A central tenet of neo-institutionalist approaches to the study of politics is that institutional reforms create opportunity structures that help some actors—in both civil society and the state—pursue their policy goals, but that simultaneously frustrate other actors and goals (Hall and Taylor, Reference Hall and Taylor1996; Immergut, Reference Immergut1998). As a fundamental reform to Canada's constitutional structure, the adoption of the Charter of Rights and Freedoms, and the related emergence of substantive judicial review of legislation and government actions, created such opportunities for interest groups and judges (Brodie, Reference Brodie2002; Epp, Reference Epp1998; Hein, Reference Hein, Howe and Russell2000; Morton and Knopff, Reference Morton and Knopff2000) but also governments. As others have noted (for example, Ginsburg, Reference Ginsburg2003; Graber, Reference Graber1993, Reference Graber2005; Hirschl, Reference Hirschl2004, Reference Hirschl2008; Whittington, Reference Whittington2005), there is a growing awareness within the comparative law and courts literature that governments often have incentives to empower courts and to use them to achieve policy goals, rather than by acting through the legislative branch—particularly when the latter is not possible. Hirschl's “hegemonic preservation thesis” (Reference Hirschl2004), for example, posits that political elites establish judicial review when they foresee that they will soon lose power in the electoral realm to actors who do not share the current elite's values and if the judiciary shares that elite's values, are appointed for fairly lengthy tenures, and are seen as more legitimate or trustworthy by the public than elected officials. Similarly, others see empowered, independent courts as a form of “political insurance,” which are created and sustained by political actors when regime power is diffused, or when elected authorities anticipate periodically losing power (Finkel, Reference Finkel2008; Ginsburg, Reference Ginsburg2003). Courts can also help the current government achieve its immediate policy goals, however, because, as Whittington observes, “judicial review can be used to void statutes passed by previous governing coalitions…when governing coalitions are unable or unwilling to displace the legislative baselines themselves” (Reference Whittington2005: 584).
The broader implication of such research is that the traditional view of judicial review, which sees judicial rulings that invalidate legislation as “activist” and a “counter-majoritarian dilemma” (Bickel, Reference Bickel1962), is overly simplistic. Elected governments and courts are, more accurately, engaged in a complex interaction that blurs (often intentionally, on the part of the government) responsibility for policy outcomes (Hennigar, Reference Hennigar, Kelly and Manfredi2009; Hirschl, Reference Hirschl2008). Although Hirschl observes that “it is hard to positively ascertain” (2009: 108) evidence of governments' intentionally delegating authority to courts, an example of this more complex relationship is when government lawyers concede in court that laws passed by their respective legislatures are unconstitutional under the Charter, thus facilitating judicial invalidation of legislation. Existing work (Hennigar, Reference Hennigar2007) has identified and analyzed one such type of concession, when governments decline to appeal losses in the lower courts to the Supreme Court of Canada (SCC), thus quietly allowing laws to be invalidated or reinterpreted by courts that operate in relative obscurity. This research found that the practice is not uncommon with respect to the Canadian federal government, which appealed only half of the lower appellate court rulings that found its laws violated the Charter (Hennigar, Reference Hennigar2007: 241).Footnote 1
This article focuses on a second form of concession, when governments concede in their substantive arguments before the court that legislation is unconstitutional. Several law and politics commentators and some justices of the SCC have discussed the normative implications of such concessions. Huscroft (Reference Huscroft1995, Reference Huscroft2009) and Morton and Knopff (Reference Morton and Knopff2000) criticize concessions as undemocratic and disrespectful of Parliament because, through concessions, to use Whittington's phrase, “the Court assists powerful officials within the current government in overcoming various structural barriers”—namely, opposition within the legislative branch—“to realizing their ideological objectives through direct political action” (Reference Whittington2005: 584). The example most frequently cited by Canadian critics is that of Ontario Attorney General Marion Boyd in M. v. H. (1996). Boyd conceded before the lower court that some provisions of Ontario's family law regime violated the Charter's section 15 equality rights because they failed to include same-sex spouses. This came after Boyd's NDP government had attempted to reform these provisions through the Ontario legislature but had been blocked in a free vote.Footnote 2 In contrast, several authors (Edwards, Reference Edwards and Sharpe1987; Freiman, Reference Freiman2002; McAllister, Reference McAllister2002; Roach, Reference Roach2000, Reference Roach2006; Scott, Reference Scott1989) argue that attorneys general should concede a law's unconstitutionality—even over the objections of their political superiors—when they are convinced that it is necessary to defend the public interest and the rule of law.
Supreme Court justices who have publicly addressed concessions have been critical of the practice. In Schachter v. Canada (1992), for example, Chief Justice Lamer and Justice La Forest registered their strong disapproval of the Attorney General of Canada (AG Canada)'s concession that the Unemployment Insurance Act was discriminatory because it denied biological fathers the same parental leave benefits as mothers and adoptive parents, on the grounds that the concession pre-empted and undermined the role of the court. As Chief Justice Lamer wrote:
I find it appropriate at the outset to register the court's dissatisfaction with the state in which this case came to us…. The appellants chose to concede a s.15 violation and to appeal only on the issue of remedy. This precludes this court from examining the s.15 issue on its merits…. Further, the appellants' choice not to attempt a justification under s.1 at trial deprives the court of access to the kind of evidence that an s.1 analysis would have brought to light.
All of the above essentially leaves the court in a factual vacuum with respect to the nature and extent of the violation, and certainly with respect to the legislative objective embodied in the impugned provision. This puts the court in a difficult position in attempting to determine what remedy is appropriate in the present context. (10–11)
Similarly, when the AG Canada conceded in R. v. Sharpe (2001) that Criminal Code prohibitions on child pornography violated the Charter's freedom of expression, Justices L'Heureux-Dubé, Gonthier and Bastarache complained that “it is unfortunate that the Crown conceded that the right to free expression was violated in this appeal in all respects, thereby depriving the Court of the opportunity to fully explore the content and scope of s. 2(b) as it applies in this case” (para. 151).
Notwithstanding this attention to concessions, to date there has been no attempt to assess empirically the frequency with, and conditions under, which they occur in Canada. This article—a preliminary analysis from an ongoing project on the federal government's Charter litigation—begins to fill this gap in the law and politics literature by examining concessions by the federal government of Canada in all Supreme Court Charter cases from 1982 to 2004 and considering the implications of the findings for the larger relationship between the executive and judicial branches. Two important dimensions of this relationship, both of central importance in the Canadian law and political literature, are of particular note. First, and most basically, the rate of concessions tells us the degree to which the Court's level of apparent judicial activism is, in fact, not activist. For a judicial ruling to be activist it must second-guess the actions or policy choices of the legislative or executive branches of government (Russell et al., Reference Russell, Knopff and Morton1990). A ruling which simply complies with a government's concession that its laws or actions are unconstitutional under the Charter is not, therefore, activist in the proper sense.
Second, some scholars, rejecting the simple adversarial relationship suggested by leading conceptions of “judicial activism,” contend that courts and the elected branches of government are engaged in a long-term collaborative process, to which both are genuinely committed, to unfold the meaning of the Charter. Examples include advocates of “co-ordinate construction” (Huscroft, Reference Huscroft2007; Manfredi and Kelly, Reference Manfredi and Kelly1999; Tushnet, Reference Tushnet1999), as well as those whose work examines how governments and Parliament have internalized Charter-based review and discourse into their policy-making processes (Hiebert, Reference Hiebert2002; Kelly, Reference Kelly2005). These accounts also challenge the prevailing “dialogue” metaphor for the interaction between courts and the elected branches under the Charter, in particular the presumption among its proponents (for example, Hogg and Bushell, Reference Hogg and Bushell1997; Hogg et al., Reference Hogg, Thornton and Wright2007; Roach, Reference Roach2001) that courts have a monopoly over the “correct” interpretation of rights. Hennigar complains that the existing critics of the dialogue theory, however, “define away the potential for genuine agreement resulting from inter-institutional discourse” (Reference Hennigar2004: 8), and that both sides of the dialogue debate conceive of dialogue “in discrete inter-institutional terms, with the judiciary ‘speaking’ through rulings, and government through legislation,” thus overlooking “the important, ongoing dialogue that occurs through government litigation, which brings the two institutional actors together within the court setting, with government lawyers arguing legal interpretation before judges” (3). Concessions (and non-appeals) are arguably evidence of such executive–judicial agreement, and as such provide valuable information about the prevalence of dialogue through litigation.
As we will see in the next section, concessions through legal argumentation can take a number of forms, including conceding right violations, conceding that a violation is not a reasonable limit, or both. The next section also elaborates on why governments might concede, and the means through which the federal government can do so, which are complicated by the particular way many federal laws—and in particular the Criminal Code—are enforced in Canada. The paper's methodology will then be outlined, followed by a discussion of the findings, which are principally that while concessions of rights violations are not common, neither are they exceptionally rare; in contrast, the AG Canada almost never concedes that federal legislation is an unreasonable limit on rights under section 1 analysis. Moreover, concessions are typically driven by the Court's jurisprudence or legislative action, making the concession in court moot. This said, there is some evidence that the government's partisan affiliation influences “partial” concessions.
Theoretical Background: Why Concede, and How?
Marion Boyd's actions highlight one of the key reasons for conceding rights violations, namely, to realize legislative reforms that cannot be achieved through regular parliamentary channels due to opposition within one's party or from other political parties. Alternatively, the government might opt to concede in court when it could achieve its policy goals through traditional parliamentary channels but wishes to avoid political responsibility for the policy change. While legislation on controversial issues usually attracts considerable public attention, concessions in court are typically “below the radar” of media and the public, making it an attractive strategy for governments trying to accomplish their policy goals more discreetly. As Hirschl observes, “from the politicians' point of view, delegating policy-making authority to the courts may be an effective means of shifting responsibility and thereby reducing the risks to themselves and to the institutional apparatus within which they operate. At the very least, the transfer to the courts of contested political ‘hot potatoes’ offers a convenient retreat for politicians who are unwilling or unable to settle public disputes in the political sphere” (Reference Hirschl2008: 106–07). The federal government's decision to refer the constitutionality of same-sex marriage to the SCC in 2004 falls into this category; when the Court refused to take the bait by refusing to answer the key question of whether the Charter required recognizing same-sex marriage, the Liberal government demonstrated that the change was possible through legislative means when it passed the draft bill it had referred. The higher levels of public trust in the judiciary than in the elected branches—a common, if ironic, phenomenon in democracies (Hirschl, Reference Hirschl2008: 107)—add an extra incentive for governments to seek judicial legitimization of their policy preferences. Delegation is also particularly attractive to politicians who face “no win” decisions, where the issue in question is so polarizing for the public that any legislative action the government takes will provoke considerable opposition, as with abortion (Graber, Reference Graber1993).
Concessions may also occur in less politically charged cases, however, when the government simply wants to save time and political capital by engineering what amounts to the amendment or repeal of legislation through the courts rather than through the regular legislative process. This may be especially appealing if the law in question was adopted by a previous government of a different party or involves a largely technical matter.
All of these scenarios underline the complex political dynamic between governments and courts and that sometimes “losing is winning” from the government's perspective. Moreover, while there can be advantages to conceding before any court, doing so in the Supreme Court has the added benefit of a) securing a ruling from the highest court in the land, foreclosing calls for further appeals to get that Court's opinion, and b) establishing a stronger precedent for related issues.
Huscroft (Reference Huscroft1995) identifies three options government lawyers (and, potentially, their political superiors) face when presented with a Charter-based challenge to a statute or regulation. The first is a “full Charter defence,” which sees the government arguing that the law does not violate the claimant's rights, and, should the court disagree, that the law is a “reasonable limit” under section 1 of the Charter. Section 1 permits the government to limit any of the Charter's rights, so long as the limits are “reasonable,” explicit in the statute, and “demonstrably justified,”Footnote 3 which the Court interpreted in R. v. Oakes (1986) as requiring that the government demonstrate a “pressing and substantial objective” for the violation, and that this objective be “proportional” to the violation: that is, that they be “rationally connected,” that the means used “minimally impair” rights, and that the collective benefit of the violating law outweighs the harm caused to the rightsholder(s). A government could concede under s.1 by conceding any of these parts of the Oakes test.
A second option is to give a “limited Charter defence,” which could mean either conceding the rights violation but not s.1, or, less commonly, contesting the rights violation but offering no s.1 defence. In the Sauvé (2002) case, for instance, the federal government conceded that the law—which denied the vote to prisoners sentenced to at least two years—violated the Charter's s.3 right to vote, but offered a vigorous s.1 defence based on political philosophy (social contract theory), the importance of citizenship and voting, and the fact that the prohibition ended when the prisoner was released. An example of a case where no s.1 defence was offered is Chaoulli (2005), which concerned whether Quebec's ban on private health insurance violates the Charter's s.7 right to “life, liberty, and security of the person.” As co-defendant with Quebec, the federal government in Chaoulli denied that the ban violated s.7, but did not address s.1, possibly because the AG Quebec offered a full Charter defence of its law. While some violations, if found, would be difficult to justify as “reasonable,” failing to argue s.1 is usually a questionable strategy as it gives the court no choice but to find the law unconstitutional if it does find a Charter violation—which occurs much more often than a finding of unreasonableness (Hiebert, Reference Hiebert1996; Kelly, Reference Kelly2005).
The third option is “no Charter defence,” which entails conceding both that the law violates a Charter right and that it is unreasonable. A government might choose this strategy if the law was passed a long time ago (especially if passed before the adoption of the Charter), or if it was passed by a previous government of a different party. An example of full concession is the Schachter case mentioned earlier, where the federal government conceded that its discrimination against biological fathers was unreasonable on its way to challenging (successfully) the lower court's remedy of “reading in” natural fathers to the parental leave benefits program. Ottawa's concessions in Schachter are likely explained by the fact that the government had already extended such benefits to all parents, while cutting the amount significantly to maintain the program's overall cost.
Concessions can occur when a government is a party to the case, that is, the prosecution (in criminal cases), plaintiff or defendant (in civil cases) at the initial trial, or the appellant or respondent by the time it reaches the Supreme Court. With respect to the federal government, however, concessions regarding federal laws can also occur when the AG Canada is a third-party intervener. A unique feature of Canadian law is that Criminal Code offences—which are created by the federal Parliament—are usually enforced by provincial Crown prosecutors. This is because s.92(14) of the 1867 Constitution gives the provinces jurisdiction over the administration of justice, as does the Criminal Code.Footnote 4 This division of responsibility reflects Canada's federal nature and creates an interesting dynamic: criminal offences are created at the national level but enforced “locally,” thus allowing national standards to be influenced by the values of the smaller community in which the crime took place. A similar dynamic exists in Quebec with the federal Youth Criminal Justice Act (and its predecessor, the Young Offenders Act), the former Narcotics Control Act and the Controlled Drugs and Substances Act (an example of “asymmetrical federalism”), and with some provincial aboriginal constables enforcing the Indian Act on reservations. (Notably, however, there are several offences which are usually enforced by the federal government, most notably narcotics outside Quebec, income tax fraud, illegal fishing, and since 2001, terrorismFootnote 5; historically, federal Crowns also prosecuted crimes in the Territories.) Sometimes, wide differences in enforcement emerge at the provincial level. An example arose when the federal government created its gun registry, which was deeply unpopular in rural Canada where guns and hunting are common. The several provincial governments which opposed the registry, such as Alberta, Newfoundland, and Ontario, stated that they would not prosecute individuals who committed the offence of refusing to register their guns (Lindgren and Naumetz, Reference Lindgren and Naumetz2003: A1). Where there is provincial prosecution of federal laws, the AG Canada has a right under statute and regulatory law, but not the constitution, to intervene when the constitutionality of the law is challenged, as it can often illuminate the rationale for the law and its legislative history better than the prosecuting province (and, of course, may have a stronger incentive to do so).Footnote 6 Although the federal Department of Justice's guidelines state, “The Attorney General of Canada intervenes in criminal cases selectively, not as a matter of routine” (2005: chapter 47.1), they also note that “The Attorney General of Canada intervenes frequently in the Supreme Court of Canada, occasionally in the other appellate courts, and very infrequently at the trial level (except, perhaps, in language and aboriginal rights cases)” (2005: chapter 47.3b).
Before proceeding to the details of this study, it is appropriate to consider briefly the matter of who decides whether to concede a rights violation in Court. The content of federal government facta (written arguments submitted to the Court) receive extensive scrutiny by senior officials within the AG Canada's office. Interviews with senior Justice Department officialsFootnote 7 confirmed the process that is laid out in official Department of Justice guidelines (2005: chapter 23.6), that facta are approved by a series of top bureaucratic officials, including senior directors in Justice Department regional offices, the Assistant Deputy Attorney General, and especially the Litigation Committee—an Ottawa-based committee composed of several senior Justice Department lawyers and invited client department representatives (see Hennigar, Reference Hennigar2002, Reference Hennigar2007, for more detail). We can infer from this process that concessions are quite deliberately chosen by senior officials in the AG's office. In view of the fact that high-profile or politically sensitive litigation is sometimes directed by political and bureaucratic officials at the centre of government (Hennigar, Reference Hennigar2008), concessions may even have been requested by the “client” minister or ordered by the prime minister him- or herself.
Methodology and Data
This study examines the AG Canada's facta in every case decided by the Supreme Court in which a federal law was challenged under the Charter, from 1984 (the year of the first Charter case in that court, Law Society of Upper Canada v. Skapinker) up to and including 2004. The AG Canada (or his or her agents) is the government's official representative in litigation involving virtually all line departments and agencies and is therefore the logical focus of the study. It should be noted, however, that this focus excludes those cases where an institution or official of the federal state is not represented by the AG: these include those involving the Canadian International Trade Tribunal, the Judge Advocate General (National Defence), the Canadian Human Rights Commission, the Office of the Information and Privacy Commissioners, the Senate and the House of Commons (Brunet, Reference Brunet2000: 67; MacNair, Reference MacNair2001: fn 10).
I include appearances as both a direct party to the case (appellant, respondent, or both in cross-appeals) and as an intervener when federal legislation was being challenged. Notably, there were eight instances where the AG Canada filed a single factum for multiple cases (that is, those given distinct registration numbers by the Registrar of the Supreme Court), seven of them when intervening and once as a party.Footnote 8 For counting purposes, the study focuses on facta rather than Court decisions, as befits a project examining governmental arguments. By the same token, I count as distinct entries separate facta filed in cases the Court later consolidates into a single decision (for example, in White, Ottenheimer & Baker v. Canada (Attorney General) and Lavellee, Rackel & Heinz v. Canada (Attorney General), 2002), on the grounds that the AG Canada may have had a good reason for filing separate facta. The facta were purchased (with the much-appreciated assistance of a SSHRC Standard Research Grant) from the Court Records Office within the Registry Branch of the Supreme Court of Canada.Footnote 9
The search parameters yielded 139 facta, with a fairly even split between the number of intervener facta (66 or 47.5% of the total) and facta as party (73, or 52.5%). Within the latter group, 47 (64.4%) were filed when the AG as the respondent, 23 (31.5%) as appellant, and three (4.1%) where they were both appellant and respondent due to a cross-appeal. That the AG Canada appears as a respondent twice as often as it does as appellant is consistent with previous findings that Ottawa enjoys a spectacularly high success rate in the penultimate courts of appeal (that is, the provincial and federal Courts of Appeal), and so does not have the opportunity to appeal nearly as often as it is called to appear by another party (Hennigar, Reference Hennigar2002). A provision of the Criminal Code was challenged in 59 (83%) of the interventions, but in only 12 (17%) of the appearances as party. This pattern reflects the dynamic created by provincial prosecution of the code, discussed above, with Ottawa rarely prosecuting the code but having a right to intervene in such cases.
Table 1 Numerical Breakdown of AG Canada Facta in Charter Cases Involving Federal Laws before the Supreme Court of Canada, 1984–2004
As noted above, the government may concede that the provision violates (prima facie) a Charter right, and/or concede that a law that violates a right is “unreasonable” under s.1 analysis. Concessions were determined through qualitative analysis of facta arguments by me and two research assistants, which were then coded for quantitative analysis. Federal government concessions of a violation were coded as “1,” rights claims that were contested were coded “0,” and facta which did not address the rights claim were coded “9.” As discussed below, while it was useful to distinguish instances of AG silence on the rights claim from explicit concessions, the failure to address a rights claim which challenges a legal provision is an implicit concession. Concessions under s.1 were coded similarly, and again, the absence of a s.1 defence (“9”) is an implicit concession, since it effectively requires the Court to find the violation unreasonable (Huscroft, Reference Huscroft1995: 146). It should be noted that in some cases, the government simply stated that it did not concede s.1, but offered no argumentation or evidence to that effect; such cases were coded as “not addressed” (=9), since such an approach would clearly be insufficient to persuade the Court to uphold the law.
To determine the frequency of each of Huscroft's three litigation argumentation strategies—“full Charter defence,” “limited Charter defence” and “no Charter defence” —the variables for rights concessions and s.1 concessions were recoded as follows:
• A “full Charter defence” was offered if the AG Canada contested both the rights claim and the s.1 analysis.
• A “limited Charter defence” was offered if the AG Canada EITHER contested the rights claim (=0) BUT conceded or did not address s.1 (=1 or 9), OR conceded or did not address the rights claim (=1 or 9) BUT contested s.1 (=0).
• “No Charter defence” was offered if the rights claim was conceded or not addressed (=1 or 9) AND s.1 was conceded or not addressed (=1 or 9).
Findings and Discussion
Of the 139 facta addressing a challenge to a federal law or regulation, only 19 (13.7%) conceded a Charter violation, with another one (0.7%) not addressing it.Footnote 10Table 2 summarizes the AG Canada's strategies in these cases, organized by its role in the case. While admittedly there is no existing benchmark for what constitutes “rare” or “frequent” concessions, it is immediately apparent that there are few full concessions (“no Charter defence”): only 3.6 per cent of all AG Canada facta submissions on federal laws over the first 20 years of Charter litigation in the Supreme Court. At five total, in other words, a complete concession is only made on average about every four years, while an average of over seven federal laws are challenged under the Charter per year. More detailed analysis reveals an even more surprising conclusion: there has only been one case where the AG Canada explicitly conceded a violation and s.1 (as opposed to implicitly conceding by not addressing one or the other): Schachter! That this commonly cited example by critics of concession (for example, Huscroft, Reference Huscroft1995; Morton and Knopff, Reference Morton and Knopff2000) is an isolated event suggests that the concession issue may be something of a “tempest in a teapot,” at least as regards the federal government. At the other end of the spectrum, “full Charter defence” is by far the most common strategy, representing just over 70 per cent of the facta. Regarding the other four facta offering no Charter defence—in R. v. Hamill (1987), R. v. Yorke (1993), Canadian Egg Marketing Agency v. Richardson (1998), and R. v. Johnson (2003)—there appear to be good (or mitigating) reasons for the decision to concede the violation and to not address s.1. In anticipation of a Charter-based challenge, Parliament had already repealed the offending provisions of the Narcotics Control Act (which had authorized warrantless searches by “writs of assistance”) at issue in Hamill. In Yorke, the AG Canada (as Crown) had conceded at trial that the section of the Customs Act under which a search had been executed was unconstitutional, in light of three lower court rulings which had held that the section contravened the Charter. Ironically, the Supreme Court's ruling in Yorke reversed the lower courts and upheld the provision. In Johnson, the AG Canada was explicit about his reasons for taking the approach he did: the AG felt the lower court had acted inappropriately in ruling on whether the code provision (regarding sentencing long-term offenders) violated the Charter when no right had actually been claimed, and without formal argumentation by the parties. This proved a poor strategy in the Supreme Court, which upheld the lower court's finding of unconstitutionality (again without the benefit of arguments by the parties!). Finally, Ottawa really only partially conceded a claim under the Charter's s.6 mobility rights in Canadian Egg Marketing Agency, in that it agreed that s.6 applies to the Territories since they are creatures of federal legislation; they declined to address whether s.6 was actually violated, however, or whether such a violation might be reasonable under the Oakes test.
Table 2 AG Canada Litigation Strategies in Cases Challenging Federal Laws
That said, there are 36 instances of partial concessions, as indicated by the figure for “limited Charter defence.” In 15 of these the government conceded the rights violation but defended the law under s.1; these cases are discussed in more detail below, in the course of examining which Charter rights have been conceded. In the remaining 21 partial concessions the right was contested but s.1 was conceded. As with the “no Charter defence” facta, however, most of the latter group (18) were implicit concessions as s.1 was not addressed. There is no evidence that any of these types of partial concession follow a chronological pattern. This was even so for implicit concessions, where one might have expected to find evidence of a “learning curve” as governments realized that failing to address s.1 typically ensures defeat.Footnote 11 Instead, the findings reveal that implicit concessions were scattered across the twenty-year period, and in those years where there were multiple instances (five in 1988 and 1993, for example) there were also an unusually large number of cases involving federal legislation (11 and 17 respectively, compared to a median of five and mean of six).
This leaves only three cases of explicit s.1 concession: R. v. Simmons (1988), R. v. Zundel (1992), and United Nurses of Alberta v. Alberta (1992). At issue in Simmons was a s.8 challenge to the Customs Act provision that authorizes personal searches at the border. Ottawa conceded that if the s.8 right against “unreasonable search and seizure” was violated (which it did not concede), it could not advance an argument that the law was “reasonable” under s.1. A similar logic motivated the concessions in Zundel and United Nurses. Both cases concerned Criminal Code provisions that were arguably “void for vagueness” under s.7 of the Charter: a ban on “spreading falsehoods” and the preservation of the judicially defined crime of contempt of court, respectively. The AG Canada disputed both claims, but conceded that if the Court found the provisions so vague as to violate s.7, no defence could be advanced under s.1 since the standards for the Oakes test closely overlap those for vagueness. Why implicit concessions so outnumber explicit ones is unclear, but at least one facta that did so pre-dated Oakes (Hunter v. Southam, 1984), another dealt only with draft legislation in the absence of “real” facts that would be relevant to s.1 (Reference re Same-Sex Marriage, 2004), a third with a referendum law that was effectively moot (Haig v. Chief Electoral Officer of Canada, 1993), and a fourth with such sweeping rights claims that detailed s.1 defences were impractical (Canadian Council of Churches, 1992). Another nine facta ignored s.1 even though it was available, instead focusing on the related criminal process issue and the exclusion of evidence, and another focused only on the interpretation of the defence of “duress” in the Criminal Code (R. v. Ruzic, 2001).
Table 2 reveals that the government's litigation strategy is consistent regardless of whether they are appearing as an intervener or party, with full Charter defences being slightly more likely in interventions at the expense of limited Charter defences. In a somewhat counter-intuitive twist, however, full defences are slightly less likely when the AG Canada appears as an appellant (56.5%) than as a respondent (74.2%); one might reasonably assume that filing an appeal implied a strong desire to “fight,” since a government that wished to concede could simply refuse to appeal (which, as noted earlier, it does half of the time (Hennigar, Reference Hennigar2007)). On the other hand, a respondent government could concede a case before even going to a hearing; their refusal to do so may indicate a strong resistance to the rights claim being advanced.
Within the subset of Criminal Code cases, once again full Charter defences predominate (76% of facta). Partial defences appear in 13 interventions and three party facta, and no Charter defence was offered in only one intervention (R. v. Johnson, 2003), and never as party. These findings are summarized in Table 3.
Table 3 AG Canada Litigation Strategies in Cases Challenging the Criminal Code
Having established that, although rare, concessions of rights do occur when federal laws are challenged, which rights are being conceded, and whose laws, by which party? Examining only those rights which are explicitly conceded (as opposed to claims that are not addressed), Table 4 illustrates the clear pattern: the only right Ottawa regularly concedes is s.2(b)'s freedom of expression. The reason for this is also clear: beginning in the late 1980s with its ruling in cases such as Irwin Toy (1989), and affirmed consistently thereafter (see, for example, Prostitution Reference, 1990, Zundel, 1992, and RJR-MacDonald v. Canada, 1995) the Supreme Court took a “large, liberal” approach to free speech, defining “expression” as anything that nonviolently conveys meaning, regardless of content (Moon, Reference Moon2002). This definition, as the AG Canada noted in its factum in Ramsden v. Peterborough (1993), turns s.2(b) into a “barely policed port of entry into s.1,” as virtually any restriction on free speech is a prima facie violation of s.2(b) which must be defended in the context of the Oakes test. Notwithstanding its concerns, however, the AG Canada has obviously taken a cue from the Court, and does not attempt to dispute that the right to free expression has been violated; indeed, the AG often notes that such a strategy is simply unavailable given the Court's jurisprudence. The Supreme Court itself concurs. In R. v. Sharpe, cited earlier, even though some of the justices complained about the government conceding that prohibitions on child pornography violated s.2(b), they immediately wrote: “At the same time, we recognize that, at this stage, our jurisprudence leads to the conclusion that, although harmful, the content of child pornography cannot be the basis for excluding it from the scope of the s. 2(b) guarantee” (2001: para. 151). These “concessions” should not be confused for actual support for the rights claimant, however, in view of the fact that Ottawa never subsequently concedes s.1.
Table 4 Charter Rights Conceded by AG Canada in Challenges to Federal Laws, Supreme Court of Canada 1984–2004
† Only conceded application of s.6 to Territories; rights violation itself not addressed.
* Schachter v. Canada (1992)
** Does not equal the number of “no Charter defences” in Table 2 (n=5) because no right was actually claimed in one of those cases (R. v. Johnson 2003), making a s.1 defence impossible.
Section 8's right against unreasonable search and seizure was conceded in Hamill (which also conceded s.7) and Yorke, and as explained above, s.1 was not addressed in either. Violations of section 11(d)'s right to be presumed innocent were conceded in two cases, R. v. Laba (1994) and R. v. Keegstra (1996), which involved Criminal Code offenses entailing “reverse onus” (that is, the accused had to prove their innocence). Laba concerned an offense requiring someone accused of selling precious metals to prove they were the legal owners, while Keegstra (in a follow-up to his failed 1990 appeal to the SCC) challenged the anti-hate speech law's “truth” defence on the grounds that the accused had to prove “truth.” Given the Court's well-established precedent (including in Oakes, which concerned narcotics trafficking) that reverse onus offenses violate s.11(d)'s presumption of innocence, the AG Canada's concessions are not surprising. Again, however, the government defended the laws under s.1. The right to vote in s.3 was conceded twice, both in prisoner voting rights cases involving Richard Sauvé (Sauvé v. Canada, 1993 and 2002), in which the prima facie violation of s.3's unambiguous right (“every citizen of Canada has the right to vote”) was similarly unambiguous; in both, however, the violation was vigorously contested (unsuccessfully) under s.1. The Charter's s.6 mobility rights were partly conceded in Canadian Egg Marketing Agency, discussed above. Finally, the only equality rights violation conceded was in Schachter, a case that was technically moot since the legislation in question had already been amended to address the violation, and the AG's factum focused on the question of available judicial remedies.
Finally, there is some evidence of a relationship between the federal government's partisan affiliation and the decision to concede. I looked for three possible partisan factors, by posing the following questions:
1. Is one party more likely than another to concede (either fully or partially)?
2. Are governments more likely to concede a law is unconstitutional when that law was passed by a different party?
3. Is one party more likely than the other to concede a different party's law?
On the first issue, although the federal Liberals and Progressive Conservatives conceded almost exactly the same number of times (21 and 20 respectively), the Liberals participated in fewer cases when federal laws were challenged (59) than the PCs (80). As such, the Liberals conceded in a greater proportion of possible cases, 36 per cent to the Tories' 25 per cent. The second question permitted me to test Huscroft's recent assertion that “It is inconceivable that any attorney general would concede the unconstitutionality of his or her own government's legislation,” and that “concessions of unconstitutionality will only be made in regard to the legislation of prior governments” (Reference Huscroft2009: 38).Footnote 12 The overall findings do not support Huscroft's view. Taking Liberal and PC litigation together, 23 concessions were made of a different party's laws, compared to 18 concessions of a party's own laws. Of the latter group of concessions, it is noteworthy that both parties, in their last few years in government, conceded laws passed previously under their own prime minister who was still in office. This suggests that parties are no more likely to use concessions to orchestrate the invalidation of laws passed by previous governments of a different partisan stripe than they are with respect to their own laws.
This conclusion is tempered, however, by the findings regarding the third question above: while the Tories were actually less likely to concede previous Liberal laws than their own (8 versus 12 concessions), the Liberals were two and a half times more likely to concede Tory laws than their own (15 versus 6). This partisan relationship was statistically significant (Pearson's R = .317, sig. = .044). Furthermore, there is some evidence of a chronological pattern, in that the Chrétien Liberals conceded only Tory laws in the years after their 1993 return to power, until 2002 (although many of these laws were originally Liberal Criminal Code provisions retained by the Tories during the latter's overhaul of the code in 1985). The earliest Tory concessions after their election in 1984 were of their own laws, but in 1988 they conceded several Liberal laws. Thus, although there is a more complex relationship between concession and partisanship than the simple narrative that “governments concede another party's laws,” there is some truth to this view, particularly when applied to the Liberals, although it must be remembered that almost all of these concessions were partial at best.
Conclusions
Government concessions in court that laws violate the constitution are a form of inter-branch exchange about the meaning and scope of constitutional rights that have implications for our understanding of judicial “activism” and the degree to which elected officials use courts to further their own policy goals. This study—the first to provide hard empirical evidence about the prevalence of concessions—shows that overt concessions by the government of Canada before the country's highest court are rare, and complete concessions are virtually unheard of. A comprehensive examination of all AG Canada facta submitted in Charter cases in the Supreme Court from 1984 to 2004 found only a single instance of both the violation and the law's unreasonableness being explicitly conceded, and in that case the law in question had already been amended by Parliament. In the four remaining cases where no Charter defence was offered, the law had already been amended in one, had not been explicitly challenged in the second, was not actually conceded as unconstitutional in the third, and was not defended in the fourth because it had already been repeatedly ruled unconstitutional in multiple lower courts. While only a “limited” or partial Charter defence was offered in a quarter of the facta, there were often good jurisprudential reasons to do so, as the Court has made it very difficult to contest prima facie rights violations, especially with regard to freedom of expression and the presumption of innocence. In other cases, the primary focus was on the application of a given law to police procedure, and the factum (usually via intervention) focused on that issue.
The findings permit several conclusions. The first, and most obvious, is that once the federal government has decided to argue a Charter case on the merits, it overwhelmingly chooses to defend its laws in court. This may sound intuitively self-evident, but as discussed earlier in this article, there are in fact several reasons why this might not have been the case, including governments conceding in politically sensitive policy areas, attorneys general “independently” conceding laws that they believed were constitutionally indefensible, and governments conceding laws passed by previous governments of a different party. Only the last of these receives any empirical support, and only partial support at that.
A second and related conclusion is that there is little evidence of governments' demonstrating agreement with the Court over constitutional meaning in “dialogue through litigation.” Although rights violations are conceded—mostly concerning freedom of expression—the fact that these are defended as reasonable violations in every case but one (where the law in question was moot) suggests only grudging compliance with the prevailing jurisprudence about the prima facie scope of rights. Moreover, this may be a side effect of the federal government's internal process of vetting laws for possible Charter violations (see Kelly Reference Kelly2005): if Justice Department lawyers have already reviewed the law and found it defensible, why should they concede? Related to this, the low frequency of government concessions does not challenge the existing assessments of judicial activism; it is simply not the case that court rulings which invalidate (or “rewrite”) federal legislation are the result of concessions.
A final conclusion is that critics' concerns about the practice of concession seem, as an empirical matter, overblown; the example they cite, of Schachter, was the only instance found. That is not to deny the merit of their normative argument, that concessions disrespect Parliament by intentionally avoiding the regular legislative process of repeals and amendments and make government policy making less transparent. But, the low rate of concessions to date suggests that this is not a pressing problem.
In closing, however, it must be stressed that this study is limited to the SCC, which has two limitations. First, it means we are only seeing cases that the government chose to appeal; in other words, from what previous research tells us, about half of the possible concessions have already been made by the government's decisions not to appeal from lower court losses. Second, a government that wished to concede that a law was unconstitutional could do so more discreetly in a lower court (and, then refuse to appeal the loss) than in the SCC. While these observations do not negate the value of studying the SCC, it does warn against concluding too much about the judicial–executive relationship based only on concessions in legal argumentation in the highest court and points to the need for further research on the government's interaction with lower courts.
