Democracy Watch v. Canada (Attorney General)

Reasons for Judgment — Wagner C.J. (Karakatsanis, Côté, Rowe, Martin, Kasirer, Jamal, O’Bonsawin and Moreau JJ. concurring)

I. Introduction

1The rule of law is a “fundamental postulate of our constitutional structure”: Roncarelli v. Duplessis, [1959] S.C.R. 121, at p. 142. The rule of law requires, among other things, that “the exercise of all public power must find its ultimate source in a legal rule”: Reference re Remuneration of Judges of the Provincial Court of Prince Edward Island, [1997] 3 S.C.R. 3, at para. 10. Under the rule of law, all public powers must be legal, and all legal powers have limits: D. Dyzenhaus, “The Politics of Deference: Judicial Review and Democracy”, in M. Taggart, ed., The Province of Administrative Law (1997), 279, at p. 283. These limits are set out in the delegating statute, in the common or civil law, or derived from the Constitution itself: Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190, at para. 28. An unlimited power is, definitionally, not a legal power.

2This appeal concerns the constitutional role of the courts in preserving the rule of law within the administrative state. Administrative decision-makers exercise delegated authority at every level of government and in every area of policy. Their work is essential to the proper functioning of government, and has profound implications for the day-to-day lives of Canadians.

3Since at least this Court’s decision in Crevier v. Attorney General of Quebec, [1981] 2 S.C.R. 220, it has been beyond dispute that some minimum of judicial review of administrative decision-makers is guaranteed by the Constitution. The judicature provisions in ss. 96 to 101 of the Constitution Act, 1867, interpreted in light of the preamble and the principle of the rule of law, enshrine the superior courts as guardians of the rule of law and constitutionally guarantee the core aspects of their jurisdiction that are essential to their function. This includes the power of the courts to exercise their supervisory jurisdiction over the executive through judicial review. As we recognized in Dunsmuir, “judicial review is intimately connected with the preservation of the rule of law . . . [it] is the means by which the courts supervise those who exercise statutory powers, to ensure that they do not overstep their legal authority”: paras. 27‑28. Judicial review is constitutionally guaranteed, and as such, “legislatures cannot shield administrative decision making from curial scrutiny entirely”: Canada (Minister of Citizenship and Immigration) v. Vavilov, 2019 SCC 65, [2019] 4 S.C.R. 653, at para. 24.

4But what does it mean to say that legislatures cannot shield administrative decision making from curial scrutiny entirely? The appellant, Democracy Watch, submits that the constitutional guarantee of judicial review must entail a core constitutional minimum of legality review of all aspects of an administrative decision. Democracy Watch submits that this core constitutional minimum cannot be limited by ordinary legislation, and so legislative efforts to shield any aspect of an administrative decision from review are unconstitutional. The respondent, the Attorney General of Canada, submits that the core constitutional minimum of judicial review is narrower. Canada argues that only jurisdictional review is constitutionally guaranteed, meaning that legislatures may validly oust review on questions of fact or law not touching jurisdiction.

5The Federal Court of Appeal has held that judicial review on questions of fact and law is constitutionally guaranteed, and that legislative efforts to oust review on these grounds are unconstitutional: see Canada (Attorney General) v. Best Buy Canada Ltd., 2021 FCA 161, at paras. 116-17; Canada (Attorney General) v. Pier 1 Imports (U.S.), Inc., 2023 FCA 209, at paras. 29-30. However, this Court has not yet explicitly defined the constitutional guarantee of judicial review, nor directly addressed the constitutionality of legislative mechanisms that purport to limit judicial review.

6This issue is now squarely before us. To decide Democracy Watch’s appeal, we must determine if the partial privative clause in s. 66 of the Conflict of Interest Act, S.C. 2006, c. 9, s. 2 (“COIA”), is a valid bar to judicial review on questions of fact and law, or whether it is inoperative to the extent that it trenches on the constitutionally guaranteed core minimum of judicial review.

7I would allow the appeal and remand the matter to the Federal Court of Appeal. The authority to review the legality of an exercise of delegated power is the exclusive preserve of the courts, assigned to the judicial branch by the judicature provisions in ss. 96 to 101 of the Constitution Act, 1867. Intrinsic to the very idea of a constitution is the notion that legal powers and authorities are bounded; all delegated power has limits, and no authority may be used in a manner inconsistent with the purpose for which it is granted: see Reference re Secession of Quebec, [1998] 2 S.C.R. 217, at p. 248, citing Reference re Manitoba Language Rights, [1985] 1 S.C.R. 721, at p. 750; Alford v. Canada (Attorney General), 2026 SCC 14, at para. 56; see also References re Greenhouse Gas Pollution Pricing Act, 2021 SCC 11, [2021] 1 S.C.R. 175, at para. 73.

8By virtue of these features of our constitutional system, every exercise of public authority is subject to the courts’ supervisory jurisdiction to ensure that it does not exceed its legal limits. The availability of this “legality review” is constitutionally guaranteed. When read in its historical context, our jurisprudence is clear that the Constitution guarantees the availability of legality review of all aspects of an administrative decision. Section 66 of the COIA infringes this guarantee by purporting to bar judicial review on questions of fact and law: to that extent, it is of no force and effect.

II. Background

9On May 14, 2021, the Conflict of Interest and Ethics Commissioner released a report concluding that then Prime Minister Justin Trudeau had not contravened the COIA when he participated in two funding decisions relating to the WE Charity. The report was the result of an investigation initiated under s. 44 of the COIA at the request of two members of Parliament in 2020.

10Democracy Watch applied to the Federal Court of Appeal for judicial review of the Commissioner’s decision. Democracy Watch alleged that the Commissioner made errors of fact and law in his interpretation and application of the COIA.

11The Attorney General of Canada brought a motion to strike Democracy Watch’s application for judicial review. Canada argued that Democracy Watch lacked standing and that its application was barred by s. 66 of the COIA, which prohibits judicial review of the Commissioner’s decisions on questions of fact and law.

12The motion judge determined that Democracy Watch had public interest standing to bring its application: 2022 FCA 208. However, he deferred to a full panel the question of whether s. 66 was a bar to judicial review on questions of fact and law, noting that it engaged an ongoing debate in the jurisprudence over the constitutionality of partial privative clauses. In a separate decision, he ordered the proceedings to be split into two stages. Stage 1 would determine the validity and applicability of s. 66; if Democracy Watch succeeded at Stage 1, the application would proceed to Stage 2 for consideration of other issues raised by Canada: 2023 FCA 39.

13In its decision following the Stage 1 hearing, the full panel of the Federal Court of Appeal unanimously dismissed Democracy Watch’s application. It concluded that there were adequate alternative political remedies, such that judicial review was inappropriate: 2024 FCA 158. Chief Justice de Montigny would have also held that s. 66 of the COIA was valid and effectively barred judicial review on questions of fact and law. His majority colleagues disagreed. They held that the constitutional guarantee of judicial review on questions of fact and law had been conclusively decided by prior, binding decisions of their court, and dismissed the application on the sole basis of adequate alternative remedies: paras. 94-96, citing Canada (Attorney General) v. Public Service Alliance of Canada, 2019 FCA 41, 432 D.L.R. (4th) 170; Best Buy; BCE Inc. v. Québecor Média Inc., 2022 FCA 152; Pier 1 Imports (U.S.), Inc.

III. Discussion

A. Issues

14Democracy Watch asks this Court to set aside the order of the Federal Court of Appeal and remand the application to proceed to Stage 2. It bears emphasis that this appeal does not concern the merits of Democracy Watch’s application, nor any issues relating to that application beyond the following two grounds of appeal raised by Democracy Watch:

15I address each ground of appeal in turn.

B. The Federal Court of Appeal Erred in Declining Judicial Review on the Basis of an Adequate Alternative Remedy

16With respect for the contrary view of the Federal Court of Appeal, there is no adequate alternative to judicial review of the Commissioner’s decision. While the political process is tasked with the enforcement of the COIA, there is no mechanism other than judicial review whereby the legality of the Commissioner’s conclusions may be challenged directly. The Federal Court of Appeal erred in dismissing Democracy Watch’s application on this basis.

(1) Declining Judicial Review on the Basis of Adequate Alternative Remedies

17As this Court explained in Yatar v. TD Insurance Meloche Monnex, 2024 SCC 8, persons with standing to challenge administrative decisions have a right, as a matter of constitutional principle, to apply for judicial review: para. 49; see also Immeubles Port Louis Ltée v. Lafontaine (Village), [1991] 1 S.C.R. 326, at p. 360. As a corollary, the courts must at a minimum consider applications for judicial review to determine whether judicial review is appropriate in each case: para. 49. However, the decision to grant relief on an application for judicial review is discretionary; where there is a basis to decline relief, a court need not engage with the merits of the application: para. 54.

18Among the discretionary bases for refusing judicial review is the availability of an adequate alternative remedy: Yatar, at paras. 49, 54 and 56; Strickland v. Canada (Attorney General), 2015 SCC 37, [2015] 2 S.C.R. 713, at paras. 40-45. An alternative remedy is adequate only if it provides an alternative “channel for oversight of the legality . . . of administrative action”; otherwise, it will not be appropriate to defer the supervision of an impugned decision to the alternative forum: Yatar, at para. 61, citing Vavilov, at para. 24; P. Daly, Understanding Administrative Law in the Common Law World (2021), at p. 188.

19For an alternative remedy to be adequate, its process and the remedies it offers need not be identical to those available on judicial review. The question is whether the alternative remedy is adequate in all the circumstances to address the applicant’s concern: Strickland, at para. 42. A court must “consider not only the available alternative, but also the suitability and appropriateness of judicial review in the circumstances”: para. 43; see also Yatar, at para. 56. Proper consideration must be given to all relevant factors, including the relative advantages of each forum, their accessibility to the applicant, and the remedies available: Strickland, at para. 42.

20In short, where a statutory regime makes available an alternative forum in which the legality of an impugned administrative decision may be adequately reviewed, and a commensurate remedy ordered, judicial review is unnecessary to vindicate the right of the applicant. In such circumstances it is appropriate for a court to decline to review the impugned decision: Yatar, at paras. 54 and 61-62.

21The discretionary decision of a court not to consider an application for judicial review on its merits is not immune from appellate oversight, though it is owed deference: Yatar, at para. 41; Strickland, at para. 39; Canadian Pacific Ltd. v. Matsqui Indian Band, [1995] 1 S.C.R. 3, at para. 39. An exercise of discretion to decline judicial review will be vulnerable to appellate review if it was based on an error of law or a palpable and overriding error of fact, on consideration of irrelevant factors, on a failure to consider relevant factors, or on a “wrong principle”: Yatar, at para. 41, citing Canadian Pacific Ltd., at para. 112; see also Canada (Transportation Safety Board) v. Carroll‑Byrne, 2022 SCC 48, [2022] 3 S.C.R. 515, at para. 41.

(2) Reasons of the Federal Court of Appeal, 2024 FCA 158

22The Federal Court of Appeal held that the COIA provided for political oversight of the Commissioner’s activities and that this oversight was an adequate alternative to judicial supervision. It held that the COIA was intended to provide for “dual parliamentary and judicial oversight”: para. 92. The Commissioner “is an independent Officer of Parliament, and the position he occupies is firmly within the legislative branch of government . . . he acts under the direction of the House of Commons when enforcing the code to its members”: para. 80. The Commissioner is required to submit annual reports on the administration of the COIA to the Standing Committee on Access to Information, Privacy and Ethics. The Federal Court of Appeal held that this reporting mechanism provided an avenue for political oversight that was an adequate alternative to judicial review: paras. 83 and 88.

23In reaching its conclusion, the Federal Court of Appeal relied on this Court’s decision in Canada (Auditor General) v. Canada (Minister of Energy, Mines and Resources), [1989] 2 S.C.R. 49. In that case, the Auditor General sought judicial review of the Governor in Council’s decision not to order certain information be made available to him for the purposes of performing an audit. This Court declined to review the decision of the Governor in Council. Under s. 7(1) of the Auditor General Act, S.C. 1976-77, c. 34, as then in force, the Auditor General reported annually to the House of Commons “on whether, in carrying on the work of his office, he received all the information . . . required”: Auditor General, at p. 98. This Court reasoned that the Auditor General’s access to Parliament provided an adequate venue for the resolution, by political means, of any grievances the Auditor General might have about the Governor in Council’s decisions. Parliament “designated itself as final arbiter of any disputes over the Auditor General’s access to information”: p. 100. The Court emphasized that:

24According to the Federal Court of Appeal, “the same can be said of the reporting mechanism found in the COIA”: para. 88. As such, and so as to respect the “proper role” of courts in a system where the “separation of powers goes hand in hand with the rule of law”, the political remedies available under the COIA should be judged adequate, and the courts should “exercise judicial restraint and adhere to the limits prescribed by section 66 of the COIA”: paras. 89 and 92.

(3) Political Oversight of the Commissioner Is Not an Adequate Alternative to Judicial Review

25Democracy Watch submits that the court below misapplied the doctrine of adequate alternative remedy, misunderstood the statutory scheme of the COIA, and relied on distinguishable authorities to reach its conclusion. Canada submits that the court below made no error, and that the questions of fact and law raised by Democracy Watch can be adequately addressed by political mechanisms provided under the COIA.

26I agree with Democracy Watch.

27Under the COIA and the Parliament of Canada Act, R.S.C. 1985, c. P-1 (“PCA”), the Commissioner administers both the COIA and the Standing Orders of the House of Commons, App. I, Conflict of Interest and Ethics Code for Members of the House of Commons (April 2026) (“MP Code”). In this latter capacity, the PCA is explicit that the Commissioner acts under the “general direction” of the House of Commons: s. 86(3). But the PCA is equally clear that the Commissioner does not act under the direction of the House of Commons where, as here, he is administering the COIA with respect to “ministers of the Crown, ministers of state or parliamentary secretaries acting in their capacity as ministers of the Crown, ministers of state or parliamentary secretaries”: s. 86(4).

28Canada agrees that parliamentary privilege applies only to the Commissioner’s work in administering the MP Code, and not when he administers the COIA with respect to public office holders, a separate responsibility: R.F., at para. 59. But Canada endorses the Federal Court of Appeal’s analysis on the basis that it evinces the appropriate respect for Parliament’s institutional design choices and the separation of powers.

29The Federal Court of Appeal’s reasons do not demonstrate the required analysis of the appropriateness and suitability of judicial review in the circumstances: Strickland, at para. 43. Missing is any consideration of the Strickland factors, particularly the availability of an adequate alternative remedy commensurate with what is available on judicial review. As I have explained, the Federal Court of Appeal was required to satisfy itself that the COIA provided an alternative forum in which the legality of the Commissioner’s decision could be adequately reviewed: Yatar, at para. 61.

30The Federal Court of Appeal’s analysis places great weight on Auditor General. But Auditor General concerned an application to the courts made by an officer of Parliament against a decision of the Governor in Council. It was not, as here, an application brought by a private party challenging the exercise of delegated power by an officer of Parliament. Under his enabling statute, the Auditor General could go before Parliament and seek, in effect, the same remedy he sought from the courts. Under s. 7(1) he was empowered, and indeed required, to report to the House of Commons on whether he was receiving the information he needed to fulfill his mandate: p. 98. Parliament could then require that the Governor in Council make available to the Auditor General that information which he needed to fulfill his mandate, exercising its sovereign authority over the executive. Because this avenue for redress was available to him, the Court rightly declined to usurp the arbitral role Parliament had assigned itself: p. 100.

31While the Commissioner has a facially similar reporting obligation under s. 90(1)(b) of the PCA, it is hard to see how such an annual report by the Commissioner to Parliament is of any assistance to Democracy Watch. This case does not concern an application by the Commissioner; it concerns an application by Democracy Watch challenging the legality of the Commissioner’s report. Unlike in Auditor General, where an alternative to the remedy sought by the Auditor General was available, here Democracy Watch has no recourse under the COIA to ask Parliament, or anyone other than the courts, to review the Commissioner’s report.

32Indeed, the COIA is explicit that the Commissioner’s conclusions are final and “may not be altered by anyone”: s. 47. Even if Democracy Watch could bring its concerns before Parliament, it is hard to see what political remedy Democracy Watch could expect, since the report itself, if valid, cannot be altered.

33Finally, I note Chief Justice Dickson’s concluding remarks in Auditor General, where he took pains to limit his holdings to the facts of that case:

34There is nothing in the COIA that provides interested parties like Democracy Watch with an alternative forum in which to seek the remedy available on judicial review: consideration of the legality of the Commissioner’s report. While s. 47 of the COIA makes clear that the report is not determinative of enforcement measures that will be taken on the basis of the Commissioner’s conclusions, it provides explicitly that these conclusions cannot be changed by anyone. Unless the Commissioner’s decision is quashed following judicial review, his conclusions cannot be altered.

35I wish to emphasize that this conclusion does not usurp Parliament’s role as it relates to the enforcement of the COIA. Parliament retains a meaningful role in supervising the conduct of the Commissioner’s activities, and in choosing whether and how to give effect to his recommendations. But where it is alleged that the Commissioner has exercised his statutory powers in a manner that exceeds their limits, the rule of law requires that the courts be able to exercise their supervisory jurisdiction to remedy any unlawful exercise of legal authority.

36The Federal Court of Appeal erred in concluding that political oversight of the Commissioner’s activities, generally, was an adequate alternative to judicial review of his report. With respect, this conclusion was based on a flawed approach to the doctrine of adequate alternative remedy. This is an error of law that requires appellate intervention. There is no adequate alternative remedy that could justify the Federal Court of Appeal’s exercise of discretion not to consider Democracy Watch’s application for judicial review.

C. Section 66 of the COIA Cannot Bar Judicial Review on Questions of Fact and Law

37Because it concluded that there was an adequate alternative remedy, the majority of the Federal Court of Appeal did not go on to consider in any detail the question the motion judge had referred to the panel: the applicability and validity of s. 66 of the COIA. Canada submits that, even if the Federal Court of Appeal erred in concluding there was an adequate alternative remedy, it was still right to dismiss the application. Canada endorses de Montigny C.J.’s minority opinion, which concluded that s. 66 of the COIA validly ousts judicial review on questions of fact and law. Democracy Watch submits that legality review on questions of fact and law is constitutionally guaranteed, and as such s. 66 cannot bar judicial review of the Commissioner’s decision.

38I agree with Democracy Watch.

39The judicature provisions in ss. 96 to 101 of the Constitution Act, 1867 guarantee the existence of superior courts of inherent jurisdiction and protect the core, essential aspects of that jurisdiction: see Reference re Code of Civil Procedure (Que.), art. 35, 2021 SCC 27, [2021] 2 S.C.R. 291 (“CCP Reference”), at para. 63; Reference re Remuneration of Judges of the Provincial Court, at para. 124; MacMillan Bloedel Ltd. v. Simpson, [1995] 4 S.C.R. 725, at paras. 36-37; Toronto (City) v. Ontario (Attorney General), 2021 SCC 34, [2021] 2 S.C.R. 845, at para. 55; Trial Lawyers Association of British Columbia v. British Columbia (Attorney General), 2014 SCC 59, [2014] 3 S.C.R. 31, at paras. 28-29 and 38-39; U.E.S., Local 298 v. Bibeault, [1988] 2 S.C.R. 1048, at p. 1090.

40The core constitutional jurisdiction of the superior courts protected under ss. 96 to 101 encompasses “the powers and jurisdiction essential to [the superior courts’] role as the cornerstone of the unitary justice system and the primary guardians of the rule of law”: CCP Reference, at para. 63. The “review of the legality and constitutional validity of laws, enforcement of its orders, control over its own process, and its residual jurisdiction as a court of original general jurisdiction” are all included within this constitutionally protected core of the courts’ jurisdiction: para. 68.

41The supervision of the administrative state through judicial review of the exercise of public powers is one of these essential aspects of the courts’ constitutional function: Dunsmuir, at para. 31; Vavilov, at para. 24; Bibeault, at p. 1090; Attorney General of Quebec v. Farrah, [1978] 2 S.C.R. 638, at p. 654; Noël v. Société d’énergie de la Baie James, 2001 SCC 39, [2001] 2 S.C.R. 207, at para. 27. This supervisory jurisdiction allows the courts to fulfill their role as guardians of the rule of law, and ensure that Canadians are “protected from arbitrary government action”: CCP Reference, at paras. 46-51. Legislative efforts to limit the core jurisdiction of the superior courts are ultra vires: they trench on the powers assigned exclusively to the courts by ss. 96 to 101 of the Constitution Act, 1867.

42For this reason, this Court has held that a privative clause that purports to completely insulate an administrative decision-maker from judicial review is ineffective. In Crevier, Chief Justice Laskin, writing for the unanimous Court, held that a privative clause cannot shield an administrative decision-maker from judicial review on questions of jurisdiction. However, he observed that it was “now unquestioned that privative clauses may, when properly framed, effectively oust judicial review on questions of law and, indeed, on other issues not touching jurisdiction”: p. 236.

43Canada says this language from Crevier means that judicial review on questions of fact and law fall outside the scope of the constitutional guarantee, and can be barred by a partial privative clause without trenching on the constitutionally protected jurisdiction of the courts. Democracy Watch disagrees. It submits that Crevier must be read in its historical context, and that the jurisdictional review which Crevier identified as constitutionally guaranteed encompasses legality review on all aspects of an administrative decision, including questions of fact and law.

44This historical context is critical to understanding Crevier, and to identifying the proper role of privative clauses under the modern regime of administrative law. This Court’s administrative law jurisprudence has evolved alongside its understanding of the scope of s. 96 and the constitutional role of the courts. Crevier is not an outlier: it fits neatly into the long arc of Canadian administrative law away from a strict conception of judicial review as correctness review towards deference and the presumption of reasonableness review set out in Vavilov. This history illuminates the core constitutional minimum of judicial review. Our jurisprudence has long recognized that the power of the courts to review exercises of public power, so as to guard against unlawful government action in any form, is constitutionally guaranteed. Efforts to shield any aspect of an administrative decision from legality review are therefore unconstitutional.

(1) The Evolution of Jurisdictional Questions in Canadian Administrative Law

45This Court’s decision in Canadian Union of Public Employees, Local 963 v. New Brunswick Liquor Corp., [1979] 2 S.C.R. 227 (“CUPE”), is widely recognized as the demarcation between the ancien régime of Canadian administrative law and the modern era: see, e.g., Vavilov, at paras. 26 and 198; Canada (Citizenship and Immigration) v. Khosa, 2009 SCC 12, [2009] 1 S.C.R. 339, per Rothstein J., concurring; National Corn Growers Assn. v. Canada (Import Tribunal), [1990] 2 S.C.R. 1324, at pp. 1331-32, per Wilson J., concurring; Canada (Attorney General) v. Public Service Alliance of Canada, [1991] 1 S.C.R. 614, per Cory J., dissenting, but not on this point. Prior to CUPE, Canadian courts largely followed the “jurisdictional error” model of administrative review encapsulated in the House of Lords’ decision in Anisminic Ltd. v. Foreign Compensation Commission, [1969] 2 A.C. 147, which was endorsed by this Court: see, e.g., Service Employees’ International Union, Local No. 333 v. Nipawin District Staff Nurses Association, [1975] 1 S.C.R. 382, at pp. 388-89. Anisminic was the “canonical example” of the orthodox approach to judicial review, which germinated in the long shadow of Professor A. V. Dicey’s court-centric view of the rule of law: Vavilov, at paras. 206-7; see H. W. Arthurs, “Rethinking Administrative Law: A Slightly Dicey Business” (1979), 17 Osgoode Hall L.J. 1; A. V. Dicey, Introduction to the Study of the Law of the Constitution (10th ed. 1959).

46The jurisdictional error model regarded administrative decisions taken within the scope of a body’s statutory jurisdiction as unreviewable; these decisions were, in the Diceyan imagination, not “law” at all, but simply exercises of arbitrary discretion, illegible to courts and antithetical to the rule of law: see In re Ontario Labour Relations Board, [1953] 2 S.C.R. 18, at pp. 27-30; National Corn Growers Assn., at pp. 1333-34, citing Lord Hewart of Bury, The New Despotism (1945). To curb this “despotic” power, courts jealously guarded the boundaries of administrative jurisdiction; where an administrative body exceeded its legislative authority — by, for instance, erring in law, making an unreasonable finding of fact, or breaching the laws of natural justice — its decision was a nullity and the courts would adjudicate the question de novo: Anisminic, at p. 171; Executors of the Woodward Estate v. Minister of Finance, [1973] S.C.R. 120, at pp. 127-28; P. Daly, A Culture of Justification: Vavilov and the Future of Administrative Law (2023), at p. 21. A jurisdictional error also entitled the courts to ignore a privative clause, no matter how clearly or expansively worded. Where an exercise of delegated power by an administrative body exceeded the boundaries of its enabling statute, it could not rely on a privative clause within that same statute to shield its jurisdictional error from review. As Lord Wilberforce put the point in Anisminic, “[w]hat would be the purpose of defining by statute the limit of a tribunal’s powers if, by means of a clause inserted in the instrument of definition, those limits could safely be passed?”: p. 208.

47With CUPE, this Court began its move towards a “more sophisticated understanding of the role of administrative tribunals in the modern Canadian state”: National Corn Growers Assn., at p. 1336. Justice Dickson (as he then was), writing for the unanimous Court in CUPE, observed that “what is and is not jurisdictional is often very difficult to determine” and cautioned against classifying as “jurisdictional, and therefore subject to broader curial review, that which may be doubtfully so”: p. 233. He held that the legislature had intended to grant a broad authority to the administrative tribunal — a labour board — to inquire into complaints under its home statute, and that it was entitled by its privative clause to make errors of law in so doing. The only question for the reviewing court was whether, in adjudicating a complaint within its purview, the board had reached a conclusion “so patently unreasonable” as to result in a loss of jurisdiction and demand judicial correction: pp. 236-37.

48Reasonableness had long been recognized as one of the boundaries of an administrative tribunal’s jurisdiction: Kruse v. Johnson, [1898] 2 Q.B. 91 (Div. Ct.), at pp. 99-100, per Lord Russell of Killowen, C.J.; Associated Provincial Picture Houses, Ltd. v. Wednesbury Corp., [1947] 2 All E.R. 680 (C.A.), at p. 685. It is a “fundamental assumption derived from the rule of law that a legislature does not intend the power it delegates to be exercised unreasonably, or in some cases, incorrectly”: Catalyst Paper Corp. v. North Cowichan (District), 2012 SCC 2, [2012] 1 S.C.R. 5, at para. 12; see also Dunsmuir, at para. 131, per Binnie J., concurring. As Lamer J. (as he then was) explained in Blanchard v. Control Data Canada Ltd., [1984] 2 S.C.R. 476, an administrative body “has the necessary jurisdiction to make a mistake, and even a serious one, but not to be unreasonable”: p. 494.

49Nothing in CUPE displaced the principle that administrative decision-makers lacked the jurisdiction to act unreasonably. However, by focusing on reasonableness as the presumptive standard for assessing the legality of administrative action, CUPE set this Court’s jurisprudence on a path that led away from Anisminic, through Dunsmuir, to Vavilov.

50Since CUPE, the dominant approach has been to treat reasonable decisions as valid even where they are “incorrect”, insofar as they reflect an approach different from the one the court itself might have taken. The scope of issues subject to correctness review — the last vestiges of the old Anisminic approach — steadily declined in the post-CUPE era. After Vavilov, only a few narrow categories of issues attract correctness review, and then only for reasons having to do with the need to preserve consistency across Canada’s unitary legal system: Vavilov, at para. 53; Society of Composers, Authors and Music Publishers of Canada v. Entertainment Software Association, 2022 SCC 30, [2022] 2 S.C.R. 303, at para. 33; see also Paul v. British Columbia (Forest Appeals Commission), 2003 SCC 55, [2003] 2 S.C.R. 585, at para. 22.

51For decades, patently unreasonable factual and legal findings were recognized as exceeding the scope of administrative bodies’ jurisdiction: see CUPE; Blanchard, at pp. 492-93; Toronto (City) Board of Education v. O.S.S.T.F., District 15, [1997] 1 S.C.R. 487, at paras. 35-37. This Court consistently affirmed that patently unreasonable exercises of delegated power, including on questions of both fact and law, were still jurisdictional errors after CUPE. Over time, however, more and more questions came to be reviewed on that standard, rather than on a standard of correctness.

52In Canada (Director of Investigation and Research) v. Southam Inc., [1997] 1 S.C.R. 748, the Court introduced an intermediary standard of “reasonableness simpliciter”. Until Dunsmuir, developments in this area of administrative law were largely confined to the evolving approach to determining which of these three standards — patent unreasonableness, reasonableness simpliciter, or correctness — ought to be applied in a given case: Dunsmuir, at paras. 36-38.

53But the distinction between an unreasonable decision and a patently unreasonable decision proved difficult to ascertain on any consistent or principled basis. The flaws in this approach were evident early on: Dunsmuir, at para. 41, citing the comments of Justice LeBel in Toronto (City) v. C.U.P.E., Local 79, 2003 SCC 63, [2003] 3 S.C.R. 77, and Justice Abella in Council of Canadians with Disabilities v. Via Rail Canada Inc., 2007 SCC 15, [2007] 1 S.C.R. 650. The Dunsmuir Court recognized that rationality was a binary consideration; a decision either was rational or it was not, and “[a]ttaching the adjective ‘clearly’ to irrational is surely a tautology”: para. 41, citing D. J. Mullan, “Recent Developments in Standard of Review”, in Canadian Bar Association (Ontario), Taking the Tribunal to Court: A Practical Guide for Administrative Law Practitioners (2000), at p. 25.

54The Dunsmuir Court endorsed LeBel J.’s concurring comments in Toronto (City) v. C.U.P.E., at para. 108:

55While not framed in terms of “jurisdiction”, there are clear echoes of CUPE in this formulation of the standard of reasonableness. An irrational decision is an invalid decision because it is not the kind of decision an administrative decision-maker is empowered to make. Legal powers are given for purposes, and subject to limits; rationality is among the most essential of those limits.

56As reasonableness review overtook the jurisdictional error approach, questions of “jurisdiction” began to be understood more narrowly as just another category attracting correctness review. The Dunsmuir Court held that “true” questions of jurisdiction attracted correctness review, but noted that it did not “wish nor intend to return to the jurisdiction/preliminary question doctrine that plagued the jurisprudence in this area for many years” before CUPE: para. 59. Even on this narrower view, however, the slipperiness of the concept of true questions of jurisdiction made the category uncertain, and the Court increasingly questioned its utility: see Alberta (Information and Privacy Commissioner) v. Alberta Teachers’ Association, 2011 SCC 61, [2011] 3 S.C.R. 654, at paras. 33-34; Quebec (Attorney General) v. Guérin, 2017 SCC 42, [2017] 2 S.C.R. 3, at paras. 32-36; Canada (Canadian Human Rights Commission) v. Canada (Attorney General), 2018 SCC 31, [2018] 2 S.C.R. 230, at paras. 31-41.

57The critical move came in Vavilov, where this Court aimed to fully realize “Dunsmuir’s promise of simplicity and predictability” by recognizing that reasonableness is the presumptive standard of review, except where it can be shown that a question falls into the limited set of categories that attract correctness review: para. 7. With the paradigm shift to deferential review, the Court concluded that the category of “true questions of jurisdiction” was no longer useful and did away with it as a category of questions attracting correctness review: paras. 65-68.

58Vavilov also tied the standard of reasonableness back to the underlying concerns which had led CUPE to depart from the Anisminic orthodoxy. Because the legislature intends to give decision-making power to its administrative delegate, not the courts, reasonableness review balances the constitutional duty of the courts to police the boundaries of legality with the constitutional imperative to respect legislative intent: Vavilov, at paras. 33 and 68.

(2) The Evolution of the Approach to Privative Clauses

59The role of privative clauses in administrative law evolved alongside the role of jurisdictional review. Privative clauses were key to the CUPE-era move from Anisminic-style interventionism to deferential review. Privative clauses were incontrovertible evidence of legislative intent to leave certain questions with the legislature’s delegate, rather than with the courts. It was to respect these legislative signals that the Court began its move towards deference: CUPE, at pp. 235-36; Crevier, at p. 237.

60But as deferential review overtook the “jurisdictional error” orthodoxy, privative clauses became decreasingly relevant. Again, it was always acknowledged that a sufficiently serious error — quintessentially, an unreasonable exercise of delegated power — could not be shielded from review by any privative clause: see Gleason J.A.’s review in Best Buy, at paras. 78-86; and see, e.g., CUPE, at pp. 236-37; Crevier, at pp. 237-38; Blanchard, at pp. 480-81 and 492-94; Toronto (City) Board of Education, at paras. 47-48.

61In the post-CUPE era, the presence of a privative clause became a signal of legislative intent that militated in favour of reasonableness review over correctness review: Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R. 982, at paras. 30-31; Dunsmuir, at para. 52. But this Court in Dunsmuir was clear, at para. 52, that

62By the time Vavilov overtook Dunsmuir, privative clauses no longer served an independent or additional function in identifying the standard of review: Vavilov, at para. 49. Under the modern approach, reasonableness review is presumed, with or without a privative clause. This approach aims to respect the legislative intent inherent in the delegation of decision-making power. As Vavilov makes clear, legislative intent remains the “polar star” of judicial review, as it informs the scope of the grant of power to the legislature’s delegate: para. 33. The scheme of the empowering statute, which may include privative clauses, can indicate how broad a discretion the administrative decision-maker is intended to have, and so inform the “legal constraints that bear on the decision”: paras. 99 and 108-10; see also P. Daly, “Unresolved Issues after Vavilov” (2022), 85 Sask. L. Rev. 89, at p. 115. Privative clauses have gone from the paper tigers of the Anisminic era, ignored almost entirely, to an integrated part of the statutory scheme, irrelevant to the scope of the courts’ supervisory jurisdiction.

(3) The Availability of Legality Review Is Constitutionally Guaranteed

63Crevier concerned a provincial tribunal with powers so broad they contravened the prohibition in s. 96 of the Constitution Act, 1867 on provincially appointed superior courts. However, it is Laskin C.J.’s broader dictum on the constitutional guarantee of judicial review that has attracted the most attention over the years. When read in its historical context, Crevier’s statement that judicial review on jurisdictional questions is constitutionally protected must be understood to mean that the courts’ ability to conduct legality review is constitutionally guaranteed. Crevier was decided shortly after CUPE, when the jurisdictional error approach was still firmly entrenched. But it reflected the new direction of Canadian administrative law. In line with the post-CUPE move from strict, Anisminic-style correctness review to deferential review that gave greater effect to legislative intent, Crevier observed that questions of law and other questions not touching jurisdiction could be shielded from judicial review by a privative clause: pp. 236-37.

64But, as I have explained, “judicial review” in the Anisminic era meant correctness review — that is, de novo consideration of the issue before the administrative tribunal which substituted the court’s perspective for that of the tribunal. Crevier, following CUPE, meant that reasonable errors of law were not jurisdictional errors and could be shielded from correctness review. Crevier cannot have intended that unreasonable findings of fact or law could be shielded from judicial correction by a privative clause. Such a holding would have been incompatible with the whole weight of centuries of English and Canadian jurisprudence. It would also be incompatible with subsequent statements in our jurisprudence to the effect that unreasonable errors of fact and law are jurisdictional errors: see, e.g., Blanchard, at pp. 479-81, per Beetz J., and at pp. 493-95, per Lamer J.; see also Toronto (City) Board of Education, at paras. 44-46, citing Douglas Aircraft Co. of Canada v. McConnell, [1980] 1 S.C.R. 245, at p. 277, per Estey J., dissenting in part.

65Nothing in Crevier suggests that it is a departure from CUPE or out of step with the broader trend of this Court’s move from jurisdictional review to deference. Chief Justice Laskin was explicit that privative clauses could not preclude jurisdictional review. He observed that giving effect to privative clauses, particularly on questions of law, “balanced the competing interests of a provincial Legislature in its enactment of substantively valid legislation and of the courts as ultimate interpreters of the British North America Act and s. 96 thereof”: p. 237. He cautioned that:

66In the years since Crevier, our s. 96 jurisprudence has continued to undergo a “‘process of liberalization’ to adapt to modern realities”: CCP Reference, at para. 54, citing Re Residential Tenancies Act, 1979, [1981] 1 S.C.R. 714, at p. 730. This evolution has clarified those features of the judicial branch of government which are constitutionally guaranteed. It is now well established that the existence of the superior courts is guaranteed by s. 96 of the Constitution Act, 1867, and judicial review for legality — to ensure that all exercises of public power are sourced in law — is entrenched by necessary implication in the core supervisory jurisdiction of those courts: CCP Reference, at paras. 63-69; MacMillan Bloedel, at paras. 36-37; Trial Lawyers Association of British Columbia, at paras. 28-29 and 38-39.

67It is important to be clear what, precisely, is constitutionally guaranteed — that is, what the core constitutional minimum of judicial review entails. Certain parties and interveners in this case have expressed the view that reasonableness review, as defined by this Court in Vavilov, is the constitutional minimum. I would not endorse that view. As products of our Court’s jurisprudence, the common law standards of review can be and have been modified over time. Certainly, broader considerations going to the rule of law and Canada’s constitutional “culture of justification” informed this Court’s restatement of reasonableness review in Vavilov: para. 14; see also D. Dyzenhaus, “Dignity in Administrative Law: Judicial Deference in a Culture of Justification” (2012), 17:1 Rev. Const. Stud. 87; the Rt. Hon. B. McLachlin, “The Roles of Administrative Tribunals and Courts in Maintaining the Rule of Law” (1998), 12 C.J.A.L.P. 171, at p. 174; M. Cohen-Eliya and I. Porat, “Proportionality and Justification” (2014), 64 U.T.L.J. 458, at pp. 467-70. That is not to say, however, that the reasonableness standard articulated in Vavilov is constitutionally entrenched. What is clear is that the availability of legality review — that is, review to ensure that an exercise of public power has not exceeded the boundaries of an administrative decision-maker’s delegated authority — is constitutionally guaranteed.

68The courts of Canada, like the courts of England, have consistently held that irrational administrative decisions are unlawful, insofar as they exceed the inherent limits of delegated authority. Irrational decisions raise the spectre of arbitrary exercises of public power; the power of the courts to protect against such arbitrariness lies, as I have explained, at the very heart of the courts’ constitutionally protected jurisdiction under ss. 96 to 101.

69There can be no doubt, then, that rationality review has a constitutional aspect. Rationality is a boundary of legality. Moreover, deferential judicial review strikes the constitutionally necessary balance between respect for the democratic sovereignty of the legislature, respect for the competence and expertise of the executive, and respect for the courts as arbiters of legality: CCP Reference, at paras. 46-48; Cooper v. Canada (Human Rights Commission), [1996] 3 S.C.R. 854, at paras. 10-13, per Lamer C.J., concurring; see also T. Bingham, The Rule of Law (2010), at p. 61. Judicial deference to the legislature and the executive evinces respect without submission — the role of the courts remains essential, but the rule of law is not the rule of courts. Provided that courts can exercise their supervisory jurisdiction over the legality of all exercises of public power, the rule of law does not preclude meaningful decision-making by other state actors: see D. Dyzenhaus, “Disobeying Parliament? Privative Clauses and the Rule of Law”, in R. W. Bauman and T. Kahana, eds., The Least Examined Branch: The Role of Legislatures in the Constitutional State (2006), 499, at p. 514.

70Having said that, the question of what minimum standard of review is constitutionally guaranteed is not before us, and should therefore be left to another day. Democracy Watch’s appeal does not require us to decide when a legislated standard of review will trench on the constitutional core of the courts’ supervisory jurisdiction and offend “the limits imposed by the rule of law”: Vavilov, at paras. 34-35.

71Legislative efforts to limit legality review are ultra vires not because reasonableness review, per se, is constitutionally entrenched, but because the Constitution guarantees the courts’ role in ensuring that all exercises of public power, as they manifest in all aspects of an administrative decision, are sourced in law. Where an administrative decision-maker acts in an official capacity and exercises a delegated power, it does not matter whether that exercise of power concerns questions of fact, of law, of fairness, or any other issue: see Blanchard, at p. 494, per Lamer J. As Chief Justice Khullar explained in her dissenting judgment in Northback Holdings Corporation v. Alberta Energy Regulator, 2025 ABCA 186, 510 D.L.R. (4th) 488, the notion advanced by some academic commentators that judicial review is guaranteed only on questions of law has no sound historical or jurisprudential basis: at paras. 210-22. What matters, for constitutional purposes, is that the courts are constitutionally empowered to supervise the exercise of delegated power to ensure it respects the boundaries of its inherent limits.

72Legislation that purports to oust this constitutional power of the courts oversteps the limits of the legislatures’ constitutional authority. Such efforts are ultra vires; they do not affect the scope of the courts’ supervisory jurisdiction. This is true for the s. 96 superior courts, but also where the Federal Courts have been assigned exclusive jurisdiction over judicial review under the Federal Courts Act, R.S.C. 1985, c. F-7.

73Section 101 of the Constitution Act, 1867 authorizes Parliament to transfer some of the jurisdiction of the s. 96 courts to courts constituted for “the better Administration of the Laws of Canada”, as it did through the Federal Courts Act: Canada Labour Relations Board v. Paul L’Anglais Inc., [1983] 1 S.C.R. 147, at pp. 153-54 and 162. But the transfer cannot eliminate any portion of that jurisdiction; the jurisdiction of the s. 96 courts, whether exercised by s. 96 courts themselves or by courts constituted under s. 101, is constitutionally guaranteed. Where Parliament has assigned a matter to the Federal Courts, Parliament cannot combine that allocation of jurisdiction with a limit on judicial review that would circumvent the constitutional guarantee of legality, and the limits that inhere in any delegation of public powers. In other words, while s. 101 permits Parliament to create new courts to exercise the jurisdiction protected by s. 96, it does not permit Parliament to limit or circumvent that jurisdiction entirely. A privative clause that prevents the Federal Courts from exercising a component of the s. 96 supervisory jurisdiction assigned to them is contrary to the Constitution Act, 1867 and of no force and effect.

74In determining whether a statutory provision is unconstitutional for this reason, courts must consider its substance: the question is whether, properly interpreted, the impugned provision has the effect of ousting judicial review for legality of any aspect of an administrative decision.

75To be clear, legislatures are competent to prescribe procedures for judicial review, provided that the supervisory role of the courts is not ousted. Legislation may, for example, validly set out specific timelines or fora for judicial review of certain issues without functionally foreclosing judicial review for legality. In the same vein, as already noted, legislated standards of review are valid, provided they respect the limits of the rule of law: Vavilov, at para. 35. And, of course, the common law bases to decline to consider the merits of an application for judicial review — for instance, the presence of an adequate alternative remedy, non-justiciability, mootness, or prematurity — are fully consistent with the requirements of the rule of law: they constitute an exercise of, not a limit on, the supervisory jurisdiction of the courts.

76But the legality of every aspect of an administrative decision, and every exercise of public power, is subject to the supervision of the courts. Where a statutory provision, properly construed, purports to limit any aspect of the courts’ constitutionally enshrined supervisory jurisdiction, that provision is ultra vires.

(4) Application to Section 66 of the COIA

77Section 66 of the COIA provides that:

78Section 66 is therefore explicit in limiting the scope of judicial review to the grounds set out in s. 18.1(4)(a), (b) and (e) of the Federal Courts Act. This excludes the following additional grounds that appear in the remainder of s. 18.1(4):

79Democracy Watch submits that the jurisdictional review referred to by s. 18.1(4)(a), understood in light of the history canvassed above, includes legality review on questions of fact and law. Since s. 66 permits review on the grounds included under s. 18.1(4)(a), Democracy Watch submits that its application for judicial review on questions of fact and law is not barred by s. 66. Alternatively, if s. 66 does bar judicial review on questions of fact and law, Democracy Watch submits that s. 66 is unconstitutional and should be declared of no force and effect.

80While courts have often given little or no effect to privative clauses in their analysis, they have hesitated to say explicitly that such clauses cannot constitutionally accomplish what, on their face, they were intended to accomplish. Instead, courts have tended to read these clauses into constitutional compliance. The privative clause in Dunsmuir, for example, stated explicitly that “every order, award, direction, decision, declaration or ruling of the Board, an arbitration tribunal or an adjudicator is final and shall not be questioned or reviewed in any court”: Public Service Labour Relations Act, R.S.N.B. 1973, c. P-25, s. 101(1). Despite this unambiguous language, the Dunsmuir Court held that the provision merely “gives rise to a strong indication that the reasonableness standard of review will apply”: para. 67.

81Democracy Watch’s proposed interpretation of s. 66 conforms to this tendency. Rather than state outright that s. 66 is ultra vires to the extent that it bars legality review on questions of fact and law, it invites this Court to read such questions into the ground of jurisdictional review permitted under s. 66.

82I am disinclined to take this approach. It is true that, in interpreting the text of s. 66 of the COIA in context and in light of its purpose, it is presumed that Parliament intended to respect its constitutional constraints: see R. v. Mills, [1999] 3 S.C.R. 668, at para. 56; R. v. J.J., 2022 SCC 28, [2022] 2 S.C.R. 3, at para. 18. The presumption of constitutionality is an important principle of statutory interpretation, but there are limits: where legislation is not amenable to a constitutionally compliant interpretation, the legislation must be declared invalid (Slaight Communications Inc. v. Davidson, [1989] 1 S.C.R. 1038, at p. 1078, per Lamer J. (as he then was), dissenting in part, but not on this point).

83The courts’ tendency to read privative clauses into constitutional compliance has strained the limits of the presumption of constitutionality. Courts must not assign to these provisions meanings that are fundamentally incompatible with the intent of the enacting legislature.

84Here, I cannot accept that s. 66, read together with s. 18.1(4) of the Federal Courts Act, permits legality review on questions of fact and law. By expressly separating out review on questions of fact and law from review on jurisdictional grounds, s. 66 of the COIA suggests that Parliament’s intent was to invoke “jurisdiction” in its narrower, modern sense, analogous to the category of “true questions of jurisdiction” with which this Court grappled until Vavilov. Parliament’s clear intent in drafting s. 66 so as to expressly exclude review on questions of fact and law cannot be ignored.

85The better approach is to recognize that s. 66, properly interpreted, purports to oust judicial review on questions of fact and law. To the extent it does so, s. 66 is inconsistent with ss. 96 to 101 of the Constitution Act, 1867 and, pursuant to s. 52(1) of the Constitution Act, 1982, should be declared of no force and effect.

IV. Conclusion

86All legal powers are subject to inherent limits. Under a constitutional order based in the rule of law, there is no such thing as an unfettered discretion: Roncarelli, at p. 140. The courts are tasked with ensuring that legal powers are exercised in accordance with their limits: Vavilov, at para. 67. Parliament and the legislatures are not constitutionally competent to shield administrative exercises of delegated power from judicial review for legality on any ground; such efforts trench on the core jurisdiction of the courts and are ultra vires.

87The privative clause in s. 66 of the COIA should be declared of no force and effect to the extent that it purports to bar judicial review on questions of fact and law. It poses no obstacle to Democracy Watch’s application for judicial review. Furthermore, the Federal Court of Appeal erred in dismissing the application on the basis that there is an adequate alternative to judicial review.

88For these reasons, I would declare that s. 66 of the COIA is of no force and effect to the extent it bars judicial review on questions of fact and law. I would allow Democracy Watch’s appeal with costs, set aside the order of the Federal Court of Appeal dismissing the application, and remand the matter to the court below to proceed to Stage 2.