prevention of Canadians’ subjective fears and apprehension of harm is a goal of s. 329 , evidence of those subjective fears must be taken as important.  As for the nature of the infringed activity, while political expression lies at the core of the guarantee of free expression,  the right at issue is the putative right to receive election results before the polls close; restricting access to such information before polls close carries less weight than after they close. [10] [14] [16] [19] [23] [25] [27] [30]

Section 329, by virtue of its objective of ensuring informational equality among voters, is a reasonable limit on s. 2 (b) of the Charter .  Under the first branch of the s. 1  analysis, the government has established that ensuring informational equality is a pressing and substantial objective on the basis of logic and reason applied to the evidence brought by the Attorney General.  Further, the Court has already determined in a previous decision that the objective of maintaining public confidence in the fairness of the electoral system is a pressing and substantial one. [35] [37] [53]

Under the second branch of the s. 1  analysis, the government has also demonstrated that the s. 329 ban meets the proportionality test.  Logic and reason, combined with the Lortie Commission’s report and the 2005 poll produced as fresh evidence sufficiently establish that s. 329 is rationally connected to the objective of maintaining public confidence in the electoral system.  To allow some voters to access the results of voting in other districts would patently violate that objective.  The suggestion that such information could be available to voters in spite of the ban, through private communications, does not detract from this conclusion.  With respect to minimal impairment, the Lortie Commission’s report also supports the view that maintaining public confidence in the electoral system requires some method of restraining publication of election results until most or all Canadians have voted.  Staggered hours cannot alone perfectly address the problem of voter confidence.  Parliament debated the advantages and disadvantages of various approaches to the problem, including the alternative options proposed by the Lortie Commission, and determined the s. 329 scheme to be the most effective and least intrusive.  There is also sufficient evidence in the particular context of this case showing that Parliament’s policy choice in adopting s. 329 is a rational and justifiable solution to the problem of informational imbalance.  Finally, the salutary effects of s. 329 outweigh the deleterious effects.  The salutary effects of s. 329 are clear.  Of primary importance is the fact that s. 329 maintains the integrity of the principle of informational equality, and is the only effective legislative response available to Parliament to address that objective.  Secondarily, logic and reason suggest that, given that 70 percent of Canadians surveyed in the 2005 poll believe in the importance of informational equality in elections, s. 329 contributes to the maintenance of public confidence in the electoral system.  These salutary effects are real, not merely potential.  Moreover, s. 329 contributes in a positive way to the fairness and reputation of the electoral system as a whole.  Regarding deleterious effects, there is no evidence that s. 329 harms the electoral process or the general right of Canadians to be informed that is manifestly superior to the evidence of its salutary effects in promoting electoral fairness.  The s. 329 ban is only operative for a matter of two to three hours, only on election day, and it is only the late voters who will be affected.  While the ban may be inconvenient for the media, this argument cannot be allowed to override as important a goal as the protection of Canada’s electoral democracy. [40-41] [45-52]

Per Fish J.:  Section 329  of the Canada Elections Act  is but one element of a comprehensive, elaborate electoral system that temporarily restricts various forms of expression.  Care must be taken not to usurp Parliament’s role in determining the rules of the electoral game most appropriate for Canada as a whole.  The role of the courts is simply to decide whether Parliament’s impugned preference passes constitutional muster.  Here, the government has discharged its burden under s. 1  of the Charter .  The object of s. 329 , which is to avoid or restrict the information imbalance that would otherwise result, relates to a pressing and substantial concern.  Without the limit imposed by s. 329 , voters in Central and Western Canada would have access to the results from the Atlantic provinces before casting their ballots, while voters in the Atlantic provinces would lack similar information.  The s. 329  ban also meets the proportionality test.  There is a rational connection between the objective and the limitation, and the limitation minimally impairs the affected Charter  right.  The deleterious effects of the limitation are also outweighed by its salutary effects.  The justification invoked by the government must be assessed in light of the inherent resistance of the relevant harm to precise measurement.  In this regard, the Lortie Commission’s report and the 2005 poll provide a sound basis for concluding that the information imbalance is a real and significant harm and that Canadians value the principle of information equality.  Further, they support the government’s assertion that the information imbalance alone creates a perception of unfairness in the electoral system, which is itself a harm that Parliament may address.  While modern communications technology diminishes the delay’s effectiveness, and thereby its salutary effects, s. 329  does curb widespread dissemination of this information and thus contributes materially to its objective of information equality between voters in different parts of the country.  The deleterious effects of the delay, in contrast, are minimal.  The s. 329  limitation on freedom of expression involves no suppression of any information at all, but only a brief delay in its communication to voters who have not yet cast their ballots.  Lastly, while political expression is at the core of the s. 2 (b) guarantee, restricting the publication of election results in other provinces before all votes have been cast carries less weight than would a similar restriction after the close of polls.  As most voters have no access to electoral results until after the close of polls, these results, before one has cast one’s own ballot, cannot form an important part of the political discourse. [58-62] [68] [71] [78-81]

Per Deschamps, Charron and Rothstein JJ.:  The analytical approaches adopted and the result reached by Bastarache and Fish JJ. were agreed with.  The appeal is resolved at the proportionality stage of the Oakes test, and in this regard, their reasons are complementary. [83]

Per McLachlin C.J. and Binnie, LeBel and Abella JJ. (dissenting):  The s. 329  publication ban is an excessive response to an insufficiently proven harm  and a violation of s. 2 (b) of the Charter  that cannot be justified under s. 1 .  The government’s s. 1  justification falters fatally in its submission that the benefits of the limitation on the freedom of expression are proportional to its harmful effects.  While the government may not be required to demonstrate that its policy judgment is justified with evidence that is amenable to precise measurement, the social science evidence supported by reason and logic must convincingly establish the consequences of imposing or failing to impose the limit.  Here, when the harm at which the blackout period in s. 329  is aimed is considered in the context of staggered hours, there is only speculative and unpersuasive evidence to support the government’s claim that the information imbalance is of sufficient harm to voter behaviour or perceptions of electoral unfairness that it outweighs any damage done to a fundamental and constitutionally protected right. [103] [106] [108] [110] [133]

At issue are the core democratic rights of the media to publish and of Canadians to receive election results in a timely fashion.  The possibility that some Western voters might be influenced by results from Atlantic Canada cannot be completely discounted, but the question is whether the impact will be a harmful one.  The inference that the information imbalance created by lifting the ban in s. 329  would have a harmful impact of any kind on voter turnout, choice or perception is highly theoretical and unsubstantiated by cogent evidence.  The evidence of the expert witness  at trial indicates that there is no impact on voter turnout unless the outcome of the election is known or knowable, something that cannot realistically occur based on knowing the results from 32 ridings.  Moreover, the evidence from the Lortie Commission suggests that there is no harm to public perception from knowing the results from the 32 ridings in Atlantic Canada.  The real harm perceived by the Commission — the influence on voter perception generated by knowing the Ontario and Quebec results, since these two provinces have the potential to determine who will form the government — is alleviated by staggered hours.  Further, any potential benefits of the publication ban are diminished by the reality that it has been rendered obsolete by telecommunications technology. [110] [117] [120-126]

The evidence adduced by the government fails to demonstrate either that the ban in s. 329 , in the context of staggered hours, promotes public confidence in the fairness of elections or that harm will occur in the absence of the limitation.  On the other hand, the s. 329  ban impairs the right both to disseminate and receive election results at a crucial time in the electoral process.  To suggest that the limitation at issue involves only a delay, not the suppression of information, unduly minimizes the significance both of the information and of the delay.  Canadians are entitled to know, as soon as possible, who their elected representatives are. [127-129] [131]

Furthermore, the media’s role in disseminating election information to the public is critical.  The technical challenges they face when implementing the ban are considerable, and mean that individuals in certain areas may not receive radio or television election coverage even though the polls have closed in their area of the country.  All of this harm to the Charter  right is demonstrable; the benefits of the ban are not. [130-131 ] [133]

Cases Cited

By Bastarache J.

Applied:  Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, 2004 SCC 33; referred to:  Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877; R. v. Oakes, [1986] 1 S.C.R. 103; Libman v. Quebec (Attorney General), [1997] 3 S.C.R. 569; Harvey v. New Brunswick (Attorney General), [1996] 2 S.C.R. 876; R. v. Wholesale Travel Group Inc., [1991] 3 S.C.R. 154; McKinney v. University of Guelph, [1990] 3 S.C.R. 229; Edmonton Journal v. Alberta (Attorney General), [1989] 2 S.C.R. 1326; R. v. Butler, [1992] 1 S.C.R. 452; R. v. Keegstra, [1990] 3 S.C.R. 697; RJR‑MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199; R. v. Malmo‑Levine, [2003] 3 S.C.R. 571, 2003 SCC 74.

By Fish J.

Referred to:  Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, 2004 SCC 33; R. v. Oakes, [1986] 1 S.C.R. 103; RJR‑MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199; Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877.

By Deschamps, Charron and Rothstein JJ.

Referred to:  R. v. Oakes, [1986] 1 S.C.R. 103.

By Abella J. (dissenting)

Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877; Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, 2004 SCC 33; R. v. Oakes, [1986] 1 S.C.R. 103; RJR‑MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199; Sauvé v. Canada (Chief Electoral Officer), [2002] 3 S.C.R. 519, 2002 SCC 68; McKinney v. University of Guelph, [1990] 3 S.C.R. 229; Dagenais v. Canadian Broadcasting Corp., [1994] 3 S.C.R. 835.

Statutes and Regulations Cited

Act to amend the Canada Elections Act , the Parliament of Canada Act  and the Referendum Act, S.C. 1996, c. 35, s. 44.1.

Canada Elections Act, S.C. 2000, c. 9, ss. 128 , 323 , 328 , 329 , 495(4) , 500(4) .

Canadian Charter of Rights and Freedoms, ss. 1 , 2 (b), (d).

Constitution Act, 1982, s. 52 .

Criminal Code, R.S.C. 1985, c. C‑46, s. 813 .

Dominion Elections Act, 1938, S.C. 1938, c. 46, s. 107.

Authors Cited

Canada.  House of Commons.  House of Commons Debates, vol. 134, 2nd Sess., 35th Parl., November 26, 1996, p. 6723.

Canada.  Royal Commission on Electoral Reform and Party Financing.  Final Report.  Reforming Electoral Democracy, vol. 2.  Ottawa:  The Commission, 1991.

Choudhry, Sujit.  “So What Is the Real Legacy of Oakes?  Two Decades of Proportionality Analysis under the Canadian Charter ’s Section 1 ” (2006), 34 S.C.L.R. (2d) 501.

Decima Research/Carleton University, School of Journalism and Communication.   “Most Canadians Prefer Election Night Results Blackout”, 2006.

Hogg, Peter W.  Constitutional Law of Canada, vol. 2, loose‑leaf ed.  Scarborough, Ont.:  Carswell, 1997 (updated 2006, release 1).

APPEAL from a judgment of the British Columbia Court of Appeal (Finch C.J.B.C., Rowles and Saunders JJ.A.) (2005), 253 D.L.R. (4th) 137, 213 B.C.A.C. 52, 352 W.A.C. 52, 196 C.C.C. (3d) 369, 130 C.R.R. (2d) 348, [2005] B.C.J. No. 1130 (QL), 2005 BCCA 285, reversing a decision of Kelleher J. (2003), 233 D.L.R. (4th) 745, 112 C.R.R. (2d) 189, [2003] B.C.J. No. 2479 (QL), 2003 BCSC 1499.  Appeal dismissed, McLachlin C.J. and Binnie, LeBel and Abella JJ. dissenting.

Donald J. Jordan, Q.C., and Rodney W. Sieg, for the appellant.

Andrew I. Nathanson and Brook Greenberg, for the respondent Her Majesty the Queen.

Graham Garton, Q.C., and Sean Gaudet, for the respondent the Attorney General of Canada.

Joseph J. Arvay, Q.C., Brent Olthuis and Daniel Henry, for the interveners Canadian Broadcasting Corporation, CTV Inc., TVA Group Inc., Rogers Broadcasting Limited, CHUM Limited, Sun Media Corporation, Sun Media (Toronto) Corporation, Canadian Press, Globe and Mail, CanWest MediaWorks Inc., CanWest MediaWorks Publications Inc. and Canoe Inc.

Mahmud Jamal and Colin Feasby, for the intervener the Canadian Civil Liberties Association.

The following are the reasons delivered by

Bastarache J. —

I.       Introduction

1                                   The sole issue on this appeal is whether s. 329  of the Canada Elections Act, S.C. 2000, c. 9 , which prohibits the broadcasting of election results on election day until polling stations are closed in all parts of Canada, constitutes a violation of the freedom of expression guaranteed by s. 2 (b) of the Canadian Charter of Rights and Freedoms  that cannot be justified under s. 1 .  As such, this case is a direct application of this Court’s decision in Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, 2004 SCC 33.

II.      Facts and Judicial History

2                                   During the federal general election of November 27, 2000, the appellant transmitted the election results from Atlantic Canada while polling stations remained open in other parts of Canada by posting the results on a website.  The results were therefore available to the public in every electoral district in Canada. The appellant had made his intention to do so public before the election, and the Commissioner of Canada Elections had warned him that such publication would be contrary to s. 329  of the Canada Elections Act .

3                                   The appellant was charged under s. 329 of the Act.  The appellant brought an application in the Provincial Court of British Columbia challenging the constitutional validity of ss. 329, 495(4) and 500(4) of the Act on the basis that the provisions infringed ss. 2 (b) and 2 (d) of the Charter , and that they were not saved by s. 1  of the Charter .  The Commissioner of Canada Elections, who was the respondent to the appellant’s application, took no position on the application, but the Attorney General of Canada appeared before the Provincial Court as an intervener to defend the constitutionality of s. 329 .

4                                   Judge Smith of the B.C. Provincial Court held, in two separate judgments ([2003] B.C.J. No. 542 (QL), 2003 BCPC 65, and (2003), 104 C.R.R. (2d) 364, 2003 BCPC 39), that s. 329  infringed s. 2 (b) of the Charter , but that the Attorney General had established that it was a demonstrable limit on s. 2 (b) and so was justified under s. 1 .  The trial judge based this second conclusion on two pieces of evidence.  One was a government report, Reforming Electoral Democracy (1991), the Report of the Royal Commission on Electoral Reform and Party Financing (the “Lortie Report”), and the other was the report and evidence of Dr. Robert MacDermid, a professor of political science at York University. Dr. MacDermid was qualified as an expert witness; this qualification was not challenged by the appellant.  It is also important to note that the appellant brought no evidence of his own.  The Lortie Report and the evidence of Dr. MacDermid remained essentially the only evidence in this case until I granted leave to adduce new evidence to the group of media organizations which intervened before this Court and subsequently to the Attorney General by Orders dated March 22, 2006 and July 25, 2006 respectively.  I will refer to the evidence in some detail in my analysis below.

5                                   The appellant’s constitutional application having been dismissed, a summary conviction trial took place on the basis of an agreed statement of facts.  The appellant admitted to having committed the elements of the offence.  He was convicted of the offence and fined $1,000.

6                                   Pursuant to s. 813  of the Criminal Code, R.S.C. 1985, c. C-46 , the appellant appealed the summary conviction on the basis that the trial judge erred in dismissing the constitutional challenge. That appeal was allowed by Kelleher J. of the B.C. Supreme Court on the basis that the evidence before the trial judge did not support the conclusion that s. 329  was supported by a pressing and substantial objective ((2003), 233 D.L.R. (4th) 745, 2003 BCSC 1499).  Kelleher J. also held that, if his conclusion regarding a pressing and substantial objective was incorrect, the Attorney General had failed to meet its burden of establishing minimal impairment and proportionality.  By judgment dated October 23, 2003, the conviction was overturned and the appellant was acquitted.

7                                   On May 18, 2004, this Court released its judgment in Harper.  In light of this Court’s ruling in that case that the Canada Elections Act ’s third party advertising limits were constitutional, the Attorney General sought, and received, leave to appeal the decision of Kelleher J.

8                                   The British Columbia Court of Appeal unanimously agreed that s. 329 of the Act violated s. 2 (b) of the Charter  and that the only real issue was whether s. 329  could be justified under s. 1  ((2005), 253 D.L.R. (4th) 137, 2005 BCCA 285).  A majority of the Court of Appeal held that the Attorney General had met its burden under s. 1  and that s. 329  was a justified limit on freedom of expression.  The conviction ordered at trial was restored.

III.    Analysis

9                                   This Court’s decision in Harper contains two important principles that are applicable to this case.  First, it establishes that courts ought to take a natural attitude of deference toward Parliament when dealing with election laws:  “Given the right of Parliament to choose Canada’s electoral model and the nuances inherent in implementing this model, the Court must approach the justification analysis with deference” (Harper, at para. 87).

10                               Second, it reaffirms that, in determining the nature and sufficiency of evidence required for the Attorney General to establish that a violation of s. 2 (b) is saved by s. 1 , the impugned provision must be viewed in its context: see Harper, at paras. 75-76, and Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877, at para. 88.  This context can be best established by reference to the four factors which this Court set out in Thomson Newspapers and Harper:  (i) the nature of the harm and the inability to measure it, (ii) the vulnerability of the group protected, (iii) subjective fears and apprehension of harm, and (iv) the nature of the infringed activity.

11                               However, the contextual factors must be understood as being about the provision.  As this Court noted at para. 87 of Thomson Newspapers:

The analysis under s. 1  of the Charter  must be undertaken with a close attention to context. This is inevitable as the test devised in R. v. Oakes, [1986] 1 S.C.R. 103, requires a court to establish the objective of the impugned provision, which can only be accomplished by canvassing the nature of the social problem which it addresses. Similarly, the proportionality of the means used to fulfil the pressing and substantial objective can only be evaluated through a close attention to detail and factual setting. In essence, context is the indispensable handmaiden to the proper characterization of the objective of the impugned provision, to determining whether that objective is justified, and to weighing whether the means used are sufficiently closely related to the valid objective so as to justify an infringement of a Charter  right. [Emphasis added.]

I take this passage to mean that only once the objectives of the impugned provision are stated can we turn to an examination of the context of those objectives to determine the nature and sufficiency of the evidence required under s. 1 .

12                               The Attorney General has asserted two distinct but related objectives of s. 329 .  First, the Attorney General claims that s. 329  is directed at ensuring informational equality among voters. The Attorney General’s expert witness in this case, Dr. MacDermid, refers to informational equality as “a central assumption of electoral democracy”.  The Attorney General has presented the promotion of informational equality as an inherently worthy goal; by this he means that the mere fact that one voter could have general access to information about election results that another voter does not have is in and of itself problematic.  The Attorney General’s second proposed objective is, as I said, related to the first: the Attorney General asserts that s. 329  promotes public confidence in the electoral system, and a principal reason that it does so is because such public confidence depends on the public’s belief that the principle of informational equality is maintained by the electoral system.

13                               The Attorney General argues that these two objectives, while causally quite closely linked, are in fact different in nature.  The first objective, ensuring informational equality, is presented as an inherently important goal: as such its existence and importance are to some degree conceptual matters, and evidence to establish them will be hard to come by.  The second objective, on the other hand, is presented as based on actual sociological fact: the Attorney General claims that confidence in the electoral system will (or would) actually decrease if some voters were to have general access to information about the results of the election that other voters cannot have. As this second objective is framed in terms of a sociological argument about the link between informational equality and voter confidence, it is the kind of objective for which the Attorney General should be expected to lead some evidence establishing the claimed link.

14                               As I see it, the true objective of the Canada Elections Act  in the context of the provisions under scrutiny is to ensure informational equality by adopting reasonable measures to deal with the perception of unfairness created when some voters have general access to information that is denied to others, and the further possibility that access to that information will affect voter participation or choices.  In Harper, this Court unanimously held that “ensuring that all voters receive the same information where possible”, was an important objective: see para. 47, per McLachlin C.J. and Major J. and para. 133, per Bastarache J. This objective is measurable to some degree, as we shall see.

15                               With this understanding of the objective of s. 329 , I now turn to a consideration of the four contextual factors from Harper and Thomson Newspapers to the facts of this case.

A.     Contextual Factors

1.    The Nature of the Harm and the Inability to Measure It

(a)   Maintaining Public Confidence in the Electoral System

16                               In Harper, this Court held, at para. 77, that when social science evidence of a harm is conflicting or inconclusive, “the court may rely on a reasoned apprehension of . . . harm”.  I noted in that case that, absent determinative social science evidence, logic and common sense could be relied upon to assist in the s. 1  analysis.  In Thomson Newspapers,  I relied on logic and common sense as an aid to interpretation of the uncertain social science evidence about the influence of polls on voters, and held that the possibility of such influence was a matter that the government was legitimately concerned to remedy:  see paras. 104-7.  Similarly in Harper,  I noted that the presence of several factors, such as the subtle influence of advertising on individual decision makers, the presence of other influencing factors and the complexity of electoral decisions, meant that the harm at issue there was difficult, if not impossible to measure and so concluded that “logic and reason assisted by some social science evidence [were] sufficient proof of the harm”:  see para. 79.

17                               The situation with respect to the maintenance of public confidence in the electoral system is no less complicated. Public confidence is important for instrumental reasons. As this Court noted at para. 82 of Harper:

Perception is of utmost importance in preserving and promoting the electoral regime in Canada. Professor Aucoin emphasized that “[p]ublic perceptions are critical precisely because the legitimacy of the election regime depends upon how citizens assess the extent to which the regime advances the values of their electoral democracy” (emphasis in original). Electoral fairness is key. Where Canadians perceive elections to be unfair, voter apathy follows shortly thereafter.

The Attorney General, following the Court’s logic in Harper, asserts that public confidence in the electoral system is important because, if public confidence is lost, voting patterns could change, and, ultimately, the outcomes of elections could be affected.

18                               The Attorney General did provide some evidence on this point. Dr. MacDermid’s testimony was as follows:

. . . I think, based on the evidence in the States, based on some partial aspects about Canada that the — that the result would be the same as we’ve seen in the United States, and that’s most — most clearly a decline in — in participation, a decline in voting rates. Also — it may also affect, differentially, partisans of different parties. And while these would not be enormous, as they are not in the United States, I think it’s important to note that we are talking about effects from one to five percent. That is certainly an ample number to change — to have a very important effect in close races, which can never be predicted from one election to another. [Emphasis added.]

Similarly, the Lortie Report based some of its recommendations on the “perception that elections are decided before [Western] voters have even finished voting” (Lortie Report,  vol. 2, at p. 83).

19                               The somewhat speculative nature of this evidence is not unexpected, as some form of the s. 329  ban has been in place since 1938.  Thus, the effect of the s. 329  ban on voting patterns and election results is almost impossible to measure.    I am therefore forced to resort to logic and common sense applied to the Attorney General’s evidence as proof of the harm of loss of public confidence in the electoral system as a result of premature release of results.

(b)   Ensuring Informational Equality

20                               In a series of cases on freedom of expression, this Court gradually reached the recognition that the paucity of social science evidence in some cases required that a reasonable apprehension of harm could be sufficient as a grounding to a s. 1  argument: see R. v. Butler, [1992] 1 S.C.R. 452, at p. 503; R. v. Keegstra, [1990] 3 S.C.R. 697, at pp. 768 and 776; RJR-MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199, at para. 137; Thomson Newspapers, at paras. 104-7. In Harper, the Court extended this line of reasoning to the realization that some harms are “difficult, if not impossible, to measure scientifically” (at para. 79), and that in such cases logic and common sense become all the more important. At least one commentator has suggested that the impetus for this move lies in the origins of the Oakes test itself (R. v. Oakes, [1986] 1 S.C.R. 103), in that the test was formulated without explicit contemplation of situations such as those discussed above, in which “cogent and persuasive” evidence does not exist:  see S. Choudhry, “So What Is the Real Legacy of Oakes? Two Decades of Proportionality Analysis under the Canadian Charter ’s Section 1 ” (2006), 34 S.C.L.R. (2d) 501. 

21                               Section 1 analysis must always be done in a manner consistent with the warning set out by Dickson C.J. in Oakes, at p. 136:

A second contextual element of interpretation of s. 1  is provided by the words “free and democratic society”. Inclusion of these words as the final standard of justification for limits on rights and freedoms refers the Court to the very purpose for which the Charter  was originally entrenched in the Constitution: Canadian society is to be free and democratic. The Court must be guided by the values and principles essential to a free and democratic society which I believe embody, to name but a few, respect for the inherent dignity of the human person, commitment to social justice and equality, accommodation of a wide variety of beliefs, respect for cultural and group identity, and faith in social and political institutions which enhance the participation of individuals and groups in society. The underlying values and principles of a free and democratic society are the genesis of the rights and freedoms guaranteed by the Charter  and the ultimate standard against which a limit on a right or freedom must be shown, despite its effect, to be reasonable and demonstrably justified. [Emphasis added.]

22                               In some cases, the objective asserted by the government will be largely a matter of the “values and principles essential to a free and democratic society”.  In such cases it may not be appropriate to require proof according to the usual civil requirements.  I believe that this is such a case. The Attorney General submits that informational equality is important in particular because democracy requires that no individual  should have a general access to information, unavailable to others, that can play a role in the exercise of  his own  right to vote.   It is thus a logically direct result of the requirement that elections be fair.

23                               In Harper, this Court unanimously held that “ensuring that all voters receive the same information where possible” was an important objective: see para. 47, per McLachlin C.J. and Major J. in dissent, but not on this issue, and para. 133, per Bastarache J.  The harm associated with a breach of that principle is not of the class of harms which are easily measured.

2.    The Vulnerability of the Group Protected

24                               Section 329 is broadly addressed at protecting the Canadian electoral system, which suggests that the group protected is Canadian voters taken as a whole.  The Attorney General also claims that Western voters in particular are protected, for it is Western voters who could be most directly influenced by the release of election results from the Atlantic provinces.  But, as we noted in Thomson Newspapers, Canadian voters “must be presumed to have a certain degree of maturity and intelligence” (para. 101).  The expression at issue in this case does not seek to influence voters as did the advertising in Harper.

3.    Subjective Fears and Apprehension of Harm

25                               As noted above, the subjective perceptions of Canadian voters that the electoral system is fair is a vital element in the value of the system.  The Attorney General, with leave,  provided fresh evidence before this Court, in the form of a 2005 Decima Research/Carleton University Poll, which supports the view that Canadians subjectively believe that informational equality is an important aspect of the electoral system:  70 percent of Canadians surveyed “thought people should not be able to know election results from other provinces before their polls close”.  Moreover, the Lortie Report stated that “western Canadian voters generally may feel that their vote counts for less if the election outcome has been determined before their votes are cast, and some may have decided not to vote for that reason”: vol. 2, at p. 83.  This evidence regarding the subjective views of Canadians must be taken as doubly important in a case such as this, where the harm that the law is seeking to address is itself about those very same subjective views.

4.    Nature of the Infringed Activity: Political Expression

26                               This Court has held that “there can be no question that opinion surveys regarding political candidates or electoral issues are part of the political process and, thus, at the core of expression guaranteed by the Charter ”: Thomson Newspapers, at para. 92.  The same logic can be applied in this case: election results are of fundamental importance in a free and democratic society.

27                               At the same time, to suggest that election results are an important political form of expression in the hands of those still to vote is to prejudge the entire s. 1  inquiry. Whether the s. 2 (b) interest in receiving or disseminating political information, or both,  is at the centre of this case,  it is not at all clear that that interest can supersede the value of the countervailing principle that no voter should have general access to information about the results of the election unavailable to others.  As we noted with respect to restrictions on referendum spending in  Libman v. Quebec (Attorney General), [1997] 3 S.C.R. 569, at para. 61:

Thus, while the impugned provisions do in a way restrict one of the most basic forms of expression, namely political expression, the legislature must be accorded a certain deference to enable it to arbitrate between the democratic values of freedom of expression and referendum fairness. The latter is related to the very values the Canadian Charter  seeks to protect, in particular the political equality of citizens that is at the heart of a free and democratic society. [Emphasis added.]

The same may be said of the restrictions in this case.

5.    Summary on Contextual Factors

28                               In Harper, I referred to the contextual factors as favouring a “deferential approach to Parliament”: see para. 88. However, in my view the concept of deference is in this context best understood as being about “the nature and sufficiency of the evidence required for the Attorney General to demonstrate that the limits imposed on freedom of expression are reasonable and justifiable in a free and democratic society”: Harper, at para. 75 (emphasis added).   What is referred to in Harper and Thomson  Newspapers as a “deferential approach” is best seen as an approach which accepts that traditional forms of evidence (or ideas about their sufficiency) may be unavailable in a given case and that to require such evidence in those circumstances would be inappropriate.

29                               As Professor Choudhry aptly notes, at p. 524 of his paper:

Public policy is often based on approximations and extrapolations from the  available evidence, inferences from comparative data, and, on occasion, even educated guesses. Absent a large-scale policy experiment, this is all the evidence that is likely to be available. Justice La Forest offered an observation in [McKinney v. University of Guelph, [1990] 3 S.C.R. 229, at p.  304] which rings true: “[d]ecisions on such matters must inevitably be the product of a mix of conjecture, fragmentary knowledge, general experience and knowledge of the needs, aspirations and resources of society”.

I agree with this assessment.  The contextual factors are essentially directed at  determining to what extent the case before the court is a case where the evidence will rightly consist of “approximations and extrapolations” as opposed to more traditional forms of social science proof, and therefore to what extent arguments based on logic and reason will be accepted as a foundational part of the s. 1  case.

30                               Turning to a weighing of the factors in this case, I note that vulnerability does not play a major part in the analysis, but in light of the fact that prevention of Canadians’ subjective fears and apprehension of harm is a goal of s. 329 , evidence of those subjective fears must be taken as important.  While political expression is undoubtedly important, the right at issue is the putative right to receive election results before the polls close; restricting access to such information before polls close carries less weight than after they close. Furthermore, it has not been established that a right to such information, which is at the periphery of the s. 2 (b) guarantee, has been breached. In my view, that is the very question at issue.

B.      The Section 1  Analysis

31                               The s. 329  limit is clearly a limit prescribed by law; this was not disputed before us. I turn now to a consideration of the Oakes factors.

1.    Pressing and Substantial Objective

32                               The law is clear that the first stage of the s. 1  analysis is not an evidentiary contest.  As my colleagues recognized in Harper, “the proper question at this stage of the analysis is whether the Attorney General has asserted a pressing and substantial objective”:  Harper, at para. 25, per McLachlin C.J. and Major J.  (emphasis in original). McLachlin C.J. and Major J. went on to note that “[a] theoretical objective asserted as pressing and substantial is sufficient for purposes of the s. 1  justification analysis”: see para. 26. They based this assumption on several cases decided by this Court. A brief review of those cases is instructive at this point.

33                               In Thomson Newspapers, the Court accepted as “clearly pressing and substantial” the goals of s. 322.1  of the Canada Elections Act  on the basis that they were “directed to the realisation of the important collective goal of safeguarding the integrity of the electoral process” (para. 38 (emphasis added)).  In Harvey v. New Brunswick (Attorney General), [1996] 2 S.C.R. 876, this Court held that maintaining and enhancing the integrity of the electoral process was without doubt “always of pressing and substantial concern in any society that purports to operate in accordance with the tenets of a free and democratic society” (para. 38 (emphasis added)).  In R. v. Wholesale Travel Group Inc., [1991] 3 S.C.R. 154, Iacobucci J. (for the majority on the s. 1  issue) based his analysis on “the fundamental importance of the legislative objective as stated” (p. 259).  Similarly, the Court in McKinney v. University of Guelph, [1990] 3 S.C.R. 229, accepted the asserted objectives as valid on the basis that they were “admirable aim[s]” (p. 281).  Finally, in Edmonton Journal v. Alberta (Attorney General), [1989] 2 S.C.R. 1326, this Court accepted that protection of individual privacy was a pressing and substantial objective on the basis that our society has “cherished and given protection to” it, and on the basis of the Court’s own previous statements regarding the importance of privacy: see pp. 1343-45.

34                               In each of these cases the Court recognized that some objectives, once asserted, can be simply accepted by the Court as always pressing and substantial in any society that purports to operate in accordance with the tenets of a free and democratic society. 

35                               The Attorney General claims that informational equality is a fundamental principle of electoral democracy, as does its expert witness Dr. MacDermid.    It is a centrally important element of the concept of electoral fairness, and one which this Court has held to be “a laudable objective that will necessarily involve certain restrictions on freedom of expression”, and “a pressing and substantial objective in our liberal democracy, even in the absence of evidence that past elections have been unfair”: see Libman, at para. 84 (emphasis deleted), and Harper, at para. 26 (per McLachlin C.J. and Major J.).  I accept that it is pressing and substantial on the basis of logic and reason applied to the evidence brought by the Attorney General.

36                               As well, the Lortie Report suggests that Canadians “feel very strongly about premature release of election results” in general, and not just because the outcome of the election might be known in the West before voting closes there (vol. 2, at p. 84). Moreover, in some elections, such as the one which took place in 1993, the outcome was in fact seen to be predictable on the basis of Atlantic results alone.  Public confidence, as I have mentioned, is a complicated phenomenon which cannot easily be linked to any single cause,  but the link between confidence and information imbalances is clear.

37                               Most importantly, at this stage of the Oakes analysis, the inquiry is still into whether or not a pressing and substantial objective has been asserted by the Attorney General.  In Harper, the Court accepted that maintenance of public confidence in the electoral system was a pressing and substantial objective:

Maintaining confidence in the electoral process is essential to preserve the integrity of the electoral system which is the cornerstone of Canadian democracy. In R. v. Oakes, [1986] 1 S.C.R. 103, at p. 136, Dickson C.J. concluded that faith in social and political institutions, which enhance the participation of individuals and groups in society, is of central importance in a free and democratic society. If Canadians lack confidence in the electoral system, they will be discouraged from participating in a meaningful way in the electoral process. More importantly, they will lack faith in their elected representatives. Confidence in the electoral process is, therefore, a pressing and substantial objective. [para. 103]

I see no reason to resile from that position. I therefore accept the objective of the provision as pressing and substantial.

2.    Proportionality

38                               The inquiry into proportionality which consists in the final three stages of the Oakes test requires the Attorney General to provide more than the assertions which were acceptable at the first stage.  Instead, the inquiry is led into questions of causation and  may require more in the way of proof. 

(a)   Rational Connection

39                               The rational connection stage of the test requires the Attorney General to “show a causal connection between the infringement and the benefit sought on the basis of reason or logic”: see RJR-MacDonald, at para. 153, and Harper, at para. 104.  It is clear that logic and reason may play a large role in establishing such a causal connection. Such a connection is “often a difficult matter to establish by evidence, and the Supreme Court of Canada has not always insisted on direct proof of the causal relationship”: P. W. Hogg, Constitutional Law of Canada (loose-leaf ed.), vol. 2, at p. 35-31, cited with approval in Thomson Newspapers, at para. 39.

40                               In this case, the causal connection is eminently clear: allowing some voters to access the results of voting in other districts would patently violate the objective. The suggestion that such information could be available to voters in spite of the ban, through private communications, does not detract from this conclusion, for at least three reasons. First, at least 75 percent of Canadians surveyed for the Decima Research/Carleton University Poll mentioned above said that they would be unlikely or not at all likely to seek out such information in spite of the ban. Second, perfect enforcement is not a requirement of a law’s validity: R. v. Malmo-Levine, [2003] 3 S.C.R. 571, 2003 SCC 74, at paras. 177-78.  Finally, such an argument ignores the fundamental qualitative difference between, on the one hand, a small number (less than 25 percent) of voters seeking out information privately, and, on the other hand, a national broadcast of election results from Atlantic Canada which would be difficult to ignore for those who did not wish to see it, and would tend to cause a much more significant violation of the principle of information equality.

41                               The Attorney General has provided some evidence that public confidence depends on a perception that all Canadians have equal access to information before voting, and thus on the presence of the s. 329  ban.  This evidence includes the Lortie Report, which found that “Canadians feel very strongly about premature release of election results” (vol. 2, at p. 84), and the Decima Research/Carleton University Poll, which found that 70 percent of those surveyed believed in the principle of informational equality, suggesting that a failure to adhere to this principle would harm their view of the electoral system.  Clearly this evidence is not conclusive, but the Attorney General is not required to demonstrate an “empirical connection” between the objective and the provision: Harper, at para. 104.  Logic and reason, combined with the evidence that is available, establish that s. 329  is rationally connected to maintaining public confidence in the electoral system.

(b)   Minimal Impairment

42                               The standard for this stage of the analysis is still best encapsulated by the well-known passage from RJR-MacDonald, at para. 160:

The impairment must be “minimal”, that is, the law must be carefully tailored so that rights are impaired no more than necessary. The tailoring process seldom admits of perfection and the courts must accord some leeway to the legislator. If the law falls within a range of reasonable alternatives, the courts will not find it overbroad merely because they can conceive of an alternative which might better tailor objective to infringement . . . .

Of course, as this Court unanimously recognized in Harper, the minimal impairment analysis may be the stage of the Oakes test in which context is most important: see paras. 33 and 110.

43                               In Harper, I suggested at para. 111 that “the contextual factors indicate that the Court should afford deference to the balance Parliament has struck between political expression and meaningful participation in the electoral process”.  As noted above, deference in this context does not mean that Parliament’s decisions will be approved by this Court without scrutiny; rather, the contextual approach to s. 1  suggests that in some cases logic and reason will constitute appropriate supplements to what evidence there is.

44                               The most important evidence at this stage of the analysis is the Lortie Report and its recommendations. As noted above, the Lortie Report contained clear support for the proposition that information imbalances were considered problematic by Canadians:

Our research indicated that Canadians feel very strongly about premature release of election results and favour changes in voting hours to eliminate the problem. [Emphasis added; vol. 2, at p. 84.]

45                               The Lortie Report, therefore, supports the proposition that maintaining public confidence in the electoral system requires some method of restraining publication of election results until most or all Canadians have voted. The proposition that one effective way of achieving this objective is the staggered voting hours system (which was instituted by Parliament, even though it differs from the recommendation of the Lortie Report) is part of the Lortie Report’s recommendations. The Commission itself considered alternative recommendations, its goal being to prevent the release of election results before the close of polls.  It looked at uniform voting hours throughout the country, delaying the vote count and extending the voting period to two days, but found these possibilities to be “too disruptive for voters or election workers” (vol. 2, at p. 85). The Commission settled on partially staggered hours.  In proposing this compromise solution, it concluded that “the release of some election results before polls close in the West — specifically, results from the 32 seats in Atlantic Canada — would not constitute a major problem so long as other results from eastern Canada were not available until after the polls closed in the West” (p. 85). Parliament debated the alternatives, the advantages and disadvantages of various approaches to the problem; it then devised its own solution.

46                               Public confidence in the electoral system is dependent not only on the belief of Canadians that the election is fair in that the premature availability of returns does not affect the outcome of the election, but also on the belief in that the principle of information equality is upheld.  The Attorney General’s expert spoke of outcomes in this fashion: “[F]or there to be an effect on people, to depress turnout, there must be some knowledge of a different outcome for the election than they expected.”  He had also suggested that a media call was required to affect voter behaviour, but explained on re-examination that voters might well draw the same conclusion on their own.  Similarly, the Lortie Report concluded that “the basic problem is ensuring that voters in western Canada do not know who will form the government before the polls close there” (vol. 2, at p. 85).  This is not to say that Canadians will consider that the process is seen as perfectly fair unless the early release of electoral results in Eastern Canada is such that Western voters are convinced that the outcome of the election is all but certain.

47                               Public confidence depends on several factors.  The staggered hours solution addresses imperfectly, as noted, one of these factors — the “basic problem” of Western voters knowing who will form the government — but does not address the impact on confidence of the public in light of its knowledge that the principle of information equality is not being respected.  Thus, staggered hours, even if more extensive, as proposed in the Lortie Report, cannot alone perfectly address the problem of voter confidence.  Informational imbalance remains if s. 329 of the Act is not retained. Restricting all publication of Atlantic results  for a very short period is thus a reasonable way of trying to protect that public confidence, as agreed by Parliamentarians themselves.  Parliament considered the alternative options proposed in the Lortie Report and determined the s. 329  scheme to be the most effective and least intrusive; there is sufficient evidence in the particular context of this case showing that the policy choice of Parliament is a rational and justifiable solution to the problem of informational imbalance.

(c)   Salutary and Deleterious Effects

48                               The final stage of the Oakes analysis requires a balancing between the salutary and deleterious effects of the legislation.  At this stage, it is important to note that it is inappropriate to require a greater standard of proof for the existence of the salutary effects of the legislation than for the deleterious effects. 

49                               The salutary effects of s. 329  are clear.  Of primary importance is the fact that s. 329  maintains the integrity of the principle of informational equality, and is the only effective legislative response available to Parliament to address that objective.  Fully staggered hours were found to be unrealistic; the more extensive staggering proposed by the Lortie Commission was found to be inappropriate after a full review of the matter  before a Parliamentary committee. Secondarily, logic and reason suggest that, given that 70 percent of Canadians believe in the importance of informational equality in elections, s. 329  contributes to the maintenance of public confidence in the electoral system.  These salutary effects are real, not merely potential, as was the case in Thomson Newspapers (see para. 94).  The effect of the s. 329  ban is to actually protect the principle of informational equality, and to actually maintain public confidence in the electoral system.  In addition, it is  certainly plausible that s. 329  has the effect of preventing some voters from actually relying on prematurely released results and thus changing voting patterns in undesirable ways: the Attorney General’s expert suggested from one to five percent of voters could decide not to vote at all based on these results. 

50                               Finally, s. 329 contributes in a positive way to the fairness and reputation of the electoral system as a whole, a pillar of the Canadian democracy.   Harper established that s. 350  of the Canada Elections Act  constitutionally limits third party election advertising spending to promote electoral fairness. Moreover, electoral fairness takes on a special degree of importance on election day. The Canada Elections Act  bans the publication of election day polls (s. 328 ), including exit polls, and bans the transmission of election advertising on polling day before the closing of all the polling stations in the electoral district (s. 323 ). The s. 329  ban of premature disclosure of election results illustrates Parliament’s consistent efforts to particularly promote electoral fairness and minimize information imbalances on election day, an objective that is supported by the public at large (Decima Research/Carleton University Poll).  In my view, lifting the publication ban clearly impacts detrimentally on the pressing and substantial objective accepted in Harper, that of informational equality.  I would therefore question the proportionality analysis of Abella J. because she ignores the stated objective she has herself accepted and seems to ignore yet again the contextual and deferential approach contemplated in the recent jurisprudence of this Court, with which she has agreed.  To advance that Canadians have the right to know the election results as soon as possible (reasons of Abella J., at para. 129) begs the question.  The issue is whether the immediate publication of results is mandated, whether such a right trumps any other right or goal in light of the nature and quality of the information and its importance in the electoral context.

51                               Regarding the existence of the deleterious effects of s. 329, there is no evidence of harm to the electoral process or to the general right of Canadians to be informed that is of a quality or character manifestly superior to the evidence of the existence of salutary effects in promoting electoral fairness.  The s. 329 ban is only operative for a matter of two to three hours, and only on election day; it is only the late voters in fact that will be affected. The magnitude of the ban, it must be noted, is extremely small. The nature of the information has already been discussed. The submission of the media interveners that the ban may be ineffective and that it is very inconvenient for them (as it requires them to modify their broadcasting practices) cannot be allowed to override as important a goal as the protection of Canada’s electoral democracy.

52                               Thus the salutary effects of s. 329 outweigh the deleterious effects.

IV.    Conclusion

53                               Section 329, by virtue of its objective of ensuring informational equality among voters, is a reasonable limit on s. 2 (b) of the Charter .  I would dismiss the appeal.

The following are the reasons delivered by

Fish J. —

I

54                               Parliament has adopted two measures that together aim to ensure, insofar as reasonably possible, that electors across the country will cast their ballots without knowing how others have voted before them. 

55                               First, the voting hours have been staggered to reduce the impact of the way the world turns — and the resulting division of Canada into six separate time zones; second, election results anywhere in Canada cannot be published in an electoral district in which the polls have not yet closed.

56                               This appeal concerns the second of these complementary measures, which necessarily delays for no more than three hours the publication of results in the Atlantic provinces.  In Quebec, Ontario, the Prairie provinces, Alberta and a small part of British Columbia, publication of the results from the Maritime provinces is delayed by two hours.  Elsewhere in British Columbia, the delay is two and a half hours.  And in both instances, results from Newfoundland are delayed a half hour longer.

57                               It is conceded that this delay, though of short duration, limits the freedom of expression protected by s. 2 (b) of the Canadian Charter of Rights and Freedoms .  The decisive issue is whether this limit is constitutionally sanctioned by s. 1  of the Charter . 

58                               In addressing this issue, I find it important to bear in mind from the outset that we are dealing here with one element of a comprehensive electoral system that temporarily restricts various forms of expression, including exit polls and the ban on election day advertising upheld unanimously by the Court only recently in Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, 2004 SCC 33.  In this context, we must be particularly careful not to usurp Parliament’s role in determining the rules of the electoral game most appropriate for Canada as a whole.  And we must avoid any temptation driven by mere preference, even for objective reasons carefully explained, to tamper with those rules unless they run afoul of Canada’s constitutional requirements.

59                               For electoral arrangements of this kind, when Parliament prefers, the courts defer — except where the Constitution otherwise dictates.  Judicial deference, however, should not be mistaken for diminished constitutional vigilance, still less for judicial approval or entrenchment.  Our role is simply to decide whether Parliament’s impugned preference passes constitutional muster.  In finding that it does here, I take care to add that Parliament can of course change its mind.  Within constitutional bounds, policy preferences of this sort remain the prerogative of Parliament, not of the courts.

60                               For the reasons already given and those that follow I would, like Justice Bastarache, dismiss the appeal.

II

61                               Pursuant to s. 1  of the Charter , freedom of expression is subject in Canada only to “such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society”.  Two main criteria must be satisfied.  First, no limitation of a Charter  right can be saved by s. 1  unless its object relates to a pressing and substantial concern in a free and democratic society; second, the limitation must satisfy the proportionality requirements articulated in Oakes and refined in subsequent decisions of the Court (R. v. Oakes, [1986] 1 S.C.R. 103).

62                               I agree with Justices Bastarache and Abella that the limit in issue here easily satisfies the first requirement.  Its object is to avoid or, if that is not possible, to restrict the information imbalance that would otherwise result.  Without the limit, voters in Central and Western Canada would have access before casting their ballots to election results in the Atlantic provinces, while voters in the Atlantic provinces would lack any information of the sort.  In the government’s view, this information imbalance would create an actual or, at the very least, a perceived element of unfairness in Canada’s electoral system.

63                               In turning to the second requirement — proportionality — I emphasize that we are dealing here with an important element of Canada’s electoral system.  Dictated in large measure by Canada’s geography, the impugned provision forms part of the elaborate statutory scheme crafted by Parliament in response to the findings and  recommendations of the Royal Commission on Electoral Reform and Party Financing (the “Lortie Commission”), Reforming Electoral Democracy (1991). 

64                               The proportionality inquiry raises three questions.  The first is whether there is a “rational connection” between the pressing and substantial concern asserted by the government and the limitation Parliament has adopted to meet it.  Like my colleagues, I would without hesitation answer that question in the affirmative.

65                               The second question is whether the limitation minimally impairs the affected Charter  right.  I recognize, of course, that the burden of justification under s. 1  of the Charter  falls on the government and not on those whose rights are curtailed.  This applies as much to the minimal impairment criterion as to the others.  Nonetheless, I think it right to mention that those who would strike down the publication delay under attack here have not identified a reasonable and more minimalist alternative to the limitation, however imperfect, chosen by Parliament.  Justice Abella finds it unnecessary to do so; Justice Bastarache, on the other hand, demonstrates in his careful treatment of the issue that we are concerned in this case with a limitation that satisfies the “minimal impairment” requirement of Oakes, and I find it unnecessary to add anything more to his reasons on this point.

66                               This brings me to the third and final question on the proportionality branch of the matter.  And that question is whether the deleterious effects of the limitation are outweighed by its salutary impact.  In this regard, I emphasize that the limitation on freedom of expression in issue here involves no suppression of any information at all, but only a brief delay in its communication to voters who have not yet cast their ballots.  If this delay has any bearing on the election result, then it is to that extent effective in addressing the information imbalance that is conceded to be a pressing and substantial concern; if it has no bearing on the result, it at least has the merit of addressing the perception of unfairness caused by the information imbalance alone.

67                               The government asserts that the purpose of s. 329  of the Canada Elections Act, S.C. 2000, c. 9 , is to avoid the harm of information imbalance and the attendant perception of unfairness in the electoral process.  The scientific evidence led by the government in support of this assertion is, however, inconclusive on its own.  And the government, I repeat, is bound when it invokes s. 1  of the Charter , as it does in this case, to demonstrate that Parliament’s limitation of a constitutionally protected right or freedom is justified in a free and democratic society such as ours.  Mere assertion will not suffice.  The government must provide “a reasoned demonstration of the good which the law may achieve in relation to the seriousness of the infringement” (RJR-MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199, at para. 129).

68                               In Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877, and again in Harper, the Court recognized that “[t]he context of the impugned provision determines the type of proof that a court will require of the legislature to justify its measures under s. 1 ” (Harper, at para. 76).  I agree with Justice Bastarache that the predominant contextual factor in this case is the nature of the harm and its inherent resistance to precise measurement.  The justification invoked by the government must be assessed in this light.

69                               Absent definitive scientific evidence of the harms asserted by the government in Harper and Thomson Newspapers, the Court in both cases considered it proper to rely on “logic, reason and some social science evidence in the course of the justification analysis” (Harper, at para. 78; to the same effect, Thomson Newspapers, at paras. 104-7).  That approach is likewise appropriate here.

III

70                               The Lortie Commission found that “Canadians feel very strongly about premature release of election results and favour changes in voting hours to eliminate the problem” (vol. 2, at p. 84).  The Commission recognized that the problems associated with information imbalance were not limited to some voters knowing the results elsewhere before voting.  As Justice Bastarache mentions, it considered various remedial alternatives including uniform voting hours throughout the country, delaying the vote count and extending the voting period to two days.  These contemplated measures would have prevented all voters from obtaining any election results before the close of polls everywhere, but were found by the Commission to be “too disruptive for voters or election workers” (vol. 2, at p. 85).  The Commission therefore recommended the adoption of partially staggered hours.

71                               After due consideration, Parliament adopted instead a blended approach that supplemented the existing publication delay with partially staggered hours across different time zones.  Both measures target the information imbalance between electoral districts, and the findings of the Lortie Commission provide a sound basis for concluding that this information imbalance is a real and significant harm.

72                               According to a 2005 Decima Research/Carleton University Poll produced as fresh evidence, 70 percent of the Canadians surveyed “thought [that] people should not be able to know election results from other provinces before their polls close”.  I agree with Justice Abella that this evidence would have been more persuasive had the context of staggered hours been explained.  But the fact that a stronger case could have been made does not mean that the case made fails.

73                               In short, the Decima poll stands as uncontradicted evidence that Canadians value the principle of information equality.  And the Court has held that “[p]erception is of utmost importance in preserving and promoting the electoral regime in Canada” (Harper, at para. 82).  Accordingly, the Decima evidence does have probative value and cannot be rejected or disregarded on the ground that it could have been more probative still.

74                               The government also called an expert witness, Dr. Robert MacDermid, a professor of political science at York University.  As Justice Bastarache notes, Dr. MacDermid characterizes information equality as “a central assumption of electoral democracy”.  Dr. MacDermid also testified that knowledge of early election results could have a negative impact on voter behaviour.

75                               It is true that Dr. MacDermid, on cross-examination, appeared to concede that a media prediction (or “call”) of the ultimate election outcome is a necessary condition for results elsewhere to affect the voting behaviour of those who have yet to cast their ballots.  On re-examination, however, Dr. MacDermid later explained that voters might well draw their own conclusions in the absence of a formal media call, and that the effect could therefore prove to be the same.

76                               Finally, Dr. MacDermid testified that if the publication delay were removed, broadcast networks would endeavour to construct models upon which election results could be predicted, based on Atlantic results alone.  This, too, is evidence that the publication delay has a salutary effect and that its repeal or annulment might well have an adverse consequence.

77                               Again, I agree that the outcome in 32 Atlantic ridings does not a government make.  Nonetheless, its premature disclosure might well determine how — or whether — electors elsewhere will vote, particularly where the results in the 32 ridings cause surprise or raise subjective concerns.  Indeed, the media interveners in this case, who oppose the delay, take for granted that knowledge of early results will have some impact on those who have access to them.  Those who have not yet voted, it is argued, are entitled to this information for the purpose of deciding how to cast their ballots — or, put differently, in order to make the most informed choice.

78                               Moreover, even were we to assume that the publication of early results would have no impact at all on those who have not yet voted, the Lortie Commission’s report and the Decima poll support the government’s assertion that the information imbalance alone creates a perception of unfairness in the electoral system, which is itself a harm that Parliament may address.

79                               I recognize, of course, that modern communications technology diminishes the delay’s effectiveness and thereby its salutary effects.  Section 329  cannot and does not entirely prevent voters in Central or Western Canada who are determined to learn before casting their ballots what has transpired in the Atlantic provinces from obtaining that information by telephone or e-mail, for example.  But it does, at the very least, curb widespread dissemination of this information and it contributes materially in this way to its objective — information equality between voters in different parts of the country.

80                               The deleterious effects of the delay, in contrast, are minimal.  As I mentioned at the outset, the limitation on freedom of expression in issue here involves no suppression of any information at all, but only a brief delay in its communication to voters who have not yet cast their ballots.  This delay is of three hours’ duration at most — a delay far shorter than the 20-hour embargo on political advertising which this Court unanimously upheld in Harper.

81                               Finally, while political expression is undoubtedly at the core of the s. 2 (b) guarantee, restricting the publication of information regarding election results in other provinces before all votes have been cast carries less weight than would a similar restriction after the close of polls.  Most voters have no access to results in their own electoral districts or elsewhere until after the polls have closed.  How, then, can those results, before one has cast one’s own ballot, form “an important part of the political discourse”? (Thomson Newspapers, at para. 91).  This does not mean that their publication enjoys no protection at all; but it does signify that a brief delay in the communication of results from the 32 Atlantic ridings carries with it only a minor deleterious consequence, outweighed by its salutary effects.

IV

82                               For these reasons, I agree with Bastarache J. and with the majority of the Court of Appeal that the government has discharged its burden under s. 1  of the Charter .  Accordingly, I would dismiss the appeal.

The following are the reasons delivered by

83                               Deschamps, Charron and Rothstein JJ. — We agree with the analytical approaches Bastarache and Fish JJ. adopt and concur in the result. The appeal is resolved at the proportionality stage of the Oakes test (R. v. Oakes, [1986] 1 S.C.R. 103), and in this regard, we find their reasons to be complementary. We would dismiss the appeal.

The reasons of McLachlin C.J. and Binnie, LeBel and Abella JJ. were delivered by

84                               Abella J. (dissenting) — The Canada Elections Act  prohibits the dissemination of voting results from one electoral district to another if the polling stations in that other district are still open.  This means that the election results from the 32 ridings in the Atlantic region cannot be communicated to other parts of Canada before the polls close there, resulting in a one and a half to three hour media blackout period.  The issue is whether this ban unjustifiably infringes the right to freedom of expression guaranteed by s. 2 (b) of the Canadian Charter of Rights and Freedoms .

Background

85                               In 1938, Parliament penalized the public transmission of election results to electoral districts where the polls were still open by enacting s. 107 of the Dominion Elections Act, 1938, S.C. 1938, c. 46.  This ban was imposed because Canadian electoral districts span multiple time zones and, without the ban, voters in the West could learn the likely result of the election before they went to the polls.

86                               Just over 50 years later, a Royal Commission was appointed in 1989 to review Canada’s electoral system.  Known as the Lortie Commission, it released its report in 1991 (Royal Commission on Electoral Reform and Party Financing, Reforming Electoral Democracy).  Of most relevance to the issue in this appeal, the Lortie Commission acknowledged that it was important to ensure that voters in Western Canada not know who will form the government before they have had a chance to vote.  The Commission concluded, however, that the existing remedy for this problem — the publication ban — had “been rendered obsolete by developments in broadcasting and telecommunications technology” (vol. 2, at p. 84).

87                               In place of a publication ban, therefore, the Commission recommended that Parliament adopt staggered voting hours.  Although there would still be some time between the close of Atlantic and Western polls even with staggered hours, the Commission concluded:

The challenge is . . . to devise a solution that responds to concerns in western Canada, is fair to different groups and regions, and is not too disruptive for voters or election workers. We believe this can be achieved if we recognize that the basic problem is ensuring that voters in western Canada do not know who will form the government before the polls close there.  This means guarding against premature release of election results from Ontario and Quebec, whose 174 constituencies constitute more than half the seats in the House of Commons. We have concluded that the release of some election results before polls close in the West — specifically, results from the 32 seats in Atlantic Canada — would not constitute a major problem so long as other results from eastern Canada were not available until after the polls closed in the West.

The time difference between the eastern and Pacific time zones is three hours, and it takes about half an hour for poll workers to begin to report results once voting ends. . . . [Emphasis added; vol. 2, at p. 85.]

88                               In 1996, in response to this recommendation, Parliament  adopted a system of staggered voting hours (An Act to amend the Canada Elections Act, the Parliament of Canada Act  and the Referendum Act, S.C. 1996, c. 35, s. 44.1).  They are a slight variation on the actual staggered hours recommended by the Lortie Commission, and are now set out in s. 128  of the Canada Elections Act, S.C. 2000, c. 9 :

128. (1) The voting hours on polling day are

(a)     from 8:30 a.m. to 8:30 p.m., if the electoral district is in the Newfoundland, Atlantic or Central time zone;

(b)     from 9:30 a.m. to 9:30 p.m., if the electoral district is in the Eastern time zone;

(c)      from 7:30 a.m. to 7:30 p.m., if the electoral district is in the Mountain time zone; and

(d)     from 7:00 a.m. to 7:00 p.m., if the electoral district is in the Pacific time zone.

89                             The ban on the publication of election results formerly contained in s. 107 of the Dominion Elections Act was, however, retained.  It is currently found in s. 329  of the Canada Elections Act  and states:

329.  No person shall transmit the result or purported result of the vote in an electoral district to the public in another electoral district before the close of all of the polling stations in that other electoral district.

Allowing for the half hour minimum that the Lortie Commission identified as being the length of time it takes poll workers to begin to report results once voting ends, the combination of the staggered hours and the ban in s. 329  results in a one to two and a half hour blackout of election results from the 32 ridings in the Atlantic region in Quebec to Alberta, and a two and a half to three hour blackout of Atlantic results in the Pacific time zone. There is no dispute that the only results that would be known to voters outside Atlantic Canada if there were no ban would be from these 32 ridings.  There are at present 308 federal ridings in Canada.

90                             On October 27, 2000, Paul Charles Bryan publicized his intention to post the federal election results from electoral districts in Atlantic Canada on his website before all the polls closed in Western Canada.  In response, the Commissioner of Canada Elections sent him a letter, dated November 21, 2000, warning him that his intended conduct was contrary to s. 329  of the Canada Elections Act .

91                             Despite this warning, on November 27, 2000, the night of the federal election, Mr. Bryan posted the Atlantic election results on his website notwithstanding that the polls were still open in parts of Western Canada.  He was charged with  contravening s. 329  of the Canada Elections Act  and argued, at trial, that the ban in s. 329  infringed s. 2 (b) of the Charter .

92                             Smith Prov. Ct. J. upheld the ban ((2003), 104 C.R.R. (2d) 364, 2003 BCPC 39).  He concluded that it was justified under s.  1  of the Charter , based particularly on the evidence of Professor Robert Hugh MacDermid, a York University political science professor.  Professor MacDermid explained that according to American studies, knowing the election results from the rest of the country, especially when this knowledge is combined with a media prediction of the election’s outcome, can have an impact on voter behaviour, including lower voter turnout and strategic voting.  Smith Prov. Ct. J. concluded that early disclosure of Eastern election results could have an adverse impact on voter behaviour in the rest of Canada, stating:

Common sense tells me there could be undesirable consequences from the early disclosure of results, particularly if they show an unexpected sweep for a particular party.  Individuals could decide not to vote because their favoured party is either secure or in a hopeless position.  Voters who had not intended to vote might change their minds and cast a ballot.  People are entitled to vote strategically.  Strategic voting means a supporter of Party A actually votes for Party B in the hope of preventing the election of the candidate for Party C.  Knowing eastern results prior to the closure of local polls would give the voter information . . . which was not available to other citizens.  The Crown uses the term information imbalance to describe the whole situation of some voters having more information than others.  That description seems reasonably accurate.  [para. 9]

Accordingly, he dismissed Mr. Bryan’s Charter  motion, convicted him, and fined him $1,000.

93                             On appeal to the Supreme Court of British Columbia, Kelleher J. allowed the appeal from the conviction ((2003), 233 D.L.R. (4th) 745, 2003 BCSC 1499).  He concluded that s. 329  of the Canada Elections Act  infringed s. 2 (b) of the Charter  and was not saved by s. 1  primarily because there was no evidence to support a finding that the government’s objective — “preventing an informational imbalance that could result from early disclosure of the election results from Atlantic Canada” — was a pressing and substantial objective (para. 41).  Even if the objective was sufficiently significant, in his view, the ban was not a proportional response.

94                             At the Court of Appeal for British Columbia, Rowles J.A., writing for the majority, concluded that Kelleher J. had erred by requiring direct evidence of an adverse impact on voting behaviour ((2005), 253 D.L.R. (4th) 137, 2005 BCCA 285).  In her opinion, “when the s. 329  publication ban is seen as having the same purpose or objective as the staggered voting hours, that is, to eliminate the information imbalance that can result from disclosure of results before all of the polls have closed, the respondent’s argument concerning the lack of evidence to support the ban falls away” (para. 59).  As a result, she found the objective of s. 329  to be pressing and substantial, and concluded that the salutary effects of the ban outweighed its harmful ones.  The limitation on freedom of expression created by s. 329  of the Canada Elections Act  was therefore justified under s. 1 .  Saunders J.A., in dissent, concluded that the ban in s. 329  was not justified because its harmful effects — the denial of access to election information — outweighed any beneficial consequences.

Analysis

95                             The impact of staggered hours as currently set out in s. 128  of the Canada Elections Act  is to ensure that the only results known to voters outside Atlantic Canada are the results of that region’s 32 ridings. The ban in s. 329  prevents these voters from knowing even these results when they vote.

96                             In two recent decisions, Thomson Newspapers Co. v. Canada (Attorney General), [1998] 1 S.C.R. 877, and Harper v. Canada (Attorney General), [2004] 1 S.C.R. 827, 2004 SCC 33, this Court dealt with constitutional challenges to other provisions of the Canada Elections Act  on the grounds that they violated s. 2 (b) of the Charter .  In Thomson Newspapers, this Court found that s. 322.1  of the Canada Elections Act , which prohibited the broadcasting, publication or dissemination of opinion poll results during the three final days of a federal election campaign, violated s. 2 (b) of the Charter .  Writing for the majority, Bastarache J. found that because the limitation was not minimally impairing, it was not justified under s. 1 .  In Harper, the issue was whether the third party spending limits in ss. 323(1)  and (3) , 350  to 357 , 359 , 360  and 362  of the Canada Elections Act  violated s. 2 (b) of the Charter .  Bastarache J., again writing for the majority, concluded that while third party election advertising is political expression which is infringed by spending limits, the limits were justified under s. 1 .

97                             Based on these decisions, the parties in this case do not dispute that publishing electoral results is a form of political expression.  Nor do they contest the proposition that the ban on the transmission of electoral results in s. 329  of the Canada Elections Act  infringes s. 2 (b).  The central disputed issue, therefore, is whether the infringement is justified under s. 1  of the Charter , which reads:

1.  The Canadian Charter of Rights and Freedoms  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.

In particular, the issue in this case is the sufficiency of the evidence adduced by the government to justify the infringement.

98                             The first analytical inquiry under s. 1  is into whether the objective of the limitation on the Charter  right is of sufficient importance to warrant overriding a constitutionally protected right.  Only if the government can demonstrate that the objective of the limitation is pressing and substantial is the second inquiry triggered, namely, the proportionality test. This second stage of the s. 1  analysis has three components:  is the limiting measure rationally connected to the objective; does it impair the right as little as possible; and is there proportionality between the harmful and beneficial effects of the limitation?

99                             As this Court has consistently noted, political expression is at the conceptual core of the values sought to be protected by s. 2 (b), which reads:

2.  Everyone has the following fundamental freedoms:

                                                                   . . .

(b)  freedom of thought, belief, opinion and expression, including freedom of the press and other media of communication;

The onus on the government under s. 1 , therefore, is to demonstrate that it is justified in infringing a form of expression that is at the heart of the constitutional right.  That means, as Dickson C.J. definingly stressed in R. v. Oakes, [1986] 1 S.C.R. 103, that the evidence necessary to prove the constituent elements of the s. 1  inquiry “should be cogent and persuasive and make clear to the Court the consequences of imposing or not imposing the limit” (p. 138).

100                         Significantly, however, he acknowledged “that there may be cases where certain elements of the s. 1  analysis are obvious or self-evident” (p. 138).  This acknowledgement that aspects of the government’s justification engage the obvious or self-evident, has attracted both judicial clarification and amplification, particularly in RJR-MacDonald Inc. v. Canada (Attorney General), [1995] 3 S.C.R. 199, where McLachlin J. explained the s. 1  burden on the government as follows (at paras. 127-29, 154 and 137):

First, to be saved under s. 1  the party defending the law . . . must show that the law which violates the right or freedom guaranteed by the Charter  is “reasonable”.  In other words, the infringing measure must be justifiable by the processes of reason and rationality.  The question is not whether the measure is popular or accords with the current public opinion polls.  The question is rather whether it can be justified by application of the processes of reason.  In the legal context, reason imports the notion of inference from evidence or established truths.  This is not to deny intuition its role, or to require proof to the standards required by science in every case, but it is to insist on a rational, reasoned defensibility.

Second, to meet its burden under s. 1  of the Charter , the state must show that the violative law is “demonstrably justified”.  The choice of the word “demonstrably” is critical.  The process is not one of mere intuition, nor is it one of deference to Parliament’s choice.  It is a process of demonstration.  This reinforces the notion inherent in the word “reasonable” of rational inference from evidence or established truths.

The bottom line is this.  While remaining sensitive to the social and political context of the impugned law and allowing for difficulties of proof inherent in that context, the courts must nevertheless insist that before the state can override constitutional rights, there be a reasoned demonstration of the good which the law may achieve in relation to the seriousness of the infringement.  It is the task of the courts to maintain this bottom line if the rights conferred by our constitution are to have force and meaning.  The task is not easily discharged, and may require the courts to confront the tide of popular public opinion.  But that has always been the price of maintaining constitutional rights.  No matter how important Parliament’s goal may seem, if the state has not demonstrated that the means by which it seeks to achieve its goal are reasonable and proportionate to the infringement of rights, then the law must perforce fail.

. . .

Where, however, legislation is directed at changing human behaviour, . . . the causal relationship may not be scientifically measurable.  In such cases, this Court has been prepared to find a causal connection between the infringement and benefit sought on the basis of reason or logic, without insisting on direct proof of a relationship between the infringing measure and the legislative objective: R. v. Keegstra, [1990] 3 S.C.R. 697, at pp. 768 and 777; R. v. Butler, [1992] 1 S.C.R. 452, at p. 503.

. . .

. . . the balance of probabilities may be established by the application of common sense to what is known, even though what is known may be deficient from a scientific point of view . . . . [Emphasis added; emphasis in original deleted.]

101                         In Sauvé v. Canada (Chief Electoral Officer), [2002] 3 S.C.R. 519, 2002 SCC 68, McLachlin C.J. further explained the interplay in a s. 1  analysis between evidence, inference and common sense as follows:

While some matters can be proved with empirical or mathematical precision, others, involving philosophical, political and social considerations, cannot. . . . [I]t is enough that the justification be convincing, in the sense that it is sufficient to satisfy the reasonable person looking at all the evidence and relevant considerations, that the state is justified in infringing the right at stake to the degree it has: see RJR‑MacDonald, supra, at para. 154, per McLachlin J.; R. v. Butler, [1992] 1 S.C.R. 452, at pp. 502‑3, per Sopinka J. . . . Common sense and inferential reasoning may supplement the evidence: R. v. Sharpe, [2001] 1 S.C.R. 45, 2001 SCC 2,  at para. 78, per McLachlin C.J. [Emphasis added; para. 18.]

102                         Similarly, Bastarache J. confirmed in Harper that the “court may rely on a reasoned apprehension of . . . harm” and that “logic and reason assisted by some social science evidence is sufficient proof of the harm that Parliament seeks to remedy” (paras. 77 and 79).

103                         These cases represent a judicial acknowledgment that policy judgments made by government often represent complex polycentric conclusions not easily amenable to “precise measurement” (McKinney v. University of Guelph, [1990] 3 S.C.R. 229, at p. 304).  But while scientific proof may not always be necessary or available, and social science evidence supported by reason and logic can be relied upon, the evidence must nonetheless  establish the consequences of imposing or failing to impose the limit.  As McLachlin C.J. warned in Sauvé, “one must be wary of stereotypes cloaked as common sense, and of substituting deference for the reasoned demonstration required by s. 1 ” (para. 18).

104                         The first determination is whether the government’s objective is “pressing and substantial”.  The government’s stated objective in enacting the publication ban now found in s. 329  was to prevent an informational imbalance which could have a negative impact on voter behaviour or perceptions of electoral fairness.  This Court in Harper accepted that “ensur[ing] that electors in different parts of the country have access to the same information before they go to the polls” is a pressing and substantial objective (para. 132).  I see no reason to reject it as a pressing and substantial objective in this case and acknowledge at the outset the significance of maintaining public confidence in electoral fairness and the integrity of the electoral system, as did Bastarache J. in Harper when he said:

Perception is of utmost importance in preserving and promoting the electoral regime in Canada. . . . Electoral fairness is key. [para. 82]

105                         The “rational connection” analysis is the initial step in the proportionality determination.  I think there can be little dispute that s. 329  is rationally connected to the objective of preventing an information imbalance.  As Kelleher J. observed:  “If the objective is to prevent the informational imbalance created by transmitting election results from Atlantic Canada to the west, banning these transmissions is an entirely rational way to fulfill that objective” (para. 48).  This finding is also consistent with Harper, where Bastarache J. held that a “blackout period [for election advertising] is . . . rationally connected to ensuring that all voters receive the same information where possible” (para. 133).

106                         It is difficult, however, to conclude that the limitation, a complete ban, minimally impairs the right to know election results in a timely way, that is, impairs the right no more than is reasonably necessary, or, in the language of McLachlin J. in RJR-MacDonald, is “carefully tailored so that rights are impaired no more than necessary” (para. 160).  But in my view it is unnecessary to further consider this issue because the government’s justification falters fatally in the balance of the proportionality analysis, namely, its submission that the benefits of the limitation are proportional to its harmful effects.

107                         For the reasons that follow, I find that the evidence submitted by the government in this case does not provide the requisite “reasoned demonstration” to justify infringing the right at stake to the extent that it has. Any evidence of harm to the public’s perception or conduct in knowing the election results from Atlantic Canada before they vote is speculative, inconclusive and largely unsubstantiated. The harm of suppressing core political speech, on the other hand, is profound. The benefits of the ban are, accordingly, far outweighed by its deleterious effects.

108                         In Thomson Newspapers, this Court accepted that the type of proof required under s. 1  will vary according to the “seriousness and likelihood of the harm” the limitation seeks to remedy (para. 111).  The harm at which the blackout period in s. 329  is aimed is the prevention of information imbalance to protect the perception or reality of electoral fairness.  However, this information imbalance must be considered not in the abstract, but in the context of staggered hours.  It is not sufficient to suggest that information imbalances in general are harmful to the public perception or conduct.  The proportionality exercise under s. 1  requires that the government demonstrate, by “cogent and persuasive” evidence, that this particular information imbalance is so harmful that it justifies the infringement represented by the ban.

109                         What the particular information imbalance now amounts to, whatever its aspirational provenance, is that some voters — those who vote during the final hours of election day — may know the results of 32 Atlantic region ridings before they vote.  The onus is on the government to demonstrate, therefore, that this specific information imbalance causes harm to voter perception or conduct to such a degree that the ban can be justified as a proportional response, considering the protected right at issue.

110                         The right at issue is the right of the media and others to publish election results in a timely fashion and the right of all Canadians to receive it (Thomson Newspapers, at para. 127).  Communicating and receiving election results is a core democratic right. Because the right to know election results is an essential part of the democratic process, clear and convincing evidence is required to justify limiting the availability of the information.  In my view, based on the record, there is only speculative and unpersuasive evidence to support the government’s claim that this particular information imbalance is of sufficient harm to voter behaviour or perceptions of electoral unfairness — the objects sought to be addressed by the ban — that it outweighs any damage done to a fundamental and constitutionally protected right.

111                         The government stressed the importance of maintaining public confidence in the electoral process and submitted that the s. 329  publication ban has the salutary effects of protecting the public perception that the electoral process is fair, and preventing the likelihood that premature disclosure of Atlantic election results could negatively affect voter behaviour.  In so doing, it relied primarily on three sources:  the evidence of Professor MacDermid, the findings of the Lortie Commission, and the Decima poll introduced as fresh evidence (Decima Research/Carleton University Poll, conducted during the period November 25 to December 5, 2005).

112                         Professor MacDermid’s conclusions were that information imbalance can affect voter perception and behaviour.  His evidence was based on American studies. The evidentiary utility of his conclusions are, consequently, limited by a fundamental difference between American and Canadian elections: the absence in the United States of staggered voting hours.  The consequences he anticipates for Canada are therefore hypothetical.  They derive from an election context where, unlike Canada, Western American voters could in fact know, based on results from the rest of the country and the predictions of the media, what the likely outcome of an election will be when they go to the polls.

113                         But it is instructive that all of the studies Professor MacDermid cited concluded that before there was any impact, let alone a harmful one, on voter turnout or behaviour from information imbalance, the imbalance must be of such a nature that voters know or can predict the outcome of the election. Nowhere in his report does Professor MacDermid cite a study that finds an impact on voter behaviour in the absence of such knowledge.

114                         As the following excerpts from his cross-examination demonstrate, Professor MacDermid agreed that knowing which party is going to form the government is a necessary precondition to voters deciding not to vote, a conclusion he acknowledged could not be reached based on staggered hours.  He also stressed the importance of being able to predict a winner, without which an information imbalance by itself has no effect on voter behaviour:

Q       And you tell me whether you -- this is still your evidence, or not?  “You need a prediction of a winner, must be able to vote -- and the prediction of the winner must be both made and different from that which was expected for there to be an effect on turnout.”

A       Absolutely, I stand by that.

. . .

Q       Are there any one of the studies that you have cited and relied upon . . . that say knowing some results, but not an outcome . . . with no projections . . . will result in people staying home . . . and not voting?

. . .

A       None of them.

. . .

Q       . . . All of the U.S. studies which found an effect on voter behaviour, whether that’s turnouts, strategic voting, whatever kind of voter behaviour we’re talking about, the condition precedent to finding that change in behaviour -- or that behaviour, I’m sorry, we don’t know if it was a change, if it’s aggregate data -- is the requirement that there be a media call?

A       Well --

Q       In those instances where there was no media call because the race was too close, I understood that the studies that you cited to say no effect?

A       I think I conceded when I talked about how to view these studies. I said that one couldn’t imagine there being effect if there is no media call, because that’s the -- that’s one of the variables that one is arguing is important, the presence of a media call, that has an effect on voters in certain circumstances and certain times of races, when they hear that information it causes them to change their mind. So when there is no call, ther[e] can’t really be an effect.  You’re right.

Q       [I]n order for there to be an effect on voter behaviour . . . there has to be a media call? No media call, no effect?

A       I think to the extent that that’s what you’re looking at, yes. . . . I think that’s right.

115                         In Professor MacDermid’s view, however, “[t]he important point” is not so much “whether voters engage in [strategic] voting”, since, in his words, “[m]ost probably decide to do this even before election day and how they choose to explain or arrive at reasons for voting is really their own business.”  Notwithstanding his observation that inconclusive results do not affect voter turnout or behaviour, he was nonetheless of the view that the real harm is in the fact that “some voters have information about the race that others cannot have”.  The harm is the mere fact of an information imbalance.

116                         In his report, he further explains the harm from information imbalance as being a feeling that “western voters don’t count since eastern voters have already decided the election”.  Yet it is difficult to see how this perception of unfair treatment can be even theoretically attributed to Western voters when, in the context of staggered hours, they will not know the final decision when they vote.  Whatever other conclusions voters may be able to draw for themselves about the Atlantic election results, they will not realistically be able to predict the outcome of the election and therefore, logically, the perception that they “don’t count” is unlikely, given that they will only know, at most, 11 percent of the election outcome when they do vote. 

117                         Professor MacDermid was unable to offer any support for his theory that information imbalances have an inherently harmful impact.  The possibility that some Western voters might be influenced by results from Atlantic Canada cannot be completely discounted, but the question is not, it seems to me, whether there will be any impact on Western voters of knowing the results from these 32 ridings.  It is impossible to immunize voters from all conceivable influences, as this Court concluded in Thomson Newspapers when it struck down the ban on the availability of opinion poll information immediately before an election.  The question is whether the impact will be a harmful one, and while deciding not to vote may be in this category, Professor MacDermid’s evidence was that there is no impact on voter turnout unless the outcome of the election is known or knowable.

118                         Nor is there support in the Lortie Commission for the contention that this information imbalance is so significant a public concern as to justify a complete ban no matter how inconclusive the election information is.  The government observed that the Lortie Commission found that Canadians “feel very strongly about premature release of election results and favour changes in voting hours to eliminate the problem” (vol. 2, at p. 84).  But this quotation must be read with the perspective of its timing.  It refers to the attitudes of Canadians before staggered voting hours were introduced, and explains why the Commission recommended them as an alternative to the ban on disclosing election results.  It also explains, as does the Decima poll discussed below, why I accept, unequivocally, that the government’s objective of protecting the perception of electoral fairness is a pressing and substantial one.  

119                         It does not, however, justify the impairment of the right.  Had the Lortie Commission been of the view that all information imbalances were problematic or harmful, either perceptually or behaviourally, it would have suggested that a solution like staggered hours be accompanied by the retention of the ban to prevent an information imbalance from knowing 32 riding results.

120                         The specific harm the Lortie Commission identified was in voters knowing the outcome of the election when they voted.  It never suggested that the mere fact of an information imbalance created a perceptual harm to public confidence in electoral fairness.  In fact, it observed that “the release of some election results before polls close in the West — specifically, results from the 32 seats in Atlantic Canada — would not constitute a major problem so long as other results from eastern Canada were not available until after the polls closed in the West” (vol. 2, at p. 85 (emphasis added)).  As Kelleher J. noted, “the most pressing concern [of the Lortie Commission] was ensuring that voters in the west did not hear results from Quebec and Ontario before the polls closed, since these two provinces have the potential to determine who will form the government” (para. 36). 

121                         The harm perceived by the Lortie Commission — the influence on voter perception generated by knowing the Ontario and Quebec results — is alleviated by staggered hours.  This leaves only the questionable harm of knowing the Atlantic results in the absence of a ban. With respect, therefore, I disagree with the assertion of my colleague Justice Fish that “the findings of the Lortie Commission provide a sound basis for concluding that this information imbalance is a real and significant harm” (para. 71). 

122                         The government also introduced the  Decima Research/Carleton University Poll as persuasive of the fact that, without the ban, Canadians would perceive elections to be unfair. According to this poll, 70 percent of Canadians surveyed “thought people should not be able to know election results from other provinces before their polls close”.  We do not know how these same Canadians would have responded if the purpose and effects of staggered hours had been explained.  Absent the relevant context, the answer is a very general response to a very general question and is indicative of an unsurprising public preference for the goal of electoral fairness, not persuasive evidence that elections would in fact be seen by reasonably informed Canadians to be unfair unless there was a ban prohibiting voters from learning the results of Atlantic Canada’s 32 ridings before they voted.

123                         Any potential benefits of the publication ban are also diminished by the reality that, as the Lortie Commission pointed out, the s. 329  publication ban was “rendered obsolete” by broadcasting and telecommunications technology (vol. 2, at p. 84).  The ineffectiveness of the blackout was one of the reasons the Commission recommended that it be replaced by suggested staggered hours.  Even the Decima poll submitted by the government found that among those surveyed who were considered “[i]nvolved” Canadians, 35 percent indicated that they would make an effort to find out election results despite the ban.  As Professor Chris Waddell observed in commenting on the Decima poll:

[T]hanks to technology there are more and more ways for people to get around the blackout if they wish. Unlike phone calls to a single person or web sites where content can be controlled, e-mails, instant messages and cell phone text messages can all be distributed simultaneously to a broad cross-section of people. So while a majority support the blackout, the blackout itself is . . . more a fiction for those who are really interested in how the votes are being counted on election night.

124                         The impact of technology on the unenforceability of a publication ban was considered by this Court in the context of criminal proceedings in Dagenais v. Canadian Broadcasting Corp., [1994] 3 S.C.R. 835.  Lamer C.J.’s conclusions are, in my view, particularly apposite in the circumstances of this case, when he observed that:

It should also be noted that recent technological advances have brought with them considerable difficulties for those who seek to enforce bans.  The efficacy of bans has been reduced by the growth of interprovincial and international television and radio broadcasts available through cable television, satellite dishes, and shortwave radios. It has also been reduced by the advent of information exchanges available through computer networks.  In this global electronic age, meaningfully restricting the flow of information is becoming increasingly difficult.

. . .

If the actual beneficial effects of publication bans are limited, then it might well be argued . . . that the negative impact the ban has on freedom of expression outweighs its useful effects. [pp. 886-87]

125                         All of this leads me to conclude that any inference that the information imbalance created by lifting the ban in s. 329  would have a harmful impact of any kind on voter turnout, choice or perception, is highly theoretical and far from sufficiently persuasive to justify infringing the core right at issue in this case.  It is true that scientific proof of the harmful effects of the information imbalance created by disclosing Atlantic Canada’s election results may not be possible, but there must still be a reasoned or logical basis for inferring its validity. 

126                         The evidence of Professor MacDermid is that there is no effect on voter turnout unless the voter is able to predict the election’s outcome.  The evidence from the Lortie Commission suggests that there is no harm to public perception from knowing the results of 32 ridings.  And the evidence of the Decima poll does not address, one way or another, whether this particular information imbalance would be harmful to public perceptions of electoral fairness.  This makes the government’s evidence essentially a combination of speculation and theory, unsubstantiated by cogent evidence. 

127                         It is far from clear to me that there is any evidence at all to demonstrate that the ban in s. 329 , in the context of staggered hours, is directed at a demonstrated harm and sufficiently promotes public confidence in the fairness of elections to justify infringing the right to disseminate and receive election results. This view was shared by Stephen Harper, then a Member of Parliament for the Reform Party, at the third reading in 1996 of Bill C-63 (An Act to amend the Canada Elections Act, the Parliament of Canada Act  and the Referendum Act), when he observed:

It is important to realize, and anybody who looks at the transcripts of the committee will see that no convincing evidence was provided to the committee that this is creating a serious problem in the functioning of our democracy.

. . . I asked the various witnesses whether there was any documented evidence or any serious academic study on whether knowing the results in other parts of the country had either of two effects: caused people not to vote or it caused them to vote differently than they would vote otherwise.  There is precious little evidence that either of these things are true.

(House of Commons Debates, vol. 134, 2nd Sess., 35th Parl., November 26, 1996, at p. 6723)

128                         On the other hand, the harm caused by the ban to the expressive rights in s. 2 (b) is considerable.  For the duration of the ban, the Atlantic election results are denied to all Canadians west of the Atlantic provinces, many of whom have already voted.  It is difficult to imagine a more important aspect of democratic expression than voting and learning the results of their vote.  The s. 329  ban impairs the right both to disseminate and receive election results at a crucial time in the electoral process.  To suggest that this is only a delay, not the suppression of information, unduly minimizes the significance both of the information and of the delay.

129                         The fact that the suppression of information is of limited duration is not determinative.  News is news precisely because of its immediacy, especially during an election, where a great deal of information and analysis must be passed on to the public in a short period of time.  Canadians are entitled to know, as soon as possible, who their elected representatives are.

130                         There is, in addition, harm of a more practical nature.  The technical challenges faced by the media when implementing the ban are considerable, and mean that individuals in certain areas may not receive radio or television election coverage even though the polls have closed in their area of the country.  The media’s role in disseminating election information to the public is crucial.  Election coverage is how the public learns the outcome of its defining democratic right — voting.  It is in this country’s interest to enhance, not impede, the media’s ability to transmit this evolving information as it becomes available.  Similarly, Atlantic Canadians who rely on the Internet as their primary source of news are denied any election news from their medium of choice since the transmission of election results over the Internet is prohibited until the end of the blackout.

131                         All of this harm to the Charter  right is demonstrable; the benefits of the ban are not.  To justify limiting a right, the evidence must establish the consequential harm that would occur in the absence of this limitation.  The government has offered only hypothetical and speculative evidence.  There is no reasonable basis for assuming that Canadian voters will perceive the electoral process to be unfair or will adjust their voting behaviour based on learning the results of 11 percent of the election results.  I agree with Kelleher J. who concluded:

There was no evidence before the trial judge of any effect on voter behaviour.  The evidence of Dr. MacDermid was that an information imbalance could have harmful effects:  it could reduce voter turnout, it could influence how citizens vote by causing strategic voting, it could cause “bandwagon” voting, and it could cause “underdog” voting.  But the evidence does not show that any of these harmful effects occurs in the context of staggered voting hours.  There is no evidence of reduced turnout where any early results information relates to no more than 11 per cent of the ridings.  Similarly, strategic voting, “underdog” voting or “bandwagon” voting occurs where the likely outcome of the election is known from the results published from another part of the country.  As noted above, the 32 ridings in Atlantic Canada cannot determine the likely outcome of the election . . . .  [para. 44]

132                         There is, therefore, no persuasive evidence of harm requiring the remedial attention of a publication ban.  This means that the government’s justification fails the proportionality branch since there can be nothing proportional between the benefits of a limitation and its harmful effects if there is no demonstrated benefit to the limitation.

133                         This Court concluded in Thomson Newspapers that it was constitutionally impermissible to prevent voters from learning about public opinion polls just before an election because no harm could be imputed from this knowledge.  Similarly, in this case the imputation of harm is weak, speculative and unpersuasive.  Potential harm caused by electoral informational imbalance is eliminated by staggered hours, which guarantee that voters cannot know what the overall outcome of the election is likely to be before they go to the polls, the only harmful information imbalance identified by the Lortie Commission as worthy of remedial attention.  Section 329  of the Canada Elections Act  is, therefore, a remedy for an undemonstrated problem and, as a result, an overbroad intrusion on a Charter  right. As an excessive response to an insufficiently proven harm in light of the implementation of staggered voting hours, the publication ban in s. 329  of the Canada Elections Act  is therefore a violation of s. 2 (b) of the Charter  that is not justified under s. 1 .

134                         For these reasons, I would allow the appeal, set aside Mr. Bryan’s conviction, and declare s. 329  of the Canada Elections Act  to be of no force or effect by virtue of s. 52  of the Constitution Act, 1982 .

Appeal dismissed, McLachlin C.J. and Binnie, LeBel and Abella JJ. dissenting.

Solicitors for the appellant:  Taylor Jordan Chafetz, Vancouver.

Solicitors for the respondent Her Majesty the Queen:  Fasken Martineau DuMoulin, Vancouver.

Solicitor for the respondent the Attorney General of Canada:  Attorney General of Canada, Toronto.

Solicitors for the interveners Canadian Broadcasting Corporation, CTV Inc., TVA Group Inc., Rogers Broadcasting Limited, CHUM Limited, Sun Media Corporation, Sun Media (Toronto) Corporation, Canadian Press, Globe and Mail, CanWest MediaWorks Inc., CanWest MediaWorks Publications Inc. and Canoe Inc.:  Arvay Finlay, Vancouver.

Solicitors for the intervener the Canadian Civil Liberties Association:  Osler, Hoskin & Harcourt, Toronto.