Canada (Citizenship and Immigration) v. Khosa, 2009 SCC 12, [2009] 1 SCR 339

                                                 SUPREME COURT OF CANADA

Between:

Minister of Citizenship and Immigration

Appellant

and

Sukhvir Singh Khosa

Respondent

‑ and ‑

Immigration and Refugee Board

Intervener

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

*  Bastarache J. took no part in the judgment.

______________________________

Canada (Citizenship and Immigration) v. Khosa, 2009 SCC 12, [2009] 1 S.C.R. 339

Minister of Citizenship and Immigration                                                        Appellant

v.

Sukhvir Singh Khosa                                                                                     Respondent

and

Immigration and Refugee Board                                                                    Intervener

Indexed as:  Canada (Citizenship and Immigration) v. Khosa

Neutral citation:  2009 SCC 12.

File No.:  31952.

2008:  March 20; 2009:  March 6.

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

on appeal from the federal court of appeal

Administrative law — Judicial review — Standard of review — Immigration Appeal Division denying special relief on “humanitarian and compassionate grounds” from removal order — Standard of review applicable to Immigration Appeal Division decision — Whether common law of judicial review displaced by s. 18.1  of Federal Courts Act, R.S.C. 1985, c. F‑7  — Immigration and Refugee Protection Act, S.C. 2001, c. 27, s. 67(1) (c).

K, a citizen of India, immigrated to Canada with his family in 1996, at the age of 14.  In 2002, he was found guilty of criminal negligence causing death and received a conditional sentence of two years less a day.  A valid removal order was issued to return him to India.

K appealed the order, but the majority of the Immigration Appeal Division (“IAD”) of the Immigration and Refugee Board, after considering the Ribic factors and the evidence, denied “special relief” on humanitarian and compassionate grounds pursuant to s. 67(1) (c) of the Immigration and Refugee Protection Act  (“IRPA ”).  A majority of the Federal Court of Appeal applied a “reasonableness” simpliciter standard and set aside the IAD decision.  It found that the majority of the IAD had some kind of fixation with the fact that the offence was related to street‑racing.  On the issue of the “possibility of rehabilitation”, the majority of the IAD merely acknowledged the findings of the criminal courts in that regard, which were favourable to K, and did not explain why it came to the contrary conclusion.  In the end, that court concluded that the majority of the IAD had acted unreasonably in denying relief.

Held (Fish J. dissenting):  The appeal should be allowed.

Per McLachlin C.J. and Binnie, LeBel, Abella and Charron JJ.:  This Court’s decision in Dunsmuir, which was released after the decisions of the lower courts in this case, recognized that, with or without a privative clause, a measure of deference has come to be accepted as appropriate where a particular decision has been allocated to administrative decision-makers in matters that relate to their special role, function and expertise.  A measure of deference is appropriate whether or not the court has been given the advantage of a statutory direction, explicit or by necessary implication.  These general principles of judicial review are not ousted by s. 18.1  of the Federal Courts Act  which deals essentially with grounds of review of administrative action, not standards of review. [25]

A legislature has the power to specify a standard of review if it manifests a clear intention to do so.  However, where the legislative language permits, the court (a) will not interpret grounds of review as standards of review, (b) will apply Dunsmuir principles to determine the appropriate approach to judicial review in a particular situation, and (c) will presume the existence of a discretion to grant or withhold relief based in part on Dunsmuir including a restrained approach to judicial intervention in administrative matters.  [51]

Resort to the flexibility of the general principle of judicial review is all the more essential in the case of a provision like s. 18.1  of the Federal Courts Act  which is not limited to particular issues before a particular adjudicative tribunal but covers the full galaxy of federal decision-makers who operate in different decision‑making environments under different statutes with distinct grants of decision‑making powers.  [28] [33]

The language of s. 18.1 generally sets out threshold grounds which permit but do not require the court to grant relief.  Despite a difference in the meaning of the English and French versions in the relevant language of s. 18.1(4), the provision should be interpreted to permit a court to exercise its discretion in matters of remedy depending on the court’s appreciation of the respective roles of the courts and the administration as well as the circumstances of each case.  The discretion must be exercised judicially, but the appropriate judicial basis for its exercise includes the general principles dealt with in Dunsmuir.  [36]

Dunsmuir establishes that there are now only two standards of review:  correctness and reasonableness.  No authority was cited suggesting that a “correctness” standard of review is appropriate for IAD decisions under s. 67(1) (c) of the IRPA , and the relevant factors in a standard of review inquiry point to a reasonableness standard.  These factors include:  (1) the presence of a privative clause; (2) the purpose of the IAD as determined by its enabling legislation — the IAD determines a wide range of appeals under the IRPA and its decisions are reviewable only if the Federal Court grants leave to commence judicial review; (3) the nature of the question at issue before the IAD — Parliament has provided in s. 67(1)(c) a power to grant exceptional relief and this provision calls for a fact‑dependent and policy‑driven assessment by the IAD itself; and (4) the expertise of the IAD dealing with immigration policy.  These factors must be considered as a whole, bearing in mind that not all factors will necessarily be relevant for every single case.  [53‑57]

Where, as here, the reasonableness standard applies, it requires deference.  Reviewing courts ought not to reweigh the evidence or substitute their own appreciation of the appropriate solution, but must rather determine if the outcome falls within a range of reasonable outcomes.  In this case, the question whether K had established “sufficient humanitarian and compassionate considerations” to warrant relief from his removal order was a decision which Parliament confided to the IAD, not to the courts.  [4] [59]

The IAD reasons, both the majority and dissent, disclose with clarity the considerations in support of both points of view, and the reasons for the disagreement as to outcome.  At the factual level, the IAD divided in large part over differing interpretations of K’s expression of remorse.  This is the sort of factual dispute which should be resolved by the IAD not the courts.  The majority considered each of the Ribic factors, reviewed the evidence and decided that, in the circumstances of this case, discretionary relief should be refused.  While the findings of the criminal courts on the seriousness of the offence and possibility of rehabilitation (the first and second of the Ribic factors), were properly noted, the IAD had a mandate different from that of the criminal courts.  The issue before it was not the potential for rehabilitation for purposes of sentencing, but rather whether the prospects for rehabilitation were such that, alone or in combination with other relevant factors, they warranted special discretionary relief from a valid removal order.  The IAD was required to reach its own conclusions based on its own appreciation of the evidence and it did so.  [64‑66]

In light of the deference properly owed to the IAD under s. 67(1) (c) of the IRPA , there was no proper basis for the Federal Court of Appeal to interfere with the IAD decision to refuse special relief in this case.  It cannot be said that this decision fell outside the range of reasonable outcomes.  [60] [67]

Per Rothstein J.:  Where a legislature has expressly or impliedly provided for standards of review, courts must follow that legislative intent, subject to any constitutional challenge.  With respect to s. 18.1(4)  of the Federal Courts Act , the language of para. (d) makes clear that findings of fact are to be reviewed on a highly deferential standard.  Courts are only to interfere with a decision based on erroneous findings of fact where the federal board, commission or other tribunal’s factual finding was “made in a perverse or capricious manner or without regard for the material before it”.  By contrast with para. (d), there is no suggestion that courts should defer in reviewing a question that raises any of the other criteria in s. 18.1(4) .  Where Parliament intended a deferential standard of review in s. 18.1(4) , it used clear and unambiguous language, as it has in para. (d) regarding facts.  The necessary implication is that where Parliament did not provide for deferential review, it intended the reviewing court to apply a correctness standard as it does in the regular appellate context.  [70] [72] [113] [117]

While recourse to the common law is appropriate where Parliament has employed common law terms or principles without sufficiently defining them, it is not appropriate where the legislative scheme or provisions expressly or implicitly ousts the relevant common law analysis as is the case with s. 18.1(4)  of the Federal Courts Act .  Courts must give effect to the legislature’s words and cannot superimpose on them a duplicative common law analysis.  The Dunsmuir standard of review should be confined to cases in which there is a strong privative clause.  Excepting such cases, it does not apply to s. 18.1(4) .  The application of Dunsmuir outside the strong privative clause context marks a departure from the conceptual and jurisprudential origins of the standard of review analysis.  [70] [74] [106] [136]

The deference approach emerged as a means of reconciling Parliament’s intent to immunize certain administrative decisions from review with the supervisory role of courts in a rule of law system.  The creation of expert administrative decision-makers evidenced a legislative intent to displace or bypass the courts as primary adjudicators in a number of areas, but it was only with the enactment of privative clauses, which marked the area of tribunal expertise that the legislature was satisfied warranted deference, that a legislature indicated an intent to oust, or at the very least restrict, the court’s review role.  Whereas tribunal expertise was a compelling rationale for imposing a privative clause, it was not a free‑standing basis for deference.  The approach of judicially imputing expertise which followed, even on questions of law, was a departure from earlier jurisprudence that relied on privative clauses as the manifest signal of the legislature’s recognition of relative tribunal expertise.  [79] [82‑84] [87]

There is no dispute that reviewing courts, whether in the appellate or judicial review contexts, should show deference to lower courts and administrative decision-makers on questions of fact and on questions involving mixed fact and law, where a legal issue cannot be extricated from a factual or policy finding.  However, where a legal issue can be extricated from a factual or policy inquiry, it is inappropriate to presume deference where Parliament has not indicated this via a privative clause.  It is not for the court to impute tribunal expertise on legal questions, absent a privative clause and, in doing so, assume the role of the legislature to determine when deference is or is not owed.  Recognizing expertise as a free‑standing basis for deference on questions that reviewing courts are normally considered to be expert on departs from the search for legislative intent that governs this area.  [89-93]

Concerns regarding the rigidity of the legislated standards are misplaced.  A review of the Federal Courts Act  makes clear that the focus of the analysis should be on the nature of the question under review and not on the type of administrative decision-maker.  Even given this legislative focus on the nature of the question under review, not all administrative decision-makers will be subject to the same standards of review.  Where a decision-maker’s enabling statute purports to preclude judicial review on some or all questions through a privative clause, deference will apply and a Dunsmuir standard of review analysis will be conducted.  [109‑110]

Section 18.1(4) confers on the Federal Courts the discretion to grant or deny relief in judicial review.  The remedial discretion in s. 18.1(4) goes to the question of withholding relief, not the review itself.  The traditional common law discretion to refuse relief on judicial review concerns the parties’ conduct, any undue delay and the existence of alternative remedies which is wholly distinct from the common law of standard of review analysis.  Reliance upon this discretion contained in s. 18.1(4) to support the view that it opens the door to the Dunsmuir standard of review analysis is inappropriate.  [131] [135-136]

The IAD’s decision not to grant relief in this case should be upheld.  The application of the Ribic factors to the case before it and its exercise of discretion is fact‑based.  The IAD’s factual findings were not perverse or capricious and were not made without regard to the evidence.  [137]

Per Deschamps J.:  There is agreement with Rothstein J. that since s. 18.1(4)  of the Federal Courts Act  sets legislated standards of review, those standards oust the common law. [138]

Per Fish J. (dissenting):  The standard of review applicable is “reasonableness”, and the IAD’s decision does not survive judicial scrutiny under that standard.  The IAD’s task was to look to “all the circumstances of the case” in order to determine whether “sufficient humanitarian and compassionate considerations” existed to warrant relief from a removal order.  The IAD placed the greatest weight on three factors:  K’s remorse, rehabilitation, and likelihood of reoffence.  Despite abundant evidence that K was extremely unlikely to reoffend and had taken responsibility for his actions, the IAD focussed on a single fact — K’s denial that he was “street racing” — and based its refusal to grant relief largely on that fact alone.  While K’s denial may well evidence some “lack of insight”, it cannot be said to contradict — still less to outweigh, on a balance of probabilities — all of the evidence in his favour on the issues of remorse, rehabilitation and likelihood of reoffence.  The IAD’s cursory treatment of the sentencing judge’s favourable findings on remorse and the risk of recidivism are particularly troubling.  While a criminal court’s findings are not necessarily binding upon an administrative tribunal with a distinct statutory purpose and a different evidentiary record, it was incumbent upon the IAD to consider those findings and to explain the basis of its disagreement with the sentencing judge’s decision.  K’s denial of street racing is, at best, of little probative significance in determining his remorse, rehabilitation and likelihood of reoffence.  The IAD’s conclusion that there was “insufficient evidence” upon which a determination could be made that K does not represent a risk to the public is not only incorrect, but unreasonable.  Decisions of the IAD are entitled to deference, but deference ends where unreasonableness begins.  [139‑140] [145] [147] [149‑151] [153‑154] [160]

Cases Cited

By Binnie J.

Applied:  Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190; distinguished:  R. v. Owen, 2003 SCC 33, [2003] 1 S.C.R. 779; referred to:  R. v. Khosa, 2003 BCCA 645, 190 B.C.A.C. 42; Ribic v. Canada (Minister of Employment and Immigration), [1985] I.A.B.D. No. 4 (QL); Chieu v. Canada (Minister of Citizenship and Immigration), 2002 SCC 3, [2002] 1 S.C.R. 84; Roncarelli v. Duplessis, [1959] S.C.R. 121; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235; Pezim v. British Columbia (Superintendent of Brokers), [1994] 2 S.C.R. 557; Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R. 982; Penetanguishene Mental Health Centre v. Ontario (Attorney General), 2004 SCC 20, [2004] 1 S.C.R. 498; Pinet v. St. Thomas Psychiatric Hospital, 2004 SCC 21, [2004] 1 S.C.R. 528; Mazzei v. British Columbia (Director of Adult Forensic Psychiatric Services), 2006 SCC 7, [2006] 1 S.C.R. 326; Pringle v. Fraser, [1972] S.C.R. 821; Howarth v. National Parole Board, [1976] 1 S.C.R. 453; Martineau v. Matsqui Institution Disciplinary Board, [1980] 1 S.C.R. 602; Attorney General of Canada v. Law Society of British Columbia, [1982] 2 S.C.R. 307; Harelkin v. University of Regina, [1979] 2 S.C.R. 561; R. v. Daoust, 2004 SCC 6, [2004] 1 S.C.R. 217; Slaight Communications Inc. v. Davidson, [1989] 1 S.C.R. 1038; Solosky v. The Queen, [1980] 1 S.C.R. 821; Canada (Auditor General) v. Canada (Minister of Energy, Mines and Resources), [1989] 2 S.C.R. 49; Friends of the Oldman River Society v. Canada (Minister of Transport), [1992] 1 S.C.R. 3; Canada v. Grenier, 2005  CAF 348, [2006] 2 F.C.R. 287; Devinat v. Canada (Commission de l’immigration et du statut de réfugié), [2000] 2 F.C. 212; Thanabalasingham v. Canada (Minister of Citizenship and Immigration), 2006 FCA 14, 263 D.L.R. (4th) 51; Charette v. Canada (Commissioner of Competition), 2003 FCA 426, 29 C.P.R. (4th) 1; Pal v. Canada (Minister of Employment and Immigration) (1993), 24 Admin. L.R. (2d) 68; Mugesera v. Canada (Minister of Citizenship and Immigration), 2005 SCC 40, [2005] 2 S.C.R. 100; Bell v. Ontario Human Rights Commission, [1971] S.C.R. 756; Canadian Union of Public Employees, Local 963 v. New Brunswick Liquor Corp., [1979] 2 S.C.R. 227; Prata v. Minister of Manpower and Immigration, [1976] 1 S.C.R. 376; Baker v. Canada (Minister of Citizenship and Immigration), [1999] 2 S.C.R. 817.

By Rothstein J.

Not followed:  Pezim v. British Columbia (Superintendent of Brokers), [1994] 2 S.C.R. 557; applied:  R. v. Owen, 2003 SCC 33, [2003] 1 S.C.R. 779; considered: Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190; discussed:  Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R. 982; referred to:  U.E.S., Local 298 v. Bibeault, [1988] 2 S.C.R. 1048; Pasiechnyk v. Saskatchewan (Workers’ Compensation Board), [1997] 2 S.C.R. 890; Canada (Attorney General) v. Mossop, [1993] 1 S.C.R. 554; Canadian Union of Public Employees, Local 963 v. New Brunswick Liquor Corp., [1979] 2 S.C.R. 227; Bell Canada v. Canada (Canadian Radio‑Television and Telecommunications Commission), [1989] 1 S.C.R. 1722; R. v. Robinson, [1996] 1 S.C.R. 683; Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235; C.U.P.E. v. Ontario (Minister of Labour), 2003 SCC 29, [2003] 1 S.C.R. 539; R. v. Holmes, [1988] 1 S.C.R. 914; Waldick v. Malcolm, [1991] 2 S.C.R. 456; Gendron v. Supply and Services Union of the Public Service Alliance of Canada, Local 50057, [1990] 1 S.C.R. 1298; Mugesera v. Canada (Minister of Citizenship and Immigration), 2005 SCC 40, [2005] 2 S.C.R. 100; Morneault v. Canada (Attorney General), [2001] 1 F.C. 30; Harelkin v. University of Regina, [1979] 2 S.C.R. 561; Immeubles Port Louis Ltée v. Lafontaine (Village), [1991] 1 S.C.R. 326; Ribic v. Canada (Minister of Employment and Immigration), [1985] I.A.B.D. No. 4 (QL); Chieu v. Canada (Minister of Citizenship and Immigration), 2002 SCC 3, [2002] 1 S.C.R. 84.

By Fish J. (dissenting)

Ribic v. Canada (Minister of Employment and Immigration), [1985] I.A.B.D. No. 4 (QL); Chieu v. Canada (Minister of Citizenship and Immigration), 2002 SCC 3, [2002] 1 S.C.R. 84; R. v. Khosa, 2003 BCSC 221, [2003] B.C.J. No. 280 (QL); R. v. Khosa, 2003 BCCA 644, 190 B.C.A.C. 23.

Statutes and Regulations Cited

Administrative Tribunals Act, S.B.C. 2004, c. 45, ss. 58, 59.

Back to School Act, 1998, S.O. 1998, c. 13, s. 18(3).

Canada Agricultural Products Act , R.S.C. 1985, c. 20 (4th Supp .), s. 10(1.1).

Canada Labour Code , R.S.C. 1985, c. L‑2 , s. 22(1) .

Certified General Accountants Act, C.C.S.M., c. C46, s. 22(2).

Code of Civil Procedure, R.S.Q., c. C‑25, s. 846.

Corrections and Conditional Release Act , S.C. 1992, c. 20 , s. 147(1) .

Criminal Code , R.S.C. 1985, c. C‑46 , ss. 672 , 672.38 , 672.54 , 672.78 .

Education Labour Relations Act, R.S.Y. 2002, c. 62, s. 95(1).

Employment Insurance Act , S.C. 1996, c. 23 , s. 115(2) .

Federal Courts Act , R.S.C. 1985, c. F‑7 , ss. 2 , 18 , 18.1 .

Gaming Control Act, C.C.S.M., c. G5, s. 45(2).

Health Professions Act, S.Y. 2003, c. 24, s. 29.

Human Rights Code, C.C.S.M., c. H175, s. 50(1).

Immigration and Refugee Protection Act , S.C. 2001, c. 27 , ss. 3(1) (h), 36(1) (a), 63 ,  67(1) (c), 72 , 162(1) .

Interpretation Act , R.S.C. 1985, c. I‑21 , s. 11 .

Judicial Review Act, R.S.P.E.I. 1988, c. J‑3, s. 4(1).

Judicial Review Procedure Act, R.S.O. 1990, c. J.1, s. 2.

Liquor Act, R.S.Y. 2002, c. 140, s. 118(1).

Occupational Health and Safety Act, S.N.B. 1983, c. O‑0.2, s. 26(5).

Rehabilitation Services Act, R.S.Y. 2002, c. 196, s. 7.

Residential Tenancies Act, S.N.B. 1975, c. R‑10.2, s. 27(1).

Traffic Safety Act, R.S.A. 2000, c. T‑6, s. 47.1(3).

Urban and Rural Planning Act, 2000, S.N.L. 2000, c. U‑8, s. 46(1).

Youth Protection Act, R.S.Q., c. P‑34.1, s. 74.2.

Authors Cited

Bastarache, Michel, et al.  The Law of Bilingual Interpretation.  Markham, Ont.:  LexisNexis Canada, 2008.

Bogart, W. A.  “The Tools of the Administrative State and the Regulatory Mix”, in C. M. Flood and L. Sossin, eds., Administrative Law in Context.   Toronto: Emond Montgomery, 2008, 25.

British Columbia.  Legislative Assembly.  Debates of the Legislative Assembly, 5th Sess., 37th Parl., May 18, 2004, p. 11193.

Brown, Donald J. M., and John M. Evans. Judicial Review of Administrative Action in Canada.  Toronto:  Canvasback, 1998 (loose‑leaf updated July 2008).

Canada.  House of Commons.  Standing Committee on Justice and Legal Affairs.  Minutes of Proceedings and Evidence, No. 26, 2nd Sess., 28th Parl., May 7, 1970, pp. 25‑26.

Charney, Richard J., and Thomas E. F. Brady.  Judicial Review in Labour Law.  Aurora, Ont.:  Canada Law Book, 1997 (loose-leaf updated 2008, release 11).

Côté, Pierre‑André.  The Interpretation of Legislation in Canada, 3rd ed.  Scarborough, Ont.:  Carswell, 2000.

Dicey, Albert Venn.  Introduction to the Study of the Law of the Constitution, 10th ed.  London:  Macmillan, 1959.

Dyzenhaus, David.  “Disobeying Parliament?   Privative Clauses and the Rule of Law”, in Richard W. Bauman and Tsvi Kahana, eds., The Least Examined Branch:  The Role of Legislatures in the Constitutional State.  New York: Cambridge University Press, 2006, 499.

Fuller, Lon L.  The Morality of Law, rev. ed.  New Haven:  Yale University Press, 1969.

Liston, Mary.  “Governments in Miniature:  The Rule of Law in the Administrative State”, in C. M. Flood and L. Sossin, eds., Administrative Law in Context.  Toronto: Emond Montgomery, 2008, 77.

Macklin, Audrey.  “Standard of Review:  The Pragmatic and Functional Test”, in C. M. Flood and L. Sossin, eds., Administrative Law in Context.  Toronto: Emond Montgomery, 2008, 197.

Mullan, David.  “Dunsmuir v. New Brunswick, Standard of Review and Procedural Fairness for Public Servants:  Let’s Try Again!” (2008), 21 C.J.A.L.P. 117.

Mullan, David J. “Establishing the Standard of Review:  The Struggle for Complexity?” (2004), 17 C.J.A.L.P. 59.

Raz, Joseph.  The Authority of Law:  Essays on Law and Morality.  Oxford:  Clarendon, 1979.

Saunders, Brian J., et al.  Federal Courts Practice 2009.  Toronto:  Thomson Carswell, 2008.

Sgayias, David, et al.  Federal Court Practice 1998.  Scarborough, Ont.:  Carswell, 1997.

Sossin, Lorne.  “Empty Ritual, Mechanical Exercise or the Discipline of Deference?  Revisiting the Standard of Review in Administrative Law” (2003), 27 Advocates’ Q. 478.

Sullivan, Ruth.  Sullivan on the Construction of Statutes, 5th ed.  Markham, Ont.:  LexisNexis, 2008.

APPEAL from a judgment of the Federal Court of Appeal (Desjardins, Décary and Malone JJ.A.), 2007 FCA 24, [2007] 4 F.C.R. 332, 276 D.L.R. (4th) 369, 360 N.R. 183, 59 Imm. L.R. (3d) 122, [2007] F.C.J. No. 139 (QL), 2007 CarswellNat 212, setting aside a decision of Lutfy C.J., 2005 FC 1218, 266 F.T.R. 138, 48 Imm. L.R. (3d) 253, [2005] F.C.J. No. 1465 (QL), 2005 CarswellNat 2651, dismissing the application for judicial review of the decision of the Immigration Appeal Division, [2004] I.A.D.D. No. 1268 (QL).  Appeal allowed, Fish J. dissenting.

Urszula Kaczmarczyk and Cheryl D. Mitchell, for the appellant.

Garth Barriere and Daniel B. Geller, for the respondent.

Joseph J. Arvay, Q.C., and Joel M. Rubinoff, for the intervener.

The judgment of McLachlin C.J. and Binnie, LeBel, Abella and Charron JJ. was delivered by

[1]     Binnie J. — At issue in this appeal is the extent to which, if at all, the exercise by judges of statutory powers of judicial review (such as those established by ss. 18  and 18.1  of the Federal Courts Act , R.S.C. 1985, c. F-7 ) is governed by the common law principles lately analysed by our Court in Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190.

[2]     The respondent, Khosa, applied unsuccessfully to the Immigration Appeal Division (“IAD”) of the Immigration and Refugee Board to remain in Canada, notwithstanding his conviction for criminal negligence causing death in an automobile street race.  A valid removal order had been issued to return him to India.  The majority of the IAD did not accept that there were “sufficient humanitarian and compassionate considerations [to] warrant special relief [against the removal order] in light of all the circumstances of the case” within the meaning of s. 67(1) (c) of the Immigration and Refugee Protection Act , S.C. 2001, c. 27  (“IRPA ”).  Applying the “patent unreasonableness” standard of review, the judicial review judge at first instance dismissed Khosa’s challenge to the IAD decision.  However, applying a “reasonableness” simpliciter standard of review, a majority of the Federal Court of Appeal set aside the IAD decision.  Dunsmuir (decided subsequently to both lower court decisions) did away with the distinction between “patent unreasonableness” and “reasonableness simpliciter” and substituted a more context-driven view of “reasonableness” that nevertheless “does not pave the way for a more intrusive review by courts” (para. 48).

[3]     The appellant Minister sought leave to appeal to this Court to argue that in any event s. 18.1  of the Federal Courts Act  establishes a legislated standard of review that displaces the common law altogether.  On this view, Dunsmuir is largely irrelevant to the current appeal.  However, it is apparent that while the courts below differed on the choice of the appropriate common law standard of review, neither the judge at first instance nor any of the judges of the appellate court considered the common law of judicial review to be displaced by s. 18.1  of the Federal Courts Act .  The trial court took the view that s. 18.1  of the Federal Courts Act  deals essentially with grounds of review of administrative action, not standards of review, and the Federal Court of Appeal proceeded in the same way.  I think this approach is correct although, as will be discussed, s. 18.1(4)(d) does provide legislative guidance as to “the degree of deference” owed to the IAD’s findings of fact.

[4]     Dunsmuir teaches that judicial review should be less concerned with the formulation of different standards of review and more focussed on substance, particularly on the nature of the issue that was before the administrative tribunal under review.  Here, the decision of the IAD required the application of broad policy considerations to the facts as found to be relevant, and weighed for importance, by the IAD itself.  The question whether Khosa had shown “sufficient humanitarian and compassionate considerations” to warrant relief from his removal order, which all parties acknowledged to be valid, was a decision which Parliament confided to the IAD, not to the courts.  I conclude that on general principles of administrative law, including our Court’s recent decision in Dunsmuir, the applications judge was right to give a higher degree of deference to the IAD decision than seemed appropriate to the Federal Court of Appeal majority.  In my view, the majority decision of the IAD was within a range of reasonable outcomes and the majority of the Federal Court of Appeal erred in intervening in this case to quash it.  The appeal is therefore allowed and the decision of the Immigration Appeal Division is restored.

I.     Facts

[5]     The respondent, Sukhvir Singh Khosa, is a citizen of India.  He immigrated to Canada with his family in 1996, at the age of 14.  He has landed immigrant status.  During the evening of November 13, 2000, he and an individual named Bahadur Singh Bhalru, drove their respective cars at over 100 kilometres per hour along Marine Drive through a residential and commercial area of Vancouver.  At their criminal trial, the court concluded that they were “street racing”.  Khosa was prepared to plead guilty to a charge of dangerous driving, but not to the more serious charge of criminal negligence causing death, of which he was eventually convicted.  The respondent continued to deny street racing, although he admitted that he was speeding and that his driving behaviour was exceptionally dangerous. On appeal from sentencing, the British Columbia Court of Appeal commented:

. . . it is significant that the respondents were racing.  They were driving at excessive speeds in competition with each other on a major street lined with both commercial and residential properties.  They did this at a time when other vehicles and pedestrians reasonably could be expected to be on the roads.

                                                                   . . .

The “spontaneous” nature of the race . . . mitigates the severity with which it should be assessed.  The race was not planned, did not involve vehicles specifically modified for the purpose of racing, and was of relatively short duration.  As unacceptable as the conduct of the respondents was, it represented a reckless error in judgment more than a deliberate endangerment of the public.

(2003 BCCA 645, 190 B.C.A.C. 42, at paras. 33 and 36)

As to the “moral culpability” of the respondent and his co-accused, the Court of Appeal continued:

The Crown concedes that there are several factors which mitigate the moral culpability of the respondents in this case.  Mr. Khosa and Mr. Bhalru are both young, have no prior criminal record or driving offences, have expressed remorse for the consequences of their conduct, and have favourable prospects for rehabilitation. . . . [para. 38]

[6]     The respondent received a conditional sentence of two years less a day.  The conditions included house arrest, a driving ban, and community service, all of which were complied with prior to the IAD hearing.

II.     Judicial History

A.     Immigration Appeal Division, [2004] I.A.D.D. No. 1268 (QL)

(1)   The Majority

[7]     The majority of the IAD recognized (at para. 12) that its discretionary jurisdiction to grant “special relief” on humanitarian and compassionate grounds under s. 67(1) (c) of the IRPA  should be exercised in light of the factors adopted in Ribic v. Canada (Minister of Employment and Immigration), [1985] I.A.B.D. No. 4 (QL), and endorsed by this Court in Chieu v. Canada (Minister of Citizenship and Immigration), 2002 SCC 3, [2002] 1 S.C.R. 84, at paras. 40, 41 and 90, namely:

(1)               the seriousness of the offence leading to the removal order;

(2)               the possibility of rehabilitation;

(3)               the length of time spent, and the degree to which the individual facing removal is established, in Canada;

(4)               the family and community support available to the individual facing removal;

(5)               the family in Canada and the dislocation to the family that removal would cause; and

(6)               the degree of hardship that would be caused to the individual facing removal to his country of nationality.

[8]     The majority considered that the last four Ribic factors were not particularly compelling for or against relief.  As to the first two factors, the offence in question was “extremely serious” (para. 14) and the majority expressed particular concern over Khosa’s refusal to accept without reservation the finding that he had been street racing.  The IAD majority considered that this refusal “reflects a lack of insight into his conduct” (para. 15).  As to Khosa’s prospects for rehabilitation, the majority decided that there was insufficient evidence upon which to make a finding one way or the other (paras. 15 and 23).  However, even if Khosa had good prospects for rehabilitation, “balancing all the relevant factors, . . . the scale does not tip in [Khosa’s] favour” (para. 23).  Accordingly, “special relief” was denied.

(2)   The Dissent

[9]     The dissenting member of the IAD would also have denied the appeal, but she would have stayed the execution of the deportation order pending a further review in three years.  She acknowledged the seriousness of the offence for which Khosa was convicted but found that it was mitigated by matters not considered important by the majority.  Evidence of remorse and rehabilitation favoured relief.  Having regard to the criminal proceedings, she noted that no penitentiary term was considered appropriate.  The crime of which Khosa was convicted is not one of intent.  There was no evidence of criminal propensity.  The race was spontaneous and short.  All sentencing conditions had been complied with.  In the circumstances, she concluded that relief ought to be granted on humanitarian and compassionate grounds.

B.    Federal Court, 2005 FC 1218, 266 F.T.R. 138

[10] Lutfy C.J. dismissed Khosa’s application for judicial review. He found that considerable deference was required, given the broad nature of the discretion vested in the IAD and its expertise in applying Ribic factors in appeals under s. 67(1) (c) of the IRPA .  The appropriate standard of review is patent unreasonableness.  Whether or not the IAD majority erred in its appreciation of the evidence in light of the Ribic factors is “substantially, if not completely, factual” (para. 29).

[11] Lutfy C.J. said that the crux of Khosa’s argument was that the majority of the IAD erred by placing inordinate emphasis on his denial that his admittedly dangerous driving took place in the context of a street race, but the judge declined to reweigh the evidence, saying (at para. 36):

In assessing Mr. Khosa’s expression of remorse, they [the majority] chose to place greater weight on his denial that he participated in a “race” than others might have.  The IAD conclusion on the issue of remorse appears to differ from that of the criminal courts.  The IAD, however, unlike the criminal courts, had the opportunity to assess Mr. Khosa’s testimony.

[12] In the result, Lutfy C.J. held that there was no basis for concluding that “the majority opinion is patently unreasonable or, in the words of paragraph 18.1(4) (d) of the Federal Courts Act , one which was based on an erroneous finding of fact ‘made in a perverse or capricious manner or without regard for the material’” (para. 39).

C.       Federal Court of Appeal, 2007 FCA 24, [2007] 4 F.C.R. 332

(1)   The Majority

[13] Décary J.A. (Malone J.A. concurring) disagreed with Lutfy C.J. on the appropriate standard of review.  In his view, the applicable standard was “reasonableness”.  Accordingly, “[s]ince the applications Judge applied the wrong standard of review, it is my duty, on appeal, to review the Board’s decision on the correct standard of review, that is, on the standard of reasonableness” (para. 14).

[14] With respect to the second Ribic factor, Décary J.A. said that the “possibility of rehabilitation” is a criminal law concept with which the IAD does not have particular expertise.  It should be wary of questioning findings of the criminal courts on matters falling squarely within their expertise.  The majority “merely acknowledges the findings of the British Columbia courts in that regard, which are favourable to [Khosa], and does not explain why it comes to the contrary conclusion . . . .  The whole of the evidence with respect to the conduct of [Khosa] after his sentencing undisputedly strengthens the findings of the criminal courts.  Yet, the Board ignores that evidence and those findings” (para. 17).  As to the “street racing” issue, Décary J.A. said:

It clearly appears from the transcripts of the hearing that the presiding member — who wrote the majority decision — and counsel for the Crown, had some kind of fixation with the fact that the offence was related to street racing, to such a point that the hearing, time and time again, was transformed into a quasi-criminal trial, if not into a new criminal trial. [para. 18]

For these reasons, Décary J.A. concluded that the majority had acted unreasonably.

(2)   The Dissent

[15] Desjardins J.A. concluded that the applications judge was right to apply the “patent unreasonableness” standard.  She emphasized that the IAD has expertise in applying the Ribic factors in decisions under s. 67(1)(c) of the IRPA and that this exercise is “highly fact-based and contextual” (para. 36).  Desjardins J.A. also emphasized the broad discretion conferred upon the IAD by s. 67(1) (c) of the IRPA .  In her view, Lutfy C.J. had made no reviewable error.  She would have dismissed the appeal.

III.     Relevant Statutory Provisions

[16]                           Immigration and Refugee Protection Act , S.C. 2001, c. 27 

3. (1)   The objectives of this Act with respect to immigration are

                                                                   . . .

(h)    to protect the health and safety of Canadians and to maintain the security of Canadian society;

36. (1) A permanent resident or a foreign national is inadmissible on grounds of serious criminality for

(a)    having been convicted in Canada of an offence under an Act of Parliament punishable by a maximum term of imprisonment of at least 10 years, or of an offence under an Act of Parliament for which a term of imprisonment of more than six months has been imposed;

67. (1) To allow an appeal, the Immigration Appeal Division must be satisfied that, at the time that the appeal is disposed of,

                                                                   . . .

(c)    other than in the case of an appeal by the Minister, taking into account the best interests of a child directly affected by the decision, sufficient humanitarian and compassionate considerations warrant special relief in light of all the circumstances of the case.

Federal Courts Act , R.S.C. 1985, c. F-7 

18.1 (1) An application for judicial review may be made by the Attorney General of Canada or by anyone directly affected by the matter in respect of which relief is sought.

(2)    An application for judicial review in respect of a decision or an order of a federal board, commission or other tribunal shall be made within 30 days after the time the decision or order was first communicated by the federal board, commission or other tribunal to the office of the Deputy Attorney General of Canada or to the party directly affected by it, or within any further time that a judge of the Federal Court may fix or allow before or after the end of those 30 days.

(3)    On an application for judicial review, the Federal Court may

(a)    order a federal board, commission or other tribunal to do any act or thing it has unlawfully failed or refused to do or has unreasonably delayed in doing; or

(b)    declare invalid or unlawful, or quash, set aside or set aside and refer back for determination in accordance with such directions as it considers to be appropriate, prohibit or restrain, a decision, order, act or proceeding of a federal board, commission or other tribunal.

(4)    The Federal Court may grant relief under subsection (3) if it is satisfied that the federal board, commission or other tribunal

(a)    acted without jurisdiction, acted beyond its jurisdiction or refused to exercise its jurisdiction;

(b)    failed to observe a principle of natural justice, procedural fairness or other procedure that it was required by law to observe;

(c)    erred in law in making a decision or an order, whether or not the  error appears on the face of the record;

(d)    based its decision or order on an erroneous finding of fact that it made in a perverse or capricious manner or without regard for the material before it;

(e)    acted, or failed to act, by reason of fraud or perjured evidence; or

(f)    acted in any other way that was contrary to law.

(5)    If the sole ground for relief established on an application for judicial review is a defect in form or a technical irregularity, the Federal Court may

(a)    refuse the relief if it finds that no substantial wrong or miscarriage of justice has occurred; and

(b)    in the case of a defect in form or a technical irregularity in a decision or an order, make an order validating the decision or order, to have effect from any time and on any terms that it considers appropriate.

IV.     Analysis

[17] This appeal provides a good illustration of why the adjustment made by Dunsmuir was timely.  By switching the standard of review from patent unreasonableness to reasonableness simpliciter, the Federal Court of Appeal majority felt empowered to retry the case in important respects, even though the issues to be resolved had to do with immigration policy, not law.  Clearly, the majority felt that the IAD disposition was unjust to Khosa.  However, Parliament saw fit to confide that particular decision to the IAD, not to the judges.

[18] In cases where the legislature has enacted judicial review legislation, an analysis of that legislation is the first order of business.  Our Court had earlier affirmed that, within constitutional limits, Parliament may by legislation specify a particular standard of review: see R. v. Owen, 2003 SCC 33, [2003] 1 S.C.R. 779.  Nevertheless, the intended scope of judicial review legislation is to be interpreted in accordance with the usual rule that the terms of a statute are to be read purposefully in light of its text, context and objectives. 

[19] Generally speaking, most if not all judicial review statutes are drafted against the background of the common law of judicial review.  Even the more comprehensive among them, such as the British Columbia Administrative Tribunals Act, S.B.C. 2004, c. 45, can only sensibly be interpreted in the common law context because, for example, it provides in s. 58(2)(a) that “a finding of fact or law or an exercise of discretion by the tribunal in respect of a matter over which it has exclusive jurisdiction under a privative clause must not be interfered with unless it is patently unreasonable”.  The expression “patently unreasonable” did not spring unassisted from the mind of the legislator.  It was obviously intended to be understood in the context of the common law jurisprudence, although a number of indicia of patent unreasonableness are given in s. 58(3).  Despite Dunsmuir, “patent unreasonableness” will live on in British Columbia, but the content of the expression, and the precise degree of deference it commands in the diverse circumstances of a large provincial administration, will necessarily continue to be calibrated according to general principles of administrative law.  That said, of course, the legislature in s. 58 was and is directing the B.C. courts to afford administrators a high degree of deference on issues of fact, and effect must be given to this clearly expressed legislative intention.

A.     A Difference of Perspective

[20] As Rand J. commented in Roncarelli v. Duplessis, [1959] S.C.R. 121, at p. 140, “there is always a perspective within which a statute is intended to operate”.  This applies to s. 18.1  of the Federal Courts Act  as much as it does to any other enactment.

[21] My colleague Justice Rothstein adopts the perspective that in the absence of a privative clause or statutory direction to the contrary, express or implied, judicial review under s. 18.1 is to proceed “as it does in the regular appellate context” (para. 117).  Rothstein J. writes:

On my reading, where Parliament intended a deferential standard of review in s. 18.1(4), it used clear and unambiguous language.  The necessary implication is that where Parliament did not provide for deferential review, it intended the reviewing court to apply a correctness standard as it does in the regular appellate context.  [Emphasis added.]

I do not agree that such an implication is either necessary or desirable.  My colleague states that “where a legal question can be extricated from a factual or policy inquiry, it is inappropriate to presume deference where Parliament has not indicated this via a privative clause” (para. 90), citing Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235, at paras. 8 and 13.  Housen, of course, was a regular appeal in a civil negligence case.

[22] On this view, the reviewing court applies a standard of review of correctness unless otherwise directed to proceed (expressly or by necessary implication) by the legislature.

[23] Rothstein J. writes, at para. 87, that the Court “depart[ed] from the conceptual origin of standard of review” in Pezim v. British Columbia (Superintendent of Brokers), [1994] 2 S.C.R. 557.  Pezim was a unanimous decision of the Court which deferred to the expertise of a specialized tribunal in the interpretation of provisions of the Securities Act, S.B.C. 1985, c. 83, despite the presence of a right of appeal and the absence of a privative clause.

[24] The conceptual underpinning of the law of judicial review was “further blurred”, my colleague writes, by Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R. 982, which treated the privative clause “simply as one of several factors in the calibration of deference (standard of review)” (para. 92).  In my colleague’s view, “[i]t is not for the court to impute tribunal expertise on legal questions, absent a privative clause and, in doing so, assume the role of the legislature to determine when deference is or is not owed” (para. 91).

[25] I do not share Rothstein J.’s view that absent statutory direction, explicit or by necessary implication, no deference is owed to administrative decision-makers in matters that relate to their special role, function and expertise.  Dunsmuir recognized that with or without a privative clause, a measure of deference has come to be accepted as appropriate where a particular decision had been allocated to an administrative decision-maker rather than to the courts.  This deference extended not only to facts and policy but to a tribunal’s interpretation of its constitutive statute and related enactments because “there might be multiple valid interpretations of a statutory provision or answers to a legal dispute and that courts ought not to interfere where the tribunal’s decision is rationally supported” (Dunsmuir, at para. 41).  A policy of deference “recognizes the reality that, in many instances, those working day to day in the implementation of frequently complex administrative schemes have or will develop a considerable degree of expertise or field sensitivity to the imperatives and nuances of the legislative regime” (Dunsmuir, at para. 49, quoting Professor David J. Mullan, “Establishing the Standard of Review: The Struggle for Complexity?” (2004), 17 C.J.A.L.P. 59, at p. 93).  Moreover, “[d]eference may also be warranted where an administrative tribunal has developed particular expertise in the application of a general common law or civil law rule in relation to a specific statutory context” (Dunsmuir, at para. 54).

[26] Dunsmuir stands against the idea that in the absence of express statutory language or necessary implication, a reviewing court is “to apply a correctness standard as it does in the regular appellate context” (Rothstein J., at para. 117).  Pezim has been cited and applied in numerous cases over the last 15 years.  Its teaching is reflected in Dunsmuir.  With  respect, I would reject my colleague’s effort to roll back the Dunsmuir clock to an era where some courts asserted a level of skill and knowledge in administrative matters which further experience showed they did not possess.

B.      Section 18.1  of the Federal Courts Act 

[27] Given the differing perspectives that Rothstein J. and I bring to judicial review, it is not surprising that we differ on the role and function of s. 18.1  of the Federal Courts Act .

[28] In my view, the interpretation of s. 18.1  of the Federal Courts Act  must be sufficiently elastic to apply to the decisions of hundreds of different “types” of administrators, from Cabinet members to entry-level fonctionnaires, who operate in different decision-making environments under different statutes with distinct grants of decision-making powers.  Some of these statutory grants have privative clauses; others do not.  Some provide for a statutory right of appeal to the courts; others do not.  It cannot have been Parliament’s intent to create by s. 18.1  of the Federal Courts Act  a single, rigid Procrustean standard of decontextualized review for all “federal board[s], commission[s] or other tribunal[s]”, an expression which is defined (in s. 2) to include generally all federal administrative decision-makers.  A flexible and contextual approach to s. 18.1  obviates the need for Parliament to set customized standards of review for each and every federal decision-maker.

[29] The Minister’s reliance on Owen is misplaced.  At issue in that case was the standard applicable to the highly specific task of judicial review of decisions of Review Boards set up under s. 672.38  of the Criminal Code , R.S.C. 1985, c. C-46 , to deal with individuals found not criminally responsible (“NCR”) on account of a mental disorder.  The mandate of these Boards is to determine the “least onerous and least restrictive” limits on the liberty of NCR individuals who remain a “significant threat to the safety of the public” (s. 672.54).  On a statutory appeal (s. 672.78), the Court of Appeal is authorized to set aside a Review Board order on a number of grounds, namely

(a) the decision is unreasonable or cannot be supported by the evidence; or,

(b) the decision is based on a wrong decision on a question of law (unless no substantial wrong or miscarriage of justice has occurred); or

(c) there was a miscarriage of justice.

[30] The Owen court held that where Parliament has shown a clear intent then, absent any constitutional challenge, that is the standard of review that is to be applied (para. 32).  This approach was affirmed in Dunsmuir where the majority said that “determining the applicable standard of review is accomplished by establishing legislative intent” (para. 30).

[31] However, in Owen itself, even in the context of a precisely targeted proceeding related to a named adjudicative board, the  standard of review was evaluated by reference to the common law of judicial review, as was made clear in the following paragraph:

The first branch of the test corresponds with what the courts call the standard of review of reasonableness simpliciter, i.e., the Court of Appeal should ask itself whether the Board’s risk assessment and disposition order was unreasonable in the sense of not being supported by reasons that can bear even a somewhat probing examination . . . . [para. 33]

And in the next paragraph:

Resort must therefore be taken to the jurisprudence governing judicial review on a standard of reasonableness simpliciter . . . . [para. 34]

See also Penetanguishene Mental Health Centre v. Ontario (Attorney General), 2004 SCC 20, [2004] 1 S.C.R. 498.

[32] In Pinet v. St. Thomas Psychiatric Hospital, 2004 SCC 21, [2004] 1 S.C.R. 528, the Court dealt with the second branch of s. 672.78(1)(b) (“error of law”) on ordinary administrative law principles (clearly applying a correctness standard, at para. 25).  As to the saving proviso (i.e., the decision may be set aside for an error of law unless “no substantial wrong or miscarriage of justice has occurred”), the Pinet court held that the party seeking to uphold the Review Board decision despite the error of law must “satisfy the appellate court that a Review Board, acting reasonably, and properly informed of the law, would necessarily have reached the same conclusion absent the legal error” (para. 28).  None of this is explicit in the statute, but the common law was necessarily called in aid to fill in interstices in the legislation.  See also Mazzei v. British Columbia (Director of Adult Forensic Psychiatric Services), 2006 SCC 7, [2006] 1 S.C.R. 326.

[33] Resort to the general law of judicial review is all the more essential in the case of a provision like s. 18.1  of the Federal Courts Act  which, unlike s. 672  of the Criminal Code , is not limited to particular issues before a particular adjudicative tribunal but covers the full galaxy of federal decision-makers.  Section 18.1 must retain the flexibility to deal with an immense variety of circumstances.

C.    Matter of Statutory Interpretation

[34] The genesis of the Federal Courts Act  lies in Parliament’s decision in 1971 to remove from the superior courts of the provinces the jurisdiction over prerogative writs, declarations, and injunctions against federal boards, commissions and other tribunals and to place that jurisdiction (slightly modified) in a new federal court.  As Donald S. Maxwell, Q.C., the then Deputy Minister of Justice, explained to the House of Commons Standing Committee on Justice and Legal Affairs:

Clause 18 is based on the philosophy that we want to remove the jurisdiction and prerogative matters from the Superior Courts of the provinces and place them in our own federal Superior Court.

                                                                   . . .

. . . Having got them there, we think they are not entirely satisfactory.  We feel that there should be improvements made on these remedies of certiorari and prohibition.  This is what we are endeavouring to do in Clause 28. 

(See Minutes of Proceedings and Evidence of the Committee, No. 26, 2nd Sess., 28th Parl., May 7, 1970, at pp. 25‑26.)

This transfer of jurisdiction was recognized and accepted in Pringle v. Fraser, [1972] S.C.R. 821; Howarth v. National Parole Board, [1976] 1 S.C.R. 453, at pp. 470-72, and Martineau v. Matsqui Institution Disciplinary Board, [1980] 1 S.C.R. 602, at p. 637, with the proviso that such transfer does not deprive the provincial superior courts of their jurisdiction to determine the constitutional validity and applicability of legislation: Attorney General of Canada v. Law Society of British Columbia, [1982] 2 S.C.R. 307.  Subsequent amendments to the Act in 1990 (when s. 18.1 was added) clarified and simplified its expression and implementation, but did not have the effect of excluding the common law.  R. Sullivan, Sullivan on the Construction of Statutes (5th ed. 2008), notes that “courts readily assume that reform legislation is meant to be assimilated into the existing body of common law” (p. 432; see also pp. 261-62).

[35] My colleague Rothstein J. writes that “to say (or imply) that a Dunsmuir standard of review analysis applies even where the legislature has articulated the applicable standard of review directly contradicts Owen” (para. 100).  This assumes the point in issue, namely whether as a matter of interpretation, Parliament has or has not articulated the applicable standard of review in s. 18.1. 

[36] In my view, the language of s. 18.1 generally sets out threshold grounds which permit but do not require the court to grant relief.  Whether or not the court should exercise its discretion in favour of the application will depend on the court’s appreciation of the respective roles of the courts and the administration as well as the “circumstances of each case”:  see Harelkin v. University of Regina, [1979] 2 S.C.R. 561, at p. 575.  Further, “[i]n one sense, whenever the court exercises its discretion to deny relief, balance of convenience considerations are involved” (D. J. M. Brown and J. M. Evans, Judicial Review of Administrative Action in Canada (loose-leaf),  at p. 3-99).  Of course, the discretion must be exercised judicially, but the general principles of judicial review dealt with in Dunsmuir provide elements of the appropriate judicial basis for its exercise.

[37] On this point, as well, my colleague Rothstein J. expresses disagreement.  He cites a number of decisions dealing with different applications of the Court’s discretion.  He draws from these cases the negative inference that other applications of the discretion are excluded from s. 18.1(4).  In my view, with respect, such a negative inference is not warranted.  Decisions that address unrelated problems are no substitute for a proper statutory analysis of s.  18.1(4) itself which in the English text provides that

18.1 . . .

(4)    The Federal Court may grant relief under subsection (3) if it is satisfied that the federal board, commission or other tribunal

. . .

[38] A different concern emerges from the equally authoritative French text of s. 18.1(4) which reads:

18.1 . . .

(4)    Les mesures prévues au paragraphe (3) sont prises si la Cour fédérale est convaincue que l’office fédéral, selon le cas :

. . .

Generally speaking, the use of the present indicative tense (sont prises) is not to be read as conferring a discretion: see s. 11 of the French version of the Interpretation Act , R.S.C. 1985, c. I-21 , and P.-A. Côté, The Interpretation of Legislation in Canada (3rd ed. 2000), p. 72, fn. 123 (in the French edition, the point is canvassed by Professor Côté, at p. 91, fn. 123).  It has been truly remarked in the context of bilingual legislation that “Canadians read only one version of the law at their peril”: M. Bastarache et al., The Law of Bilingual Interpretation (2008), at p. 32.  However, the text of s. 18.1(4) must be interpreted not only in accordance with the rules governing bilingual statutes but within the larger framework of the modern rule that the words of an Act are to be read in their entire context and in their grammatical and ordinary sense harmoniously with the scheme of the Act, the object of the Act, and the intention of Parliament.

[39] The English version of s. 18.1(4) is permissive; the court is clearly given discretion.  In the French version, the words “sont prises” translate literally as “are taken” which do not, on the face of it, confer a discretion.  A shared meaning on this point is difficult to discern.  Nevertheless, the linguistic difference must be reconciled as judges cannot be seen to be applying s. 18.1(4) differently across the country depending on which language version of s. 18.1(4) they happen to be reading.  In R. v. Daoust, 2004 SCC 6, [2004] 1 S.C.R. 217, at para. 26, the Court cited with approval the following approach:

Unless otherwise provided, differences between two official versions of the same enactment are reconciled by educing the meaning common to both.  Should this prove to be impossible, or if the common meaning seems incompatible with the intention of the legislature as indicated by the ordinary rules of interpretation, the meaning arrived at by the ordinary rules should be retained.

(Quoting Côté, at p. 324.)

(See also Bastarache et al., at p. 32.) Linguistic analysis of the text is the servant, not the master, in the task of ascertaining Parliamentary intention: see Slaight Communications Inc. v. Davidson, [1989] 1 S.C.R. 1038, at pp. 1071-72 (Lamer J. dissenting in part, but not on this point).  A blinkered focus on the textual variations might lead to an interpretation at odds with the modern rule because, standing alone, linguistic considerations ought not to elevate an argument about text above the relevant context, purpose and objectives of the legislative scheme: see Sullivan, at p. 116.

[40] Here the English version cannot be read so as to compel the court to grant relief: the word “may” is unquestionably permissive.  In Bastarache et al., it is said that “the clearer version provides the common meaning” (p. 67), but it cannot be said that the French text here is ambiguous.  Accordingly, the linguistic issue must be placed in the framework of the modern rules of statutory interpretation that give effect not only to the text but to context and purpose.  There is nothing in the context or purpose of the enactment to suggest a Parliamentary intent to eliminate the long-standing existence of a discretion in judicial review remedies.  As mentioned earlier, the principal legislative objective was simply to capture the judicial review of federal decision-makers for the Federal Court.  Under the general public law of Canada (then as now), the granting of declarations and the original prerogative and extraordinary remedies, and subsequent statutory variations thereof, have generally been considered to be discretionary, as discussed by Beetz J. in Harelkin.  The Federal Court’s discretion in matters of judicial review has repeatedly been affirmed by this Court:  see Solosky v. The Queen, [1980] 1 S.C.R. 821, at pp. 830-31; Canada (Auditor General) v. Canada (Minister of Energy, Mines and Resources), [1989] 2 S.C.R. 49, at pp. 92-93, and Friends of the Oldman River Society v. Canada (Minister of Transport), [1992] 1 S.C.R. 3, at pp. 77-80.  The Federal Courts themselves have repeatedly asserted, notwithstanding the problem posed by the French text of s. 18.1(4), the existence of a discretion in the exercise of their judicial review jurisdiction (and quite properly so in my opinion) both in decisions rendered in French (see, e.g., Canada v. Grenier, 2005 CAF 348, [2006] 2 F.C.R. 287, per Létourneau J.A., at para. 40, and Devinat v. Canada (Commission de l’immigration et du status de réfugié), [2000] 2 F.C. 212 (C.A.), per The Court, at para. 73) and in English (see, e.g., Thanabalasingham v. Canada (Minister of Citizenship and Immigration), 2006 FCA 14, 263 D.L.R. (4th) 51, per Evans J.A., at para. 9; Charette v. Canada (Commissioner of Competition), 2003 FCA 426, 29 C.P.R. (4th) 1, per Sexton J.A., at para. 61, and Pal v. Canada (Minister of Employment and Immigration) (1993), 24 Admin. L.R. (2d) 68, per Reed J., at para. 9).  I  conclude that notwithstanding the bilingual issue in the text, s. 18.1(4) should be interpreted so as to preserve to the Federal Court a discretion to grant or withhold relief, a discretion which, of course, must be exercised judicially and in accordance with proper principles.  In my view, those principles include those set out in Dunsmuir.

[41] With these general observations I turn to the particular paragraphs of s. 18.1(4)  of the Federal Courts Act  that, in my view, enable but do not require judicial intervention.

[42] Section 18.1(4)(a) provides for relief where a federal board, commission or other tribunal

(a) acted without jurisdiction, acted beyond its jurisdiction or refused to exercise its jurisdiction;

No standard of review is specified.  Dunsmuir says that jurisdictional issues command a correctness standard (majority, at para. 59).   The Federal Courts Act  does not indicate in what circumstances, despite jurisdictional error having been demonstrated, relief may properly be withheld.  For that and other issues, resort will have to be had to the common law.  See Harelkin, at pp. 575-76.

[43] Judicial intervention is also authorized where a federal board, commission or other tribunal

(b) failed to observe a principle of natural justice, procedural fairness or other procedure that it was required by law to observe;

No standard of review is specified.  On the other hand, Dunsmuir says that procedural issues (subject to competent legislative override) are to be determined by a court on the basis of a correctness standard of review.  Relief in such cases is governed by common law principles, including the withholding of relief when the procedural error is purely technical and occasions no substantial wrong or miscarriage of justice (Pal, at para. 9).  This is confirmed by s. 18.1(5).  It may have been thought that the Federal Court, being a statutory court, required a specific grant of power to “make an order validating the decision” (s. 18.1(5)) where appropriate.

[44] Judicial intervention is authorized where a federal board, commission or other tribunal

(c) erred in law in making a decision or an order, whether or not the error appears on the face of the record;

Errors of law are generally governed by a correctness standard.  Mugesera v. Canada (Minister of Citizenship and Immigration), 2005 SCC 40, [2005] 2 S.C.R. 100, at para. 37, for example, held that the general questions of international law and criminal law at issue in that case had to be decided on a standard of correctness.  Dunsmuir (at para. 54), says that if the interpretation of the home statute or a closely related statute by an expert decision-maker is reasonable, there is no error of law justifying intervention.  Accordingly, para. (c) provides a ground of intervention, but the common law will stay the hand of the judge(s) in certain cases if the interpretation is by an expert adjudicator interpreting his or her home statute or a closely related statute.  This nuance does not appear on the face of para. (c), but it is the common law principle on which the discretion provided in s. 18.1(4) is to be exercised.  Once again, the open textured language of the Federal Courts Act  is supplemented by the common law.

[45] Judicial intervention is further authorized where a federal board, commission or other tribunal

(d) based its decision or order on an erroneous finding of fact that it made in a perverse or capricious manner or without regard for the material before it;

The legislator would have been aware of the great importance attached by some judicial decisions to so-called “jurisdictional fact finding”; see, e.g., Bell v. Ontario Human Rights Commission, [1971] S.C.R. 756, and Canadian Union of Public Employees, Local 963 v. New Brunswick Liquor Corp., [1979] 2 S.C.R. 227.  Parliament clearly wished to put an end to the tendency of some courts to seize on a “preliminary fact” on which the administrative agency’s decision was said to be based to quash a decision.  In Bell, the “jurisdictional fact” was whether the residential accommodation in respect of which a prospective tenant claimed rental discrimination was a “self-contained dwelling unit”.  The Court disagreed with the Human Rights Commission, which had “based” its decision on this threshold fact.  Viewed in this light, s. 18.1(4)(d) was intended to confirm by legislation what  Dickson J. had said in New Brunswick Liquor Corp., namely that judges should  “not be alert to brand as jurisdictional, and therefore subject to broader curial review, that which may be doubtfully so” (p. 233).

[46] More generally, it is clear from s. 18.1(4)(d) that Parliament intended administrative fact finding to command a high degree of deference.  This is quite consistent with Dunsmuir.  It provides legislative precision to the reasonableness standard of review of  factual issues in cases falling under the Federal Courts Act .

[47] Paragraph (e) contemplates a question of mixed fact and law namely that the federal board, commission or other tribunal

(e) acted, or failed to act, by reason of fraud or perjured evidence;

The common law would not allow a statutory decision-maker to rely on fraudulent or perjured testimony.  The court would be expected to exercise its discretion in favour of the applicant under para. (e) as well.

[48] Section 18.1(4)(f) permits judicial intervention if the federal board, commission or other tribunal

(f) acted in any other way that was contrary to law.

A reference to “contrary to law” necessarily includes “law” outside the Federal Courts Act  including general principles of administrative law. Paragraph (f) shows, if further demonstration were necessary, that s. 18.1(4) is not intended to operate as a self-contained code, but is intended by Parliament to be interpreted and applied against the backdrop of the common law, including those elements most recently expounded in Dunsmuir.

[49] In Federal Courts Practice 2009 (2008), B. J. Saunders et al. state, at pp. 112-13:

Grounds for Review

Section 18.1(4) sets out the grounds which an applicant must establish to succeed on an application for judicial review.  The grounds are broadly stated and reflect, generally, the grounds upon which judicial review could be obtained under the prerogative and extraordinary remedies listed in section 18(1). 

Section 18.1(4)(f) ensures that the Court will not be hindered in developing new grounds for review. [Emphasis added.]

[50] I readily accept, of course, that the legislature can by clear and explicit language oust the common law in this as in other matters.  Many provinces and territories have enacted judicial review legislation which not only provide guidance to the courts but have the added benefit of making the law more understandable and accessible to interested members of the public.  The diversity of such laws makes generalization difficult.  In some jurisdictions (as in British Columbia), the legislature has moved closer to a form of codification than has Parliament in the Federal Courts Act .  Most jurisdictions in Canada seem to favour a legislative approach that explicitly identifies the grounds for review but not the standard of review.¹  In other provinces, some laws specify “patent unreasonableness”.²  In few of these statutes, however, is the content of the specified standard of review defined, leading to the inference that the legislatures left the content to be supplied by the common law.

[51] As stated at the outset, a legislature has the power to specify a standard of review, as held in Owen, if it manifests a clear intention to do so.  However, where the legislative language permits, the courts (a) will not interpret grounds of review as standards of review, (b) will apply Dunsmuir principles to determine the appropriate approach to judicial review in a particular situation, and (c) will presume the existence of a discretion to grant or withhold relief based on the Dunsmuir teaching of restraint in judicial intervention in administrative matters (as well as other factors such as an applicant’s delay, failure to exhaust adequate alternate remedies, mootness, prematurity, bad faith and so forth).

D.  Standard of Review Analysis

[52] Dunsmuir states that “[c]ourts, while exercising their constitutional functions of judicial review, must be sensitive not only to the need to uphold the rule of law, but also to the necessity of avoiding undue interference with the discharge of administrative functions in respect of the matters delegated to administrative bodies by Parliament and legislatures” (para. 27).

[53] The process of judicial review involves two steps.  First, Dunsmuir says that “[a]n exhaustive review is not required in every case to determine the proper standard of review” (para. 57).  As between correctness and reasonableness, the “existing jurisprudence may be helpful” (para. 57).  And so it is in this case.  Dunsmuir renders moot the dispute in the lower courts between patent unreasonableness and reasonableness.  No authority was cited to us that suggests a “correctness” standard of review is appropriate for IAD decisions under s. 67(1) (c) of the IRPA .  Accordingly, “existing jurisprudence” points to adoption of a “reasonableness” standard.

[54] This conclusion is reinforced by the second step of the analysis when jurisprudential categories are not conclusive.  Factors then to be considered include: (1) the presence or absence of a privative clause; (2) the purpose of the IAD as determined by its enabling legislation; (3) the nature of the question at issue before the IAD; and (4) the expertise of the IAD in dealing with immigration policy (Dunsmuir, at para. 64).  Those factors have to be considered as a whole, bearing in mind that not all factors will necessarily be relevant for every single case.  A contextualized approach is required.  Factors should not be taken as items on a check list of criteria that need to be individually analysed, categorized and balanced in each case to determine whether deference is appropriate or not.  What is required is an overall evaluation.  Nevertheless, having regard to the argument made before us, I propose to comment on the different factors identified in Dunsmuir, all of which in my view point to a reasonableness standard.

[55] As to the presence of a privative clause, s. 162(1)  of the IRPA  provides that “[e]ach Division of the Board has, in respect of proceedings brought before it under this Act, sole and exclusive jurisdiction to hear and determine all questions of law and fact, including questions of jurisdiction”.  A privative clause is an important indicator of legislative intent.  While privative clauses deter judicial intervention, a statutory right of appeal may be at ease with it, depending on its terms.  Here, there is no statutory right of appeal.

[56] As to the purpose of the IAD as determined by its enabling legislation, the IAD determines a wide range of appeals under the IRPA , including appeals from permanent residents or protected persons of their deportation orders, appeals from persons seeking to sponsor members of the family class, and appeals by permanent residents against decisions made outside of Canada on their residency obligations, as well as appeals by the Minister against decisions of the Immigration Division taken at admissibility hearings (s. 63).  A decision of the IAD is reviewable only if the Federal Court grants leave to commence judicial review (s. 72).

[57] In recognition that hardship may come from removal, Parliament has provided in s. 67(1)(c) a power to grant exceptional relief.  The nature of the question posed by s. 67(1)(c) requires the IAD to be “satisfied that, at the time that the appeal is disposed of . . . sufficient humanitarian and compassionate considerations warrant special relief”.  Not only is it left to the IAD to determine what constitute “humanitarian and compassionate considerations”, but the “sufficiency” of such considerations in a particular case as well.  Section 67(1)(c) calls for a fact‑dependent and policy‑driven assessment by the IAD itself.  As noted in Prata v. Minister of Manpower and Immigration, [1976] 1 S.C.R. 376, at p. 380, a removal order

establishes that, in the absence of some special privilege existing, [an individual subject to a lawful removal order] has no right whatever to remain in Canada.  [An individual appealing a lawful removal order] does not, therefore, attempt to assert a right, but, rather, attempts to obtain a discretionary privilege. [Emphasis added.]

[58] The respondent raised no issue of practice or procedure.  He accepted that the removal order had been validly made against him pursuant to s. 36(1)  of the IRPA .  His attack was simply a frontal challenge to the IAD’s refusal to grant him a “discretionary privilege”.  The IAD decision to withhold relief was based on an assessment of the facts of the file.  The IAD had the advantage of conducting the hearings and assessing the evidence presented, including the evidence of the respondent  himself.  IAD members have considerable expertise in determining appeals under the IRPA .  Those factors, considered altogether, clearly point to the application of a reasonableness standard of review.  There are no considerations that might lead to a different result.  Nor is there anything in s. 18.1(4) that would conflict with the adoption of a “reasonableness” standard of review in s. 67(1)(c) cases. I conclude, accordingly, that “reasonableness” is the appropriate standard of review.

E.      Applying the “Reasonableness” Standard

[59] Reasonableness is a single standard that takes its colour from the context.  One of the objectives of Dunsmuir was to liberate judicial review courts from what came to be seen as undue complexity and formalism.  Where the reasonableness standard applies, it requires deference.  Reviewing courts cannot substitute their own appreciation of the appropriate solution, but must rather determine if the outcome falls within “a range of possible, acceptable outcomes which are defensible in respect of the facts and law” (Dunsmuir, at para. 47).  There might be more than one reasonable outcome.  However, as long as the process and the outcome fit comfortably with the principles of justification, transparency and intelligibility, it is not open to a reviewing court to substitute its own view of a preferable outcome.

[60] In my view, having in mind the considerable deference owed to the IAD and the broad scope of discretion conferred by the IRPA , there was no basis for the Federal Court of Appeal to interfere with the IAD decision to refuse special relief in this case.

[61] My colleague Fish J. agrees that the standard of review is reasonableness, but he would allow the appeal.  He writes:

While Mr. Khosa’s denial of street racing may well evidence some “lack of insight” into his own conduct, it cannot reasonably be said to contradict — still less to outweigh, on a balance of probabilities — all of the evidence in his favour on the issues of remorse, rehabilitation and likelihood of reoffence.  [para. 149]

I do not believe that it is the function of the reviewing court to reweigh the evidence.

[62] It is apparent that Fish J. takes a different view than I do of the range of outcomes reasonably open to the IAD in the circumstances of this case.  My view is predicated on what I have already said about the role and function of the IAD as well as the fact that Khosa does not contest the validity of the removal order made against him.  He seeks exceptional and discretionary relief that is available only if the IAD itself is satisfied that “sufficient humanitarian and compassionate considerations warrant special relief”.  The IAD majority was not so satisfied.  Whether we agree with a particular IAD decision or not is beside the point.  The decision was entrusted by Parliament to the IAD, not to the judges.

[63] The Dunsmuir majority held:

A court conducting a review for reasonableness inquires into the qualities that make a decision reasonable, referring both to the process of articulating the reasons and to outcomes.  In judicial review, reasonableness is concerned mostly with the existence of justification, transparency and intelligibility within the decision-making process.  But it is also concerned with whether the decision falls within a range of possible, acceptable outcomes which are defensible in respect of the facts and law. [para. 47]

Dunsmuir thus reinforces in the context of adjudicative tribunals the importance of reasons, which constitute the primary form of accountability of the decision-maker to the applicant, to the public and to a reviewing court.  Although the Dunsmuir majority refers with approval to the proposition that an appropriate degree of deference “requires of the courts ‘not submission but a respectful attention to the reasons offered or which could be offered in support of a decision’” (para. 48 (emphasis added)), I do not think the reference to reasons which “could be offered” (but were not) should be taken as diluting the importance of giving proper reasons for an administrative decision, as stated in Baker v. Canada (Minister of Citizenship and Immigration), [1999] 2 S.C.R. 817, at para. 43.  Baker itself was concerned with an application on “humanitarian and compassionate grounds” for relief from a removal order. 

[64] In this case, both the majority and dissenting reasons of the IAD disclose with clarity the considerations in support of both points of view, and the reasons for the disagreement as to outcome.  At the factual level, the IAD divided in large part over differing interpretations of Khosa’s expression of remorse, as was pointed out by Lutfy C.J.  According to the IAD majority:

It is troublesome to the panel that [Khosa] continues to deny that his participation in a “street-race” led to the disastrous consequences. . . . At the same time, I am mindful of [Khosa’s] show of relative remorse at this hearing for his excessive speed in a public roadway and note the trial judge’s finding of this remorse . . . . This show of remorse is a positive factor going to the exercise of special relief.  However, I do not see it as a compelling feature of the case in light of the limited nature of [Khosa’s] admissions at this hearing.  [Emphasis added; para. 15.]

According to the IAD dissent on the other hand:

. . . from early on he [Khosa] has accepted responsibility for his actions.  He was prepared to plead guilty to dangerous driving causing death . . . .

I find that [Khosa] is contrite and remorseful. [Khosa] at hearing was regretful, his voice tremulous and filled with emotion. . . .

                                                                   . . .

The majority of this panel have placed great significance on [Khosa’s] dispute that he was racing, when the criminal court found he was.  And while they concluded this was “not fatal” to his appeal, they also determined that his continued denial that he was racing “reflects a lack of insight.”  The panel concluded that this “is not to his credit.”  The panel found that [Khosa] was remorseful, but concluded it was not a “compelling feature in light of the limited nature of [Khosa’s] admissions”.

However I find [Khosa’s] remorse, even in light of his denial he was racing, is genuine and is evidence that [Khosa] will in future be more thoughtful and will avoid such recklessness. [paras. 50-51 and 53-54]

It seems evident that this is the sort of factual dispute which should be resolved by the IAD in the application of immigration policy, and not reweighed in the courts.

[65] In terms of transparent and intelligible reasons, the majority considered each of the Ribic factors.  It rightly observed that the factors are not exhaustive and that the weight to be attributed to them will vary from case to case (para. 12).  The majority reviewed the evidence and decided that, in the circumstances of this case, most of the factors did not militate strongly for or against relief.  Acknowledging the findings of the criminal courts on the seriousness of the offence and possibility of rehabilitation (the first and second of the Ribic factors), it found that the offence of which the respondent was convicted was serious and that the prospects of rehabilitation were difficult to assess (para. 23).

[66] The weight to be given to the respondent’s evidence of remorse and his prospects for rehabilitation depended on an assessment of his evidence in light of all the circumstances of the case.  The IAD has a mandate different from that of the criminal courts.  Khosa did not testify at his criminal trial, but he did before the IAD.  The issue before the IAD was not the potential for rehabilitation for purposes of sentencing, but rather whether the prospects for rehabilitation were such that, alone or in combination with other factors, they warranted special relief from a valid removal order.  The IAD was required to reach its own conclusions based on its own appreciation of the evidence.  It did so.

[67] As mentioned, the courts below recognized some merit in Khosa’s complaint.  Lutfy C.J. recognized that the majority “chose to place greater weight on his denial that he participated in a race than others might have” (para. 36).  Décary J.A. described the majority’s preoccupation with street racing as “some kind of fixation” (para. 18).  My colleague Fish J. also decries the weight put on this factor by the majority (para. 141).  However, as emphasized in Dunsmuir, “certain questions that come before administrative tribunals do not lend themselves to one specific, particular result.  Instead, they may give rise to a number of possible, reasonable conclusions.  Tribunals have a margin of appreciation within the range of acceptable and rational solutions” (para. 47).  In light of the deference properly owed to the IAD under s. 67(1) (c) of the IRPA ,  I cannot, with respect, agree with my colleague Fish J. that the decision reached by the majority in this case to deny special discretionary relief against a valid removal order fell outside the range of reasonable outcomes.

V.    Disposition

[68] The appeal is allowed and the decision of the IAD is restored.

The following are the reasons delivered by

[69] Rothstein J. — I have had the benefit of reading the reasons of my colleague Justice Binnie allowing this appeal.  While I concur with this outcome, I respectfully disagree with the majority’s approach to the application of the Dunsmuir standard of review analysis under s. 18.1  of the Federal Courts Act , R.S.C. 1985, c.  F-7  (“FCA ”).

I.     Introduction

[70] The central issue in this case is whether the FCA  expressly, or by necessary implication, provides the standards of review to be applied on judicial review, and if so, whether this displaces the common law standard of review analysis recently articulated in Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190.  The majority is of the view that the Dunsmuir standard of review analysis is to be read into s. 18.1(4)  of the FCA .  In my view, courts must give effect to the legislature’s words and cannot superimpose on them a duplicative common law analysis.  Where the legislature has expressly or impliedly provided for standards of review, courts must follow that legislative intent, subject to any constitutional challenge.

[71] Section 18.1(4)  of the FCA  states:

(4) [Grounds of review] The Federal Court may grant relief under subsection (3) if it is satisfied that the federal board, commission or other tribunal

(a)    acted without jurisdiction, acted beyond its jurisdiction or refused to exercise its jurisdiction;

(b)    failed to observe a principle of natural justice, procedural fairness or other procedure that it was required by law to observe;

(c)    erred in law in making a decision or an order, whether or not the error appears on the face of the record;

(d)    based its decision or order on an erroneous finding of fact that it made in a perverse or capricious manner or without regard for the material before it;

(e)    acted, or failed to act, by reason of fraud or perjured evidence; or

(f)    acted in any other way that was contrary to law.

[72] The language of s. 18.1(4)(d) makes clear that findings of fact are to be reviewed on a highly deferential standard.  Courts are only to interfere with a decision based on erroneous findings of fact where the federal board, commission or other tribunal’s factual finding was “made in a perverse or capricious manner or without regard for the material before it”.  By contrast with para. (d), there is no suggestion that courts should defer in reviewing a question that raises any of the other criteria in s. 18.1(4).  Where Parliament intended a deferential standard of review in s. 18.1(4), it used clear and unambiguous language.  The necessary implication is that where Parliament did not provide for a deferential standard, its intent was that no deference be shown.  As I will explain, the language and context of s. 18.1(4), and in particular the absence of deferential wording, demonstrates that a correctness standard is to be applied to questions of jurisdiction, natural justice, law and fraud.  The  language of s. 18.1(4)(d) indicates that deference is only to be applied to questions of fact.

[73] Dunsmuir reaffirmed that “determining the applicable standard of review is accomplished by establishing legislative intent” (para. 30).  The present majority’s insistence that Dunsmuir applies even where Parliament specifies a standard of review is inconsistent with that search for legislative intent, in my respectful view. 

[74] Standard of review developed as a means to reconcile the tension that privative clauses create between the rule of law and legislative supremacy:  see U.E.S., Local 298 v. Bibeault, [1988] 2 S.C.R. 1048.  “Full” or “strong” privative clauses that purport to preclude the judicial review of a question brought before a reviewing court give rise to this judicial-legislative tension, which deference and standard of review were developed to resolve: see Pasiechnyk v. Saskatchewan (Workers’ Compensation Board), [1997] 2 S.C.R. 890, at para. 17, for a discussion of the nature of privative clauses.  In my opinion, the application of Dunsmuir outside the strong privative clause context marks a departure from the conceptual and jurisprudential origins of the standard of review analysis.

[75] In my view, the question of whether the Dunsmuir standard of review analysis applies to judicial review under s. 18.1  of the FCA  becomes clear when one examines the conceptual basis for the common law standard of review analysis.  As explained in part II, standard of review emerged as a means to reconcile the judicial-legislative tension to which privative clauses gave rise.  The legislature’s desire to immunize certain administrative decisions from judicial scrutiny conflicted with the constitutional supervisory role of the courts and, as such, required a juridical response that could reconcile these competing requirements.  Deference and standard of review was the result.  It was the departure from this conceptual origin that blurred the role of the privative clause as the legislature’s communicative signal of relative expertise, and in doing so, the Court moved away from the search for legislative intent that governs this area.  In part III, I refer to this Court’s jurisprudence on the judicial recognition of legislated standards of review.  That jurisprudence is clear that courts must give effect to legislated standards of review, subject to any constitutional challenges.  In part IV, I explain that having regard to the conceptual origin of standard of review and the jurisprudence on legislated standards of review, s. 18.1(4)  of the FCA  occupies the field of standard of review and therefore ousts the common law on that question, excepting in cases of a strong privative clause.   In part V, I conclude by briefly considering the Immigration Appeal Division (“IAD”) decision in this case.  Like the majority, I would allow the appeal.

II.     The Place of Standard of Review: Reconciling the Judicial-Legislative Tension of the Privative Clause

A.     The Judicial-Legislative Tension

[76] Absent a privative clause, courts have always retained a supervisory judicial review role.  In the provinces, provincial superior courts have inherent jurisdiction and in most, if not all, cases statutory judicial review jurisdiction.  In the federal context, the FCA  transferred this inherent jurisdiction from the provincial superior courts to the Federal Courts.  Where applicable, statutory rights of appeal also grant affected parties the right to appeal an administrative decision to court.  This residual judicial review jurisdiction means that courts retained authority to ensure the rule of law even as delegated administrative decision making emerged.  La Forest and Iacobucci JJ. acknowledged this in Canada (Attorney General) v. Mossop, [1993] 1 S.C.R. 554, when La Forest J. wrote that

[i]n the absence of other provisions indicating a disposition to limit judicial review, the normal supervisory role of the courts remains.  The administrative tribunal, of course, is authorized to make determinations on these questions, but they are not to be insulated from the general supervisory role of the courts. [p. 584]

The legislature was well aware that parties who perceived an administrative injustice would still have recourse to the courts. 

[77]               The question is, however, whether the creation of expert tribunals automatically meant that there was to be some limitation on the judicial review role of the courts, in particular on questions of law.  Where the legislature enacted strong privative clauses precluding review for legal error, there is no doubt that this was the legislative intent.  In my opinion, the same limit on judicial review cannot be inferred merely from the establishment of a tribunal when the legislature did not seek to immunize the tribunal’s decisions from judicial review.  In those cases, the creation of an administrative decision-maker did not by itself give rise to a tension with the supervisory role of the courts. 

[78] In contrast, the majority appears to understand the judicial review of administrative decisions as automatically engaging a judicial-legislative tension, which the standard of review analysis seeks to resolve.  In Dunsmuir, Bastarache and LeBel JJ., writing for the majority, described this as follows:

Judicial review seeks to address an underlying tension between the rule of law and the foundational democratic principle, which finds an expression in the initiatives of Parliament and legislatures to create various administrative bodies and endow them with broad powers. Courts, while exercising their constitutional functions of judicial review, must be sensitive not only to the need to uphold the rule of law, but also to the necessity of avoiding undue interference with the discharge of administrative functions in respect of the matters delegated to administrative bodies by Parliament and legislatures. [Emphasis added; para. 27.]

As I understand this reasoning, the legislature displaced (or attempted to displace) judicial decision making in some areas by creating administrative bodies.  From this viewpoint, the standard of review functions as a necessary balancing exercise between the courts’ constitutional exercise of judicial review and the legislative desire to delegate certain powers to administrative bodies.

[79] In my opinion, in the absence of a strong privative clause such as existed in Dunsmuir, there are important reasons to question whether this view is applicable.  Broadly speaking, it is true of course that the creation of expert administrative decision-makers evidenced a legislative intent to displace or bypass the courts as primary adjudicators in a number of areas.  As Professor W. A. Bogart notes, “[t]he core idea was that the legislature wanted to regulate some area but wished someone else, an administrative actor, to carry out the regulation for reasons of expertise, expediency, access, independence from the political process, and so forth” (“The Tools of the Administrative State and the Regulatory Mix”, in C. M. Flood and L. Sossin, eds., Administrative Law in Context (2008), 25, at p. 31).  It was only with the enactment of privative clauses, however, that the legislature evidenced an intent to oust, or at the very least restrict, the court’s review role.   

[80] The most obvious case was labour relations.  Labour relations boards were created during the First and Second World Wars, in part to stave off labour unrest: see R. J. Charney and T. E. F. Brady, Judicial Review in Labour Law (loose-leaf), at pp. 2-1 to 2-17.  In order to protect the boards from judicial intervention, the legislature enacted strong privative clauses.  Professor Audrey Macklin notes that “[f]rustrated with judicial hostility toward the objectives of labour relations legislation, the government not only established a parallel administrative regime of labour relations boards, but also enacted statutory provisions that purported to preclude entirely judicial review of the legality of administrative action”: “Standard of Review: The Pragmatic and Functional Test”, in Administrative Law in Context, 197, at p. 199.  While there are different types of privative clauses, the labour relations context gave rise to strong privative clauses that typically purported to preclude review not only of factual findings, but also legal and jurisdictional decisions of the tribunal: see Pasiechnyk, at para. 17 (discussing what constitutes a “full” or “true” privative clause).  

[81] In attempting to preclude judicial review, privative clauses gave rise to a tension between the two core pillars of the public law system: legislative supremacy and the judicial enforcement of law: 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. 500.  Strong privative clauses reflected the legislature’s intent to make administrative decisions final and thereby beyond the purview of judicial scrutiny.  This  conflicts with the rule of law principle of accountability, for which access to courts is necessary.  As Professor Mary Liston notes:

The risk to the accountability function of the rule of law was that these officials could behave as a law unto themselves because they would be the sole judges of the substantive validity of their own acts.  The institutional result of privative clauses was a system of competing and irreconcilable supremacies between the legislative and judicial branches of government.

(“Governments in Miniature: The Rule of Law in the Administrative State”, in Administrative Law in Context, 77, at p. 104)

Faced with these competing “supremacies”, courts were forced to develop a juridical approach that would reconcile, or at least alleviate, this tension.  In Canada, courts opted for the deference approach.

B.     The Origins of the Standard of Review Analysis: Resolving the Privative Clause Tension

[82] The deference approach emerged as a means of reconciling Parliament’s intent to immunize certain administrative decisions from review with the supervisory role of courts in a rule of law system.  This approach originated with Canadian Union of Public Employees, Local 963 v.  New Brunswick Liquor Corp., [1979] 2 S.C.R. 227 (“C.U.P.E.”).  In reviewing a labour tribunal decision, Dickson J., as he then was, wrote that the privative clause “constitutes a clear statutory direction on the part of the Legislature that public sector labour matters be promptly and finally decided by the [Public Service Labour Relations] Board” (p. 235).  The decision of the Board was protected so long as it was not “so patently unreasonable that the Board . . . did ‘something which takes the exercise of its powers outside the protection of the privative or preclusive clause’” (p. 237). 

[83] The deference approach sought to give effect to the legislature’s recognition that the administrative decision-maker had relative expertise on some or all questions.  The privative clause indicated the area of tribunal expertise that the legislature was satisfied warranted deference.  As Professor Dyzenhaus explains:

. . . CUPE involves more than concession.  Right at the outset of the development of the idea of deference, it was clear that there was a judicial cession of interpretative authority to the tribunal, within the scope of its expertise — the area of jurisdiction protected by the privative clause.  The cession was not total — the tribunal could not be patently unreasonable.  But it was significant because it required that judges defer to the administration’s interpretations of the law, except on jurisdictional, constitutional, or constitution like issues. [Emphasis added; p. 512.]

[84] It is clear in C.U.P.E. that the deferential approach was contingent upon and shaped by the relevant privative clause.  Interpretive authority was only ceded to tribunals in the area  “within the scope of its expertise — the area of jurisdiction protected by the privative clause”.  A strong privative clause that protected legal as well as factual and discretionary decisions meant that the legislature recognized the tribunal as having relative expertise with respect to all these questions.  Dickson J. emphasized that the legislature’s frequent use of privative clauses in the labour relations context was intimately connected to tribunal expertise.  He wrote that “[t]he rationale for protection of a labour board’s decisions within jurisdiction is straightforward and compelling.  The labour board is a specialized tribunal which administers a comprehensive statute regulating labour relations” (p. 235).  In other words, tribunal expertise was a compelling rationale for imposing a privative clause.  It was not, however, a free-standing basis for deference. 

[85] A further step in the development of the deference approach was Bibeault, when this Court introduced the pragmatic and functional approach for determining the appropriate standard of review.  The pragmatic and functional approach, now known simply as the standard of review analysis, was intended to focus “the Court’s inquiry directly on the intent of the legislator rather than on interpretation of an isolated provision” (p. 1089).  In reviewing a decision-maker protected by a strong privative clause, this more expansive analysis examined “not only the wording of the enactment conferring jurisdiction on the administrative tribunal, but the purpose of the statute creating the tribunal, the reason for its existence, the area of expertise of its members and the nature of the problem before the tribunal” (p. 1088).  Beetz J. emphasized the overarching objective of giving effect to legislative intent while upholding courts’ supervisory role in a rule of law system (see p. 1090).

[86] The reasoning of Gonthier J. in Bell Canada v. Canada (Canadian Radio-Television and Telecommunications Commission),  [1989] 1 S.C.R. 1722, further reflected this understanding that it is the privative clause that signals when deference is owed and that demarcates the area of relative expertise.  Gonthier J. made clear that:

Where the legislator has clearly stated that the decision of an administrative tribunal is final and binding, courts of original jurisdiction cannot interfere with such decisions unless the tribunal has committed an error which goes to its jurisdiction. . . .  Decisions which are so protected are, in that sense, entitled to a non‑discretionary form of deference because the legislator intended them to be final and conclusive and, in turn, this intention arises out of the desire to leave the resolution of some issues in the hands of a specialized tribunal.  [Emphasis added; p. 1744.]

Gonthier J.’s statement captured the essential role of the privative clause. Privative clauses indicate the legislature’s intent that administrative decisions made within “the hands of a specialized tribunal” be deemed final and conclusive.  It is in these cases that courts must balance their constitutional role to preserve the rule of law with the legislature’s intent to oust the courts’ jurisdiction.  Gonthier J.’s reasoning understood expertise as the underlying rationale for enacting the privative clause.  Expertise alone was not interpreted as indicating a legislative intent for finality.  If the legislature intended to protect expert decision-makers from review, it did so through a privative clause.

C.      Departure From the Origins of Standard of Review: Expertise as a Stand-Alone Basis for Deference

[87] However, with Pezim v. British Columbia (Superintendent of Brokers), [1994] 2 S.C.R. 557, there was a departure from the conceptual origin of standard of review I have described.  That case involved the judicial review of a tribunal decision that was not protected by a privative clause and in fact was subject to a statutory right of appeal.  Relying on the language of “specialization of duties” from Bell Canada, the Court in Pezim imputed relative expertise to the tribunal, including on questions of law, based on its statutory mandates.  In Pezim, the Court reviewed the constating statute of the British Columbia Securities Commission and found that “[t]he breadth of the Commission’s expertise and specialisation is reflected in the provisions of the [B.C. Securities Act]” (p. 593).  This approach of judicially imputing expertise, even on questions of law,  was a departure from earlier jurisprudence that relied on privative clauses as the manifest signal of the legislature’s recognition of relative tribunal expertise.

[88] My colleague Binnie J. writes at para. 26 of his reasons that “Pezim has been cited and applied in numerous cases over the last 15 years.”   In light of this, he rejects what he sees  as my effort “to roll back the Dunsmuir clock”.  With respect, I do not believe that the longevity of Pezim should stand in the way of this Court’s recent attempts to return conceptual clarity to the application of standard of review.  The fact that Pezim has been cited in other cases does not preclude this Court from revisiting its reasoning where there are compelling reasons to do so: R. v. Robinson, [1996] 1 S.C.R. 683, at para. 46.  In my view, Pezim’s departure from the conceptual basis for standard of review constitutes such a compelling reason.  In Dunsmuir, this Court recognized that the time had “arrived for a reassessment” of “the troubling question of the approach to be taken in judicial review of decisions of administrative tribunals” (para. 1).  Such reassessment should include a return to the conceptual basis for standard of review.

[89] I do not dispute that reviewing courts, whether in the appellate or judicial review contexts, should show deference to lower courts and administrative decision-makers on questions of fact: see Deschamps J. in concurrence in Dunsmuir at para. 161.  The principled bases articulated in Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235, at para. 13, for deference to judicial triers of fact are also relevant in the administrative review context.  Just as judicial triers of fact are better situated to make findings of fact at first instance, so too are tribunals, especially in the area of policy making.  In cases involving mixed fact and law, where the legal question cannot be extricated from a factual or policy finding, deference should be shown.

[90] However, where a legal question can be extricated from a factual or policy inquiry, it is inappropriate to presume deference where Parliament has not indicated this via a privative clause.  The basic rule in the appellate context is that questions of law are to be reviewed on a correctness standard: Housen, at para. 8.  The reasons for this are twofold.  First, “the principle of universality requires appellate courts to ensure that the same legal rules are applied in similar situations”: Housen, at para. 9.  Divergent applications of legal rules undermine the integrity of the rule of law.  Dating back to the time of Dicey’s theory of British constitutionalism, almost all rule of law theories include a requirement that each person in the political community be subject to or guided by the same general law:  see A. V. Dicey, Introduction to the Study of the Law of the Constitution (10th ed. 1959),  at p. 193; L. L. Fuller, The Morality of Law (rev. ed. 1969), at pp. 81-91 (advocating the principle of congruence between official action and declared rule); J. Raz, The Authority of Law: Essays on Law and Morality (1979), at pp. 215-17 (“[s]ince the court’s judgment establishes conclusively what is the law in the case before it, the litigants can be guided by law only if the judges apply the law correctly”).  A correctness standard on questions of law is meant, in part, to ensure this universality.  Second, appellate and reviewing courts have greater law-making expertise relative to trial judges and administrative decision-makers.  As this Court emphasized in Housen:

[W]hile the primary role of trial courts is to resolve individual disputes based on the facts before them and settled law, the primary role of appellate courts is to delineate and refine legal rules and ensure their universal application. In order to fulfill the above functions, appellate courts require a broad scope of review with respect to matters of law. [para. 9]

[91] In the administrative context, unlike the appellate context, the legislature may decide that an administrative decision-maker has superior expertise relative to a reviewing court, including on legal questions.  It signals this recognition by enacting a strong privative clause.  It is in these cases that the court must undertake a standard of review analysis to determine the appropriate level of deference that is owed to the tribunal.  It is not for the court to impute tribunal expertise on legal questions, absent a privative clause and, in doing so, assume the role of the legislature to determine when deference is or is not owed.

[92] The distinction between the judicial and legislative roles was further blurred when the privative clause was incorporated into the pragmatic and functional approach in Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 S.C.R.  982.   Pushpanathan set out the four relevant factors for the standard of review analysis:  privative clause, expertise, purpose of the act as a whole and of the provision in particular, and the nature of the problem.  Rather than being viewed as the express manifestation of legislative intent regarding deference, the privative clause was now treated simply as one of several factors in the calibration of deference (standard of review).  As Professor Macklin notes, “[i]f the privative clause was an exercise in communicating legislative intent about the role of the courts, suffice to say that the message was, if not lost, then at least reformulated in translation” (p. 225). 

D.    Legislative Intent

[93] In my opinion, recognizing expertise as a free-standing basis for deference on questions that reviewing courts are normally considered to be expert on (law, jurisdiction, fraud, natural justice, etc.)  departs from the search for legislative intent that governs this area. As Dunsmuir reaffirmed, the rationale behind the common law standard of review analysis is to give effect to legislative intent (Bastarache and LeBel JJ., at para. 30): see also Pushpanathan, at para. 26 (“[t]he central inquiry in determining the standard of review exercisable by a court of law is the legislative intent of the statute creating the tribunal whose decision is being reviewed”); C.U.P.E. v. Ontario (Minister of Labour), 2003 SCC 29, [2003] 1 S.C.R. 539, at para. 149 (standard of review as “seeking the polar star of legislative intent”).

[94] Where the recognition of relative expertise was grounded in the privative clause, the legislature’s intent was clear.  Departures from that conceptual basis  have led courts to undertake what are often artificial judicial determinations of relative expertise.  It seems quite arbitrary, for example, that courts may look at the nature of a tribunal as defined by its enabling statute, but not always conduct a full review of its actual expertise.  Should a reviewing court be required to consider the qualifications of administrative decision-makers on questions that courts are normally considered to have superior expertise?  For example, should it matter whether or not decision-makers have legal training?  In the specific context of statutory interpretation, should the reviewing court scrutinize whether or not the tribunal regularly reviews and interprets particular provisions in its home statute such that it possesses relative expertise with respect to such provisions?  See L. Sossin, “Empty Ritual, Mechanical Exercise or the Discipline of Deference? Revisiting the Standard of Review in Administrative Law” (2003), 27 Advocates’ Q. 478, at p. 491 (for a discussion of the judicial determination of expertise).

[95] Far from subscribing to the view that courts should be reviewing the actual expertise of administrative decision-makers, it is my position that this is the function of the legislature.  In my view, the discordance between imputed versus actual expertise is simply one manifestation of the larger conceptual unhinging of tribunal expertise from the privative clause.  The legislatures that create administrative decision-makers are better able to consider the relative qualifications, specialization and day-to-day workings of tribunals, boards and other decision-makers which they themselves have constituted.  Where the legislature believes that an administrative decision-maker possesses superior expertise on questions that are normally within the traditional bailiwick of courts (law, jurisdiction, fraud, natural justice, etc.), it can express this by enacting a privative clause.

[96] In my respectful view, the majority’s common law standard of review approach seeks two polar stars — express legislative intent and judicially determined expertise — that may or may not align.  While there was some attempt by the majority in Dunsmuir to reconnect these inquiries, the move has been incomplete.  Professor David Mullan notes that “expertise is no longer described as the single most important factor” in Dunsmuir and the privative clause is seen as a “strong indication” of a requirement of deference: “Dunsmuir v. New Brunswick, Standard of Review and Procedural Fairness for Public Servants: Let’s Try Again!” (2008), 21 C.J.A.L.P. 117, at pp. 125-26.  In my view, it is time for the courts to acknowledge that privative clauses and tribunal expertise are two sides of the same coin.

E.     Recognizing the Limitation of Common Law Standard of Review Analysis

[97] Standard of review has dominated so much of administrative law jurisprudence and academic writing to date that one might hope it would, by now, provide a cogent and predictable analysis of when courts should adopt a deferential approach to an administrative decision.  Dunsmuir demonstrates that this is still not the case.  In Dunsmuir, six judges of this Court said that the standard of review applicable to the adjudicator’s legal determination was reasonableness.  Three judges found that the standard was correctness.  Each group focused on different aspects of the adjudicator’s decision-making process.  The majority gave weight to the presence of a strong privative clause, that the adjudicator was imputed to have expertise in interpreting his home statute, that the purpose of the legislation was the timely and binding settlement of disputes, and that the legal question was not outside the specialized expertise of the adjudicator.  The minority focused on the relationship between the common law rules relating to dismissal and those under the Public Service Labour Relations Act, R.S.N.B. 1973, c. P-25.  Because their starting point was the common law, over which the adjudicator was not imputed to have expertise, the minority was of the view that the correctness standard applied. 

[98] What this demonstrates is that the common law standard of review analysis continues to provide little certainty about which standard will apply in a particular case.  How a court will weigh and balance the four standard of review factors remains difficult to predict and therefore more costly to litigate.  In my view, it must be recognized that the common law standard of review analysis does not provide for a panacea of rigorous and objective decision making regarding the intensity with which courts should review tribunal decisions. In attempting to reconcile the court’s constitutional role in the face of a strong privative clause, it may be the best that we have at this point.  But its application outside the privative clause context is, in my view, of highly questionable efficacy.

III.     Judicial Recognition of Legislated Standards of Review

A.      Giving Effect to Legislative Intent

[99] This Court has considered legislative language similar to that in s. 18.1(4) in previous cases and has held that a common law standard of review analysis is not necessary where the legislature has provided for standards of review.  This Court held in R. v. Owen, 2003 SCC 33, [2003] 1 S.C.R. 779, that legislative rules specifying standards of review must be given effect by courts, subject to constitutional limits.  

[100]             The majority now attempts to qualify that holding in Owen.  In my respectful view, that is ignoring the obvious.   The majority insists that although not stated, the “common law of judicial review” was still in play in Owen.  Binnie J. writes that “even in the context of a precisely targeted proceeding related to a named adjudicative board, the standard of review was evaluated by reference to the common law of judicial review” (para. 31).  In my respectful opinion, to say (or imply) that a Dunsmuir standard of review analysis applies even where the legislature has articulated the applicable standard of review directly contradicts Owen.

[101]             The majority nevertheless implies that even if the Dunsmuir standard of review analysis did not apply in Owen, this was only because of the specificity of s. 672.38  of the Criminal Code , R.S.C. 1985, c. C-46  (see para. 29).  That section sets out the standard of review to be applied on judicial review of decisions from Review Boards regarding the liberty of persons found not criminally responsible.  The majority contrasts this with s. 18.1  of the FCA , stating that “[r]esort to the general law of judicial review is all the more essential in the case of a provision like s. 18.1 of the [FCA ] which, unlike s. 672  of the Criminal Code , is not limited to particular issues before a particular adjudicative tribunal” (para. 33).  Thus, even if one rejects the view that a common law standard of review analysis was present in Owen, the majority still says that the generality of s. 18.1  of the FCA  makes it applicable in the present case. 

[102]             The problem with this reasoning is that such qualification would seriously undermine the legislature’s ability to introduce greater certainty and predictability into the standard of review process.  Drawn to its logical conclusion, in order to displace the Dunsmuir standard of review analysis, the majority’s approach would require legislatures to enact standard of review legislation with respect to every single administrative tribunal or decision-maker and perhaps in relation to every type of decision they make.  With respect, this amounts to a serious overreaching of this Court’s role.  It fails to respect the legislature’s prerogative to articulate, within constitutional limits, what standard of review should apply to decision-makers that are wholly the products of legislation. 

[103]             In discussing British Columbia’s Administrative Tribunals Act, S.B.C. 2004, c. 45 (“B.C. ATA”), Binnie J. notes that “most if not all judicial review statutes are drafted against the background of the common law of judicial review” (para. 19).  While I agree with this observation, I disagree with him as to the conclusions that should flow from it.  The majority views the common law background as providing an opening for the continued relevance of a common law standard of review analysis. In reference to s. 58(2)(a) of the B.C. ATA, Binnie J. writes:

Despite Dunsmuir, “patent unreasonableness” will live on in British Columbia, but the content of the expression, and the precise degree of deference it commands in the diverse circumstances of a large provincial administration, will necessarily continue to be calibrated according to general principles of administrative law. [Underlining added; para. 19.]

[104]             The majority would allow for recourse to the common law on several fronts.  First, Binnie J. states that the common law jurisprudence on the “content” of “patently unreasonable” will be relevant.  I agree that the common law will be a necessary interpretive tool where common law expressions are employed by the legislator and are not adequately defined: see R. Sullivan, Sullivan on the Construction of Statutes (5th ed. 2008), at pp. 434-36; R. v. Holmes, [1988] 1 S.C.R. 914; Waldick v. Malcolm, [1991] 2 S.C.R. 456.

[105]             However, the majority would also allow for recourse on a second front.  Binnie J. says that “the precise degree of deference [patently unreasonable] commands in the diverse circumstances of a large provincial administration, will necessarily continue to be calibrated according to general principles of administrative law” (para. 19).  It is unclear exactly which principles of administrative law are being referred to.  If the reference to general principles of administrative law means there is some sort of spectrum along which patent unreasonableness is to be calibrated, that would be at odds with the B.C. legislature’s codification of discrete standards of review. 

[106]             With the ATA, the B.C. legislature expressly codified the standards of review.  However, in order for legislation to be exhaustive on a particular question, legislatures are not required to expressly oust the common law by statute.  In Gendron v. Supply and Services Union of the Public Service Alliance of Canada, Local 50057, [1990] 1 S.C.R. 1298, this Court considered whether recourse to the common law duty of fair representation was appropriate where the legislature had created a statutory duty.  L’Heureux‑Dubé J., writing for a unanimous Court, emphasized that because the content of the statutory remedy was “identical to the duty at common law”, the “common law duty is therefore not in any sense additive; it is merely duplicative” (p. 1316).  The Court went on to hold that

the common law duty of fair representation is neither “necessary [nor] appropriate” in circumstances where the statutory duty applies. Parliament has codified the common law duty and provided a new and superior method of remedying a breach.  It is therefore reasonable to conclude that while the legislation does not expressly oust the common law duty of fair representation, it does however effect this end by necessary implication  . . . .  [Emphasis in original; p. 1319.]

Thus, while recourse to the common law is appropriate where Parliament has employed common law terms or principles without sufficiently defining them, it is not appropriate where the legislative scheme or provisions expressly or implicitly ousts the relevant common law analysis as is the case with s. 18.1(4)  of the FCA .

B.     The Majority’s Concern With the Rigidity of Legislated Standards Is Misplaced

[107]             The majority expresses concern with the rigidity of general legislative schemes in the judicial review context.  With respect to the B.C. ATA, Binnie J. writes of the need for a common law analysis that would account for the “diverse circumstances of a large provincial administration” (para. 19).  In the federal context, he writes: “It cannot have been Parliament’s intent to create . . . a single, rigid Procrustean standard of decontextualized review . . .” (para. 28).  By focussing on the diversity of decision-makers covered by the FCA and the B.C. ATA, the majority’s reasons make prescribed standards appear overly rigid, even arbitrary.

[108]             With respect, the image of the Procrustean bed is misplaced in the judicial review context.  The invocation of the Procrustean image with respect to legislated rules creates the impression that the contrasting common law standard of review is operating in a fluid, fully contextualized paradigm.  This is not the case.  This is not an area where Parliament is imposing rigid conformity against the backdrop of a panoply of common law standards.  The potential flexibility of a contextual common law analysis is already limited in the post-Dunsmuir world of two standards.  Regardless of what type of decision-maker is involved, whether a Cabinet minister or an entry‑level fonctionnaire (para. 28),  the Dunsmuir analysis can only lead to one of two possible outcomes:  reasonableness or correctness.  And, as the present majority makes clear, these are single standards, not moving points along a spectrum (para. 59).

[109]             Moreover, the majority’s concerns regarding legislative rigidity are only realized if one accepts that the focus of the analysis should or must be on the type of administrative decision-maker.  The majority’s argument is that it cannot have been intended for a range of decision-makers to be subject to the same standards of review.  A review of the FCA and the B.C. ATA makes clear, however, that the respective legislatures believed the focus should be on the nature of the question under review (e.g., fact, law, etc.), rather than the nature of the decision-maker.  So there is a diversity in these schemes.  It just operates according to the type of question being reviewed.

[110]             Even given this legislative focus on the type of question under review, it is still not the case that all administrative decision-makers are subject to the same standards of review.  Where a decision-maker’s enabling statute purports to preclude judicial review on some or all questions through a privative clause, deference will apply and a Dunsmuir standard of review analysis will be conducted.  This is precisely how Parliament has legislated in the FCA  context when it intends for greater deference to be shown to certain decision-makers.

[111]             The Canada Labour Code , R.S.C. 1985, c. L-2 , for example, includes a strong privative clause protecting the Canadian Industrial Relations Board from judicial review under the FCA  on questions of law and fact.  Section 22(1) states:

22. (1) Subject to this Part, every order or decision of the Board is final and shall not be questioned or reviewed in any court, except in accordance with the Federal Courts Act  on the grounds referred to in paragraph 18.1(4)(a), (b) or (e) of that Act.

Section 22(1) expressly provides for review on questions of jurisdiction, procedural fairness, fraud or perjured evidence, but excludes review for errors of law or fact through express reference to s. 18.1(4)  of the FCA .  Where the privative clause applies, i.e. with respect to s. 18.1(4)(c), (d), or (f), the court is faced with a tension between its constitutional review role and legislative supremacy.  In such cases, the Dunsmuir analysis applies.  There is no role for the Dunsmuir standard of review analysis where s. 22(1) expressly provides for review on questions of jurisdiction, natural justice and fraud.  Correctness review applies in these cases.

[112]             In contrast, the Immigration and Refugee Protection Act , S.C. 2001, c. 27  — the underlying legislation in the present case — does not contain this type of privative clause.  Section 162(1) only provides that “[e]ach Division of the Board has, in respect of proceedings brought before it under this Act, sole and exclusive jurisdiction to hear and determine all questions of law and fact, including questions of jurisdiction.”  Contrary to the implication of the majority reasons, I think it is plain that this privative clause is intended only to differentiate between different levels and tribunals within the immigration regime and provide each with exclusive jurisdiction to hear legal and factual questions.  It is not a  privative clause that seeks to restrict or preclude judicial review.

[113]             These examples indicate that Parliament has not been unmindful of the issue of standard of review in s. 18.1(4).  Where it intends that a tribunal decision reviewed under s. 18.1 be shown deference, Parliament expressly indicates this either in s. 18.1(4) itself, as it has in para. (d) with respect to facts, or in the underlying legislation such as the Canada Labour Code .  Where it does not, the courts should undertake their review according to the standards of correctness.

[114]             I would note that the B.C. legislature has also turned its mind to these concerns.  The B.C. ATA provides for more deferential standards of review where the underlying statute contains a privative clause.  By imposing different standards of review depending on whether or not the administrative decision is protected by a privative clause, the legislature differentiates between those expert decisions it wished to protect and those it did not (ss. 58 and 59).  The Honourable Geoff Plant indicated this when introducing the B.C. ATA on second reading: 

For tribunals with specialized expertise, like the Farm Industry Review Board and the Employment Standards Tribunal, this bill generally provides that a court must defer to a tribunal’s decision unless the decision is patently unreasonable or the tribunal has acted unfairly. For other tribunals — including, for example, the mental health review panels — the bill provides that with limited exceptions, a court must adopt a standard of correctness in reviewing the tribunal’s decisions.

(Debates of the Legislative Assembly, 5th Sess., 37th Parl., May 18, 2004, at p. 11193)

[115]             The record of the proceedings of the B.C. legislature also makes clear the legislature’s intent to codify standards of review that would oust a duplicative common law standard of review analysis.  The policy rationale for this move was clear.  The legislation was aimed at refocussing judicial review litigation on the merits of the case, rather than on the convoluted process of determining and applying the standard of review.

The question of what the standard of review should be on a case‑by‑case basis is often interpreted by the courts as a search for legislative intent. . . . Accordingly, searching for that intent tends to be a time‑consuming, expensive and sometimes disruptive exercise.

. . .

. . . The provisions in this bill that codify the standards of review will shift the focus from what has been largely a scholarly debate about fine points of law to matters of greater immediate concern to the parties in tribunal proceedings. [Emphasis added.]

(Debates of the Legislative Assembly, at p. 11193)

[116]             It would be troubling, I believe, to the B.C. legislature to think that, despite its effort to codify standard of review and shift the focus of judicial review to the merits of the case, this Court would re-impose a duplicative Dunsmuir-type analysis in cases arising under the B.C. ATA.

IV.     Statutory Interpretation of the Federal Courts Act 

A.      Section 18.1(4)

[117]             Section 18.1(4) appears at para. 71 above.  On my reading, where Parliament intended a deferential standard of review in s. 18.1(4), it used clear and unambiguous language.  The necessary implication is that where Parliament did not provide for deferential review, it intended the reviewing court to apply a correctness standard as it does in the regular appellate context.

[118]             In my opinion, it is useful to analyse s. 18.1(4) by first examining para. (d), which  provides for judicial review where the federal board, commission or other tribunal

(d) based its decision or order on an erroneous finding of fact that it made in a perverse or capricious manner or without regard for the material before it;

In this paragraph, Parliament has expressly instructed courts to show significant deference to the original decision-maker.  The words “in a perverse or capricious manner or without regard for the material before it” are clear and unambiguous.  They indicate that on questions of fact, courts are only to interfere in the most egregious cases of erroneous fact finding.

[119]             Binnie J. also finds that “it is clear from s. 18.1(4)(d) that Parliament intended administrative fact finding to command a high degree of deference” (para. 46).  It would seem that in recognizing that the legislature intended a high degree of deference, one would conclude that this provision speaks for itself and ousts a common law standard of review analysis.  Yet, Binnie J. still suggests that the provision is merely complementary of the common law, rather than dispositive of the standard of review issue.  He writes that s. 18.1(4)(d) “provides legislative precision to the reasonableness standard of review of factual issues” and is “quite consistent with Dunsmuir” (para. 46).  By superimposing Dunsmuir, the majority signals that factual decisions are to be reviewed on a reasonableness standard.  The question then is whether reasonableness implies the same level of deference as “capricious” and “perverse”.  Arguably, a reasonableness review might be less deferential than that intended by the words Parliament used.  Regardless of whether that is true or not, there is no justification for imposing a duplicative common law analysis where the statute expressly provides for the standard of review: see Gendron.

[120]             By contrast with para. (d), there is no suggestion that courts should defer in reviewing a question that raises any of the other criteria in s. 18.1(4).  Parliament recognized that with respect to factual determinations, a federal board, commission or other tribunal is better situated than a reviewing court.  With respect to questions of law, jurisdiction, natural justice, fraud or perjured evidence, the legislation deems courts to have greater expertise than administrative decision-makers.

[121]             There is no suggestion in the FCA  that reviewing courts should defer on questions of law.  Section 18.1(4)(c) provides for review where the federal board, commission or other tribunal

(c) erred in law in making a decision or an order, whether or not the error appears on the face of the record;

I can see no meaningful difference between the terms “the decision is based on a wrong decision on a question of law” which, in Owen, was considered to be sufficient by this Court to determine that a correctness standard of review applied, and “erred in law in making a decision or an order, whether or not the error appears on the face of the record” in s. 18.1(4)(c).  Indeed, in Mugesera v. Canada (Minister of Citizenship and Immigration), 2005 SCC 40, [2005] 2 S.C.R. 100,  a unanimous Court thought that the words of s. 18.1(4) were sufficiently clear that “[u]nder these provisions, questions of law are reviewable on a standard of correctness” (para. 37).  Mugesera, like this case, was a judicial review of a decision of the IAD. 

[122]             The majority now attempts to qualify Mugesera by writing that “[e]rrors of law are generally governed by a correctness standard” (para. 44  (emphasis added)).  With respect, Mugesera did not qualify its application of the correctness standard of review in interpreting s. 18.1(4)(c).  Paragraph 37 of Mugesera states:

Applications for judicial review of administrative decisions rendered pursuant to the Immigration Act are subject to s. 18.1  of the Federal Court Act .  Paragraphs (c) and (d) of s. 18.1(4), in particular, allow the Court to grant relief if the federal commission erred in law or based its decision on an erroneous finding of fact.  Under these provisions, questions of law are reviewable on a standard of correctness.

Moreover, contrary to what the present majority implies, the Court in Mugesera did not limit the application of the correctness standard to “the general questions of international law and criminal law at issue in that case” (para. 44 of majority).  It is clear that as a matter of statutory interpretation, the Court understood s. 18.1(4)(c) as requiring a correctness review on questions of law.  The Court saw no need to impose the common law over what the statute itself dictated. 

[123]             The majority nevertheless insists that “para. (c) provides a ground of intervention, but the common law will stay the hand of the judge(s) in certain cases if the interpretation is by an expert adjudicator interpreting his or her home statute or a closely related statute” (para. 44 (emphasis in original)).  With respect, there is no authority for this in the legislation.  The majority finds their opening in the remedial discretion of s. 18.1(4).  Binnie J. writes: “This nuance does not appear on the face of para. (c), but it is the common law principle on which the discretion provided in s. 18.1(4)  is to be exercised” (para. 44).  As I will explain, the remedial discretion in s. 18.1(4) goes to the question of withholding relief, not the review itself.  The bases upon which the remedial discretion is to be exercised are wholly distinct from the common law of standard of review analysis. 

[124]             Paragraphs (a), (b) and (e) of s. 18.1(4) provide for relief where a federal board, commission or other tribunal

(a)   acted without jurisdiction, acted beyond its jurisdiction or refused to exercise its jurisdiction;

(b)   failed to observe a principle of natural justice, procedural fairness or other procedure that it was required by law to observe;

. . .

(e)   acted, or failed to act, by reason of fraud or perjured evidence;

There is no indication in any of these provisions that the legislature intended for the reviewing court to show any deference to administrative decision-makers in determining questions of jurisdiction, natural justice, procedural fairness and fraud or perjured evidence.

[125]             Section 18.1(4)(f) contemplates judicial intervention where the federal board, commission or other tribunal

(f)   acted in any other way that was contrary to law.

The majority  writes that s. 18.1(4)(f) “necessarily includes ‘law’ outside the [FCA ]” (para. 48) and therefore demonstrates that “s. 18.1(4) is not intended to operate as a self-contained code, but is intended by Parliament to be interpreted and applied against the backdrop of the common law, including those elements most recently expounded in Dunsmuir” (para. 48).  The majority relies on the statement by the authors of Federal Courts Practice 2009 (2008), B. J. Saunders et al., that “[s]ection 18.1(4)(f) ensures that the Court will not be hindered in developing new grounds for review” (para. 49 (emphasis added by Binnie J.)).

[126]             It is not in dispute that s. 18.1(4) is not intended to operate as a self-contained code.  In judicial review of any administrative decision where a legal error is alleged, the court is required to consider whether the decision-maker was in breach of any statutory provision or common law rule that might be relevant.  In this regard, I agree that s. 18.1(4)(f) provides for potentially expanded grounds of review.  However, that is not the issue in this case.  The issue in this case is whether Parliament has legislated exhaustively on the standard of review, so as to oust the Dunsmuir standard of review analysis.  Binnie J.’s reliance on Saunders et al.’s discussion of the “grounds of review” under s. 18.1(4) does not address whether the section also provides for standards of review.  This is troubling, given that those same commentators find that s. 18.1(4) does provide for standards of review on questions of fact and law.  At p. 145 of their text under the title “Grounds for Review — Standard of Review — Generally” in commenting on Mugesera, they write:

Under section 18.1(4) (c) and (d) of the Federal Courts Act , questions of law are reviewable on a standard of correctness.  On questions of fact, the reviewing court can intervene under section 18.1(4)(d) only if it considers that the tribunal “based its decision or order on an erroneous finding of fact that it made in a perverse or capricious manner or without regard for the material before it”.

[127]             All that s. 18.1(4)(f) provides for is review of legal errors committed by a federal board, commission or other tribunal other than those “in making a decision or an order”, which are already captured under s. 18.1(4)(c): see Morneault v. Canada (Attorney General), [2001] 1 F.C. 30 (C.A.), at para. 44 (“the intent of the paragraph appears to have been to afford a ground that was not otherwise specifically mentioned in subsection 18.1(4)”).  A tribunal’s refusal to make a decision or an order, for example, would not come under para. (c).  The reference to “acted in any other way that was contrary to law” refers, then, to legal errors that are not captured by s. 18.1(4)(c).  It does not provide an opening for a Dunsmuir standard of review analysis.  With respect, the majority’s view of s. 18.1(4) ignores the obvious interpretation in search of something that is not there.

B.  Section 18 and the Origins of the Federal Courts Act 

[128]             The majority is of the view that when s. 18.1 was added to the FCA , it “did not have the effect of excluding the common law” (para. 34).  It appears that this proposition is intended to act as a platform for the applicability of the common law standard of review analysis.  With respect, it is overly broad to suggest that all elements of the common law continued to apply to s. 18.1(4) simply because there were some gaps — for example, criteria in exercising the discretion to withhold relief — which the common law continued to fill.  For the reasons I have explained, the FCA  occupies the area of standard of review and therefore ousts the application of the common law on this question.

[129]             The genesis of the FCA and its amendments is not in dispute. Section 18 was enacted to transfer jurisdiction from the provincial superior courts to the federal courts for judicial review of federal tribunals, subject to provincial courts retaining a residual jurisdiction to determine the constitutionality and applicability of legislation.  Section 18, which refers to the prerogative writs, survives, but no application for judicial review can be made under it.  Subsection (3) provides:

The remedies provided for in subsections (1) and (2) may be obtained only on an application for judicial review made under section 18.1.

Section 18.1 contains the statutory process under which judicial review may be applied for and under which the court exercises its jurisdiction. 

[130]             The 1990 amendments were intended to clarify pre-existing procedural confusion about whether the trial or appeal divisions had jurisdiction with regard to particular applications for judicial review.  The amendments also aimed to simplify the procedure for obtaining a remedy by requiring that it be sought by way of application for judicial review, rather than by way of statement of claim or originating notice of motion as had been the prior practice: D. Sgayias et al., Federal Court Practice 1998 (1997), at pp. 69-70.  As reform legislation, the amendments did not concern the standard of review. 

C.      The Implications of Section 18.1(4) Remedial Discretion

[131]             I agree with Binnie J.’s bilingual analysis and conclusion that, “notwithstanding the bilingual issue in the text, s. 18.1(4) should be interpreted so as to preserve to the Federal Court a discretion to grant or withhold relief” (para. 40).  The pertinent question is what should form the basis for the exercise of that judicial discretion.    Relief on judicial review is equitable.  The discretion in s. 18.1(4) recognizes that it may be inappropriate to grant equitable relief in some cases.  This remedial discretion allows a reviewing judge to withhold relief in certain cases.  It does not concern the review itself, however.

[132]             The majority says that the FCA  does not “indicate in what circumstances . . . relief may properly be withheld” (para. 42).  It is true that the legislation does not provide for criteria according to which reviewing courts should exercise their discretion to withhold relief.  In the context of this specific gap, I agree with the majority that “resort will have to be had to the common law” (para. 42).  The pertinent question is which part of the common law is relevant to the withholding of relief by the court on judicial review.

[133]             Binnie J. attempts to  ground the court’s remedial discretion to withhold relief in general judicial review principles.    He states at para. 36 that the court’s exercise of the s. 18.1(4) discretion “will depend on the court’s appreciation of the respective roles of the courts and the administration as well as the ‘circumstances of each case’: see Harelkin v. University of Regina, [1979] 2 S.C.R. 561, at p. 575.”   He cites Brown and Evans’ observation that “whenever the court exercises its discretion to deny relief, balance of convenience considerations are involved” (para. 36); D. J. M. Brown and J. M. Evans, Judicial Review of Administrative Action in Canada (loose-leaf), at p. 3-99.  While “the discretion must be exercised judicially”, Binnie J. finds that “the general principles of judicial review dealt with in Dunsmuir provide elements of the  appropriate judicial basis for its exercise” (para. 36).

[134]             By linking remedial discretion to Dunsmuir “general principles of judicial review”, Binnie J. conflates standard of review (deference) with the granting of relief.  In doing so, he effectively reads in an opening for recourse to the common law standard of review analysis.  He relies on the specific gap regarding the discretion to grant relief to impute a wider gap regarding standard of review.

[135]             With respect, this is not the nature of the discretion under s. 18.1(4).  The traditional common law discretion to refuse relief on judicial review concerns the parties’ conduct, any undue delay and the existence of alternative remedies: Immeubles Port Louis Ltée v. Lafontaine (Village), [1991] 1 S.C.R. 326, at p. 364.  As Harelkin  affirmed, at p. 575, courts may exercise their discretion to refuse relief to applicants “if they have been guilty of unreasonable delay or misconduct or if an adequate alternative remedy exists, notwithstanding that they have proved a usurpation of jurisdiction by the inferior tribunal or an omission to perform a public duty”.  As in the case of interlocutory injunctions, courts exercising discretion to grant relief on judicial review will take into account the public interest, any disproportionate impact on the parties and the interests of third parties.  This is the type of “balance of convenience” analysis to which Brown and Evans were referring.

[136]             Thus, the discretion contained in s. 18.1(4) speaks to  the withholding of relief in appropriate cases;  it does not engage the question of standard of review.  Reliance upon it by the majority to support the view that it opens the door to the Dunsmuir standard of review analysis is, with respect, misplaced.  In my view, the Dunsmuir standard of review should be confined to cases in which there is a strong privative clause.  Excepting such cases, it does not apply to s. 18.1(4)  of the FCA .

V.    Decision in This Case

[137]             In determining whether the respondent was eligible for the special relief available under s. 67(1) (c) of the  Immigration and Refugee Protection Act , the IAD acknowledged that its discretion should be exercised with consideration for the criteria set out in Ribic v. Canada (Minister of Employment and Immigration), [1985] I.A.B.D. No. 4 (QL) (endorsed by this Court in Chieu v. Canada (Minister of Citizenship and Immigration), 2002 SCC 3, [2002] 1 S.C.R. 84).  The majority of the IAD expressly referred to the Ribic factors and, in my view, had regard to those it considered relevant in exercising its discretion.  The actual application of the Ribic factors to the case before it and its exercise of discretion is fact-based.  I do not find that the factual findings of the IAD were perverse or capricious or were made without regard to the evidence.  I would allow the appeal.

The following are the reasons delivered by

[138]             Deschamps J. — I agree with Rothstein J. that since s. 18.1(4)  of the Federal Courts Act , R.S.C. 1985, c. F-7 , sets legislated standards of review, those standards oust the common law.  Consequently, I agree with Parts III, IV and V of his reasons and would also allow the appeal.

The following are the reasons delivered by

Fish J. (dissenting) —

I

[139]             This appeal raises two issues. The first concerns the standard of review with respect to decisions of the Immigration Appeal Division (“IAD”). In that regard, I agree with Justice Binnie that the standard of review is “reasonableness”.

[140]             The second issue is whether the majority decision of the IAD in this case survives judicial scrutiny under that standard. Unlike Justice Binnie, and with the greatest of respect, I have concluded that it does not. 

[141]             Essentially, I find that the decision of the IAD rests on what the Court of Appeal has aptly described as a “fixation” that collides with the overwhelming weight of the uncontradicted evidence in the record before it.  I agree with the majority below that the decision, for this reason, cannot stand.

[142]             Accordingly, I would dismiss the appeal.

II

[143]             In 2000, when he was 18 years old, Sukhvir Singh Khosa caused the death of Irene Thorpe by driving recklessly at more than twice the speed limit, losing control of his automobile and running it off the roadway.  He had by then been living in Canada for four years.  When his appeal to the IAD was decided in 2004, he was 22 and married.  Four more years have elapsed since then.

[144]             To order Mr. Khosa’s removal would separate him from his wife and immediate family.  It would return him to a country he has visited only once since emigrating at the age of 14  and where he appears to have few relatives.

[145]             The IAD’s task in this case is to look to “all the circumstances of the case” in order to determine whether “sufficient humanitarian and compassionate considerations” existed to warrant relief from a removal order:  Immigration and Refugee Protection Act , S.C. 2001, c. 27, s. 67(1) (c).  The IAD is bound in performing that function to consider the various factors set out  in Ribic v. Canada (Minister of Employment and Immigration), [1985] I.A.B.D. No. 4 (QL), and endorsed by this Court in  Chieu v. Canada (Minister of Citizenship and Immigration), 2002 SCC 3, [2002] 1 S.C.R. 84, at para. 90.  The IAD in this case placed the greatest emphasis on the factors of remorse, rehabilitation, and likelihood of reoffence.

[146]             With respect to these factors, the record before the IAD consisted essentially of the following uncontested and uncontradicted evidence:

·                  The sentencing judge found that “by his actions immediately after learning of Ms. Thorpe’s death and since the accident . . . he has expressed remorse” (R. v. Khosa, 2003 BCSC 221, [2003] B.C.J. No. 280 (QL), at para. 56).

·                  Mr. Khosa took responsibility for his crime early, expressing a desire to attend the funeral of the woman he had killed and offering — before any arraignment or preliminary inquiry — to plead guilty to dangerous driving causing death.

·                  The sentencing judge found that “[i]n the more than two years that have passed since the accident, Mr. Khosa has not left the house except to go to work, to school, or the Sikh temple.  He normally does not drink.  He does not take drugs.  He has no criminal record.  He has no driving record.  He has complied with all of his bail conditions and is not likely to re-offend” (para. 55 (emphasis added)).

·                  Mr. Khosa had not driven a car since the accident, even though he was permitted to do so for some months thereafter.

·                  Mr. Khosa’s probation officer concluded from close and extensive contact with Mr. Khosa that he “appears to be making a sincere effort to maintain a stable and responsible life style in which he is a contributing member of the community”.  The probation officer also reported that he has “demonstrated a very positive attitude towards community supervision [and] willingly conforms to the expectations, requirements, and restrictions of the Conditional Sentence”.  With respect to his character, the probation officer was of the view that Mr. Khosa “presents as a pro‑social young man who values work, family, community and religion”  (Appellant’s Record, at p. 355).

·                  Mr. Khosa had no previous criminal or driving convictions whatever.

·                  Mr. Khosa had complied with all provisions of his conditional sentence.

·                  Several employers wrote letters describing Mr. Khosa as conscientious and reliable.

[147]             Despite all of this evidence indicating that Mr. Khosa was extremely unlikely to reoffend and had taken responsibility for his actions, the majority at the IAD seized upon one consideration:  Mr. Khosa’s denial that he was “street-racing” at the time the accident occurred.  Apart from a brief mention of Mr. Khosa’s “show of relative remorse at [the] hearing” ([2004] I.A.D.D. No. 1268 (QL), at para. 15), and a passing allusion to the judgments of the criminal courts to his culpability (para. 14), Mr. Khosa’s denial was the only consideration that the IAD majority considered with respect to these issues.  Manifestly, this solitary fact was the decisive element — if not the sole basis — upon which the majority of the IAD denied Mr. Khosa’s basis for all humanitarian and compassionate relief.

[148]             So much cannot reasonably be made out of so little.

III

[149]             While Mr. Khosa’s denial of street racing may well evidence some “lack of insight” into his own conduct, it cannot reasonably be said to contradict — still less to outweigh, on a balance of probabilities  — all of the evidence in his favour on the issues of remorse, rehabilitation and likelihood of reoffence.

[150]             The IAD’s cursory treatment of the sentencing judge’s findings on remorse and the risk of recidivism are particularly troubling.  While findings of the criminal courts are not necessarily binding upon an administrative tribunal with a distinct statutory purpose and a different evidentiary record, it was incumbent upon the IAD to consider those findings and to explain the basis of its disagreement with the decision of the sentencing judge.  The majority decision at the IAD mentions only in passing the favourable findings of the criminal courts and does not explain at all its disagreement with them.

[151]             Moreover, Mr. Khosa’s denial of street racing is, at best, of little probative significance in determining his remorse, rehabilitation and likelihood of reoffence.  In light, particularly, of the extensive, uncontradicted and unexplained evidence to the contrary, Mr. Khosa’s denial of street racing cannot reasonably support the inference drawn from it by the majority in the IAD. 

[152]             It is also important to note that street racing was not a necessary element of  Mr. Khosa’s crime of criminal negligence causing death (R. v.  Khosa, 2003 BCCA 644, 190 B.C.A.C. 23, at para. 85).  It appears that Mr. Khosa’s refusal to accept his guilty verdict on this charge — in contrast with  his willingness to plead guilty to the less serious charge of dangerous driving causing death — is due solely to his mistaken impression that the former requires a finding that he was racing (Appellant’s Record, at p. 145).  This is therefore not a case where a person in deportation proceedings maintains his innocence, as suggested by the majority of the IAD (at para. 14), but rather a case where the immigrant simply disputes an ancillary finding of the criminal court.

[153]             Whatever the correct interpretation of Mr. Khosa’s denial that he was street-racing, it is clear that the majority at the IAD had “some kind of fixation” — to again borrow the phrase of the majority below — with this piece of evidence, and based its refusal to grant humanitarian and compassionate relief largely on this single fact.

[154]             The majority at the IAD made repeated reference to the denial.  Toward the end of its decision, it stated that in light of Mr. Khosa’s “failure . . . to acknowledge his conduct and accept responsibility for . . . street-racing . . ., there is insufficient evidence upon which I can make a determination that [Mr. Khosa] does not represent a present risk to the public” (para. 23 (emphasis added)).  I find that this conclusion is not only incorrect, but unreasonable.  There was ample evidence suggesting that he posed no risk.  The majority decision of the  IAD simply disregarded virtually all of that evidence.

[155]             Later, in justifying its decision to deny all relief rather than order a stay of removal, the majority wrote that Mr. Khosa’s “failure to acknowledge or take responsibility for his specific reckless conduct does not suggest that any purpose would be served by staying the present removal order” (para. 24).  Here, again, the decision of the IAD majority transforms a limited, specific and ancillary denial into a general failure to take responsibility.

[156]             The majority’s inordinate focus on racing and its failure to consider contrary evidence do not “fit comfortably with the principles of justification, transparency and intelligibility” that are required in order to withstand reasonableness review (reasons of Binnie J., at para. 59). 

[157]             With respect, I thus feel bound to conclude that the IAD was unreasonable in its evaluation of Mr. Khosa’s rehabilitation, remorse and likelihood of reoffence.

IV

[158]             Because the IAD’s finding on these specific factors was central to its ultimate decision to deny any and all humanitarian and compassionate relief, the IAD’s determination cannot be sustained.

[159]             To be sure, the majority at the IAD stated that even if it were to have found that Mr. Khosa did not present a risk to the public “in balancing all the relevant factors, I determine the scale does not tip in [Mr. Khosa’s] favour and decline to exercise favourable discretion” (para. 23).  This sort of conclusory statement, however,  cannot insulate the IAD’s decision from review when the rest of its reasons demonstrate that its decision rests on an unreasonable determination of central importance, as in this case.

[160]             I agree that decisions of the IAD are entitled to deference.  In my respectful view, however, deference ends where unreasonableness begins. 

V

[161]             For all these reasons, as stated at the outset, I would dismiss the appeal and affirm the judgment of the Court of Appeal returning this matter to the IAD for reconsideration before a differently constituted panel.

Appeal allowed, Fish J. dissenting.

Solicitor for the appellant:  Attorney General of Canada, Vancouver.

Solicitor for the respondent:  Garth Barriere, Vancouver.

Solicitors for the intervener:  Arvay Finlay, Vancouver.