Judgments

[2011] 2 R.C.F. 221

2010 FCA 56

A-129-08

A-130-08

Public Service Alliance of Canada (Appellant)

v.

Canada Post Corporation and Canadian Human Rights Commission (Respondents)

A-139-08

Canadian Human Rights Commission (Appellant)

v.

Canada Post Corporation and Public Service Alliance of Canada (Respondents)

Indexed as: Canada Post Corp. v. Public Service Alliance of Canada

Federal Court of Appeal, Sexton, Evans and Ryer JJ.A.—Ottawa, November 3, 2009 and February 22, 2010.

Human Rights — Appeals from Federal Court decision allowing judicial reviews of Canadian Human Rights Tribunal decision finding that Canada Post Corporation (CPC) violating Canadian Human Rights Act, s. 11 by paying employees in male-dominated Postal Operations (PO) group more than employees in female-dominated Clerical and Regulatory (CR) group for work of equal value — Tribunal examining four elements necessary for establishing, on balance of probabilities, prima facie case of wage discrimination — One such element being that work between CR, PO groups of equal value (work of equal value element) — Federal Court finding that Tribunal misapplying standard of proof by adopting standard used to assess damages, erring in law by applying standard of proof lower than balance of probabilities threshold regarding reliability of job information — Principal issues whether Tribunal failing to find that work of equal value element established; whether Federal Court erring in concluding that Tribunal failing to apply correct standard of proof — (1) Tribunal failing to find that work of equal value element established on balance of probabilities — Terminating analysis required to satisfy element after considering reliability of evidence related thereto — Not interpreting phrase “work of equal value” or explaining applicability to circumstances herein — Presence of reliable evidence related to job information, evaluation plan, evaluation process not meaning that work of equal value element established — Tribunal conflating requisite conclusion with three evidentiary matters needing to be present to permit that conclusion — Use of balance of probabilities standard, “sub-bands of reasonable reliability” unusual — Findings that evidence worthy of consideration not necessarily leading to conclusion that work of equal value element established — (2) Tribunal’s findings with respect to job information, evaluation plan, evaluation process not established on balance of probabilities — Tribunal seeming to justify relaxation of rules with respect to burden, standard of proof — Appeals dismissed — Per Evans J.A. (dissenting): (1) Tribunal not erring over selection of PO comparator group — Selection of PO group not contrary to purpose of Act, s. 11 — Presence of women within male-dominated comparator group acceptable — Well-paid women at CPC not necessarily precluding existence of systemic gender discrimination elsewhere — (2) Tribunal not diluting standard of proof when using phrases “more likely than not”, “sufficiently adequate”, but directing itself on task of weighing sufficiency of evidence — Reduction of monetary award not indication that evidence falling short of balance of probabilities — (3) Choice of methodology to determine existence, extent of wage gap within discretion of Tribunal — (4) Open to Tribunal to reduce amount of compensation.

These were three appeals from a decision of the Federal Court allowing two applications for judicial review of a decision by the Canadian Human Rights Tribunal finding that Canada Post Corporation (CPC) had engaged in a discriminatory practice in violation of section 11 of the Canadian Human Rights Act by paying employees in the male-dominated Postal Operations (PO) group more than employees in the female-dominated Clerical and Regulatory (CR) group for work of equal value. In dockets A-129-08 and A-139-08, the Public Service Alliance of Canada (PSAC) and the Canadian Human Rights Commission (CHRC) appealed the Federal Court’s decision to set aside the Tribunal’s decision. In docket A-130-08, PSAC appealed the Federal Court’s dismissal of a judicial review of the Tribunal’s decision to reduce the damages awarded against CPC by half.

In its 2005 decision upholding the complaint, the Tribunal examined four elements necessary to establish, on a balance of probabilities, a prima facie case of wage discrimination. It found that (1) the complainant occupational group, the CR group, is predominantly female, and the comparator occupational group, the PO group, is predominantly male, (2) the two groups are employed in the same establishment, (3) an assessment of the value of the work being compared between the two groups established that the work is of equal value, and (4) a comparison of the wages between the two groups demonstrated that the CR group was being paid a lesser wage than the PO group. The Federal Court examined whether the Tribunal applied the proper standard of review in regards to the third element, specifically the reliability of the job information from the occupational groups being compared. The Federal Court found that the Tribunal had recognized the balance of probabilities as the correct standard of proof required to establish the essential element of work of equal value, but had misapplied it by adopting a standard used to assess damages. The Federal Court also found that the Tribunal erred in law in applying a novel standard of proof, with respect to the reliability of the job information, that was lower than the balance of probabilities threshold and was more akin to a “reasonable basis in the evidence”.

The principal issues were (1) whether the Tribunal failed to make a finding that the third element of a prima facie case of wage discrimination had been established and, if so, whether such a failure would vitiate the Tribunal’s decision that such a case had been made out against CPC; and (2) whether the Federal Court erred in concluding that the Tribunal failed to apply the correct standard of proof with respect to its findings in relation to the elements required to establish a prima facie case of discrimination.

Held (Evans J.A. dissenting), the appeals should be dismissed.

Per Sexton and Ryer JJ.A.: (1) The Tribunal failed to make the requisite finding that the third element, or “fact in issue”, had been established. Specifically, the Tribunal terminated its analysis of the third element after considering the reliability of the evidence related to it. It failed to conclude that such element was established on a balance of probabilities. The Tribunal was required to follow three steps, or evidentiary matters, to satisfy each element: whether evidence relating to the particular element is admissible, what weight should be given to that admissible evidence, and whether the admissible evidence, taking into account its reliability, establishes the element on the appropriate standard of proof, herein the balance of probabilities. The Tribunal failed to interpret the phrase “work of equal value” for the purpose of determining whether a comparison of the CR and PO groups established that their work is equal in value, and to explain how it applies in the circumstances of the complaint. Such an explanation was necessary before the Tribunal could make a determination that the third element had been established. Notwithstanding those failures, the Tribunal determined that the evidence relating to job information, the evaluation plan and the evaluation process was reliable. However, the presence of such reliable evidence does not automatically lead to the conclusion that the third element has been established, but is a necessary precondition. Instead of concluding that the third element had been established, the Tribunal conflated the requisite conclusion with the three evidentiary matters that must be present to permit that conclusion.

Regarding the Tribunal’s assessment of the reasonable reliability of the job information, the use of the balance of probabili- ties standard as well as “sub-bands of reasonable reliability” is unusual. The Tribunal’s findings with respect to the reliability of the evidence pertaining to the job information, the evaluation plan and the evaluation process are nothing more than findings that such evidence has some probative value and is worthy of consideration. Such findings do not necessarily lead to a conclusion that the third element has been established. The absence of such a conclusion was sufficient to dismiss the complaint.

(2) The presumption that the Tribunal applied the correct standard of proof was amply rebutted. First, the balance of probability standard requires the establishment of a “fact in issue”. In the binary formulation, as described in In re B (Children) (Fc), [2008] UKHL 35, the finding is either zero or one. The necessary finding cannot be “reasonably reliably one”, or “almost one” or “closer to one than zero”. Second, the Tribunal’s adoption of the principle that when assessing damages the assessing body must do the best it can with the evidence that it has, indicates that the Tribunal had concerns with the evidentiary record before it. It is not acceptable for the Tribunal to rely upon that approach when making findings with respect to job information, the evaluation plan and the evaluation process where they constitute “facts in issue” that must be established on a balance of probabilities. Finally, by referring a number of times to the difficult, unusual or litigious circumstances of the case, the Tribunal seemed to justify a relaxation of the long-standing rules with respect to the burden of proof and the standard of proof in civil matters. This was amplified by the adoption of the “bands and sub-bands” of acceptability or reasonable reliability. The Tribunal thus failed to make the requisite finding that the third element had been established on a balance of probabilities.

Per Evans J.A. (dissenting): There were four issues herein: (1) Was the choice of the comparator group unreasonable because it included a substantial number of well-paid women? (2) Did the Tribunal apply the correct standard of proof when finding that members of the complainant and comparator groups were performing work of equal value? (3) Did the Tribunal commit a reviewable error in finding as a fact that the CR group was being paid less than the PO group for performing work of equal value? (4) Did the Tribunal err in law in awarding PSAC compensation in the amount of half of the CR group’s lost wages?

(1) The Federal Court erred in law in concluding that the Tribunal had committed reviewable error in the exercise of its discretion over the selection of the PO comparator group. The presence of a substantial number of well-paid women in the comparator group did not undermine the CR group’s complaint of systemic gender discrimination. The selection of the PO group is thus not contrary to the purpose of section 11 of the Act. The Equal Wages Guidelines, 1986 (Guidelines) specifically contemplate the presence of women within a male-dominated comparator group. As well, the presence of well-paid women at CPC does not necessarily preclude the existence of systemic gender discrimination elsewhere in the corporation. Finally, no principle was referred to suggesting that some members of an occupational group from the comparator group may be removed. The Guidelines themselves do not suggest that only part of an occupational group may be used.

(2) The Tribunal did not dilute the standard of proof when it asked whether it is “more likely than not” that the job information, the evaluation system and the process employed are “sufficiently adequate” to enable a “fair and equitable conclusion” to be reached on whether there were wage differences for work of equal value. It was merely directing itself on its task of weighing the sufficiency of the evidence in order to reach a “fair and equitable conclusion”. Nor can it be inferred that the Tribunal’s reduction of the monetary award to 50 percent of the wages lost was an indication that the evidence fell short of the balance of probabilities.

(3) In order to determine the existence and extent of a wage gap, the Tribunal was presented with several methodologies, and chose the one proposed by the CHRC, which emphasized the content of the work performed. The choice of an appropriate methodology is within the discretion of the Tribunal and is entitled to a high degree of deference. It cannot be said that only CPC’s proposed methodology was reasonably consistent with the objectives of section 11 of the Act, or that only the CHRC methodology was unreasonable.

(4) Specialized tribunals are owed a particularly high degree of deference in their exercise of a broad statutory discretion to fashion an appropriate remedy. The Tribunal directed itself correctly in law when it stated that an award of compensation should aim to make the victims whole. However, it was also open to the Tribunal to extend by analogy principles used to take into account future uncertainties to uncertainties about the past, and on this basis to reduce the amount of compensation.

STATUTES AND REGULATIONS CITED

Canadian Human Rights Act, R.S.C., 1985, c. H-6, ss. 11, 27(2) (as am. by S.C. 1998, c. 9, s. 20), 49(1) (as am. idem), 50(3) (as am. idem), 53(2)(c) (as am. idem, s. 27).

Equal Wages Guidelines, 1986, SOR/86-1082, ss. 12-15.

Federal Courts Act, R.S.C., 1985, c. F-7, ss. 1 (as am. by S.C. 2002, c. 8, s. 14), 18.1(3)(b) (as enacted by S.C. 1990, c. 8, s. 5; 2002, c. 8, s. 27), (4)(d) (as enacted by S.C. 1990, c. 8, s. 5; 2002, c. 8, s. 27), 52(b) (as am. idem, s. 50).

Ontario Human Rights Code, R.S.O. 1980, c. 340.

CASES CITED

applied:

In re B (Children) (Fc), [2008] UKHL 35; F.H. v. McDougall, 2008 SCC 53, [2008] 3 S.C.R. 41, 297 D.L.R. (4th) 193, [2008] 11 W.W.R. 414.

considered:

Public Service Alliance of Canada v. Canada (Department of National Defence), [1996] 3 F.C. 789, (1996), 27 C.H.R.R. D/3488, 199 N.R. 81 (C.A.); Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190, 323 N.B.R. (2d) 1, 291 D.L.R. (4th) 577; Prairie Acid Rain Coalition v. Canada (Minister of Fisheries and Oceans), 2006 FCA 31, [2006] 3 F.C.R. 610, 265 D.L.R. (4th) 154, 21 C.E.L.R. (3d) 175; Ontario Human Rights Commission and O’Malley v. Simpsons-Sears Ltd. et al., [1985] 2 S.C.R. 536, (1985), 52 O.R. (2d) 799, 23 D.L.R. (4th) 321; Canada (Public Service Alliance) v. Canada (Treasury Board), 1996 CanLII 1874 (C.H.R.T.).

referred to:

Canada (Human Rights Commission) v. Canadian Airlines International Ltd., 2004 FCA 113, [2004] 3 F.C.R. 663, 238 D.L.R. (4th) 255, 318 N.R. 201; Canada (Attorney General) v. Morgan, [1992] 2 F.C. 401, (1991), 85 D.L.R. (4th) 473, 135 N.R. 27 (C.A.); Chopra v. Department of National Health and Welfare, 2004 CHRT 27, affd sub nom. Chopra v. Canada (Attorney General), 2007 FCA 268, [2008] 2 F.C.R. 393, 283 D.L.R. (4th) 634, 60 C.C.E.L. (3d) 1; Singh v. Statistics Canada, [1998] C.H.R.D. No. 7 (QL); Telfer v. Canada (Revenue Agency), 2009 FCA 23, [2009] 4 C.T.C. 123, 2009 DTC 5046, 386 N.R. 212; Bell Canada v. Canadian Telephone Employees Association, 2003 SCC 36, [2003] 1 S.C.R. 884, 227 D.L.R. (4th) 193, [2004] 1 W.W.R. 1; Canada (Citizenship and Immigration) v. Khosa, 2009 SCC 12, [2009] 1 S.C.R. 339, 304 D.L.R. (4th) 1, 82 Admin. L.R. (4th) 1; Canada (Human Rights Commission) v. Canadian Airlines International Ltd., 2006 SCC 1, [2006] 1 S.C.R. 3, 263 D.L.R. (4th) 1, 343 N.R. 308.

AUTHORS CITED

Canada. Pay Equity Task Force. Pay Equity: A New Approach to a Fundamental Right, Ottawa: Public Works and Government Services, 2004.

Dyzenhaus, David. “The Politics of Deference: Judicial Review and Democracy”, in Michael Taggart, ed. The Province of Administrative Law, Oxford: Hart Publishing, 1997.

Waddams, S. M. The Law of Damages, loose-leaf ed., Toronto: Canada Law Book Inc., 2004.

appealS from a decision of the Federal Court (2008 FC 223, [2008] 4 F.C.R. 648) allowing two applications for judicial review of a decision by the Canadian Human Rights Tribunal (2005 CHRT 39) finding that Canada Post Corporation had engaged in a discriminatory practice in violation of section 11 of the Canadian Human Rights Act by paying employees in a male-dominated group more than employees in a female-dominated group for work of equal value. Appeals dismissed, Evans J.A. dissenting.

APPEARANCES

James G. Cameron, David G. Yazbeck and Kim N. Patenaude for appellant/respondent Public Service Alliance of Canada.

Philippe Dufresne and Daniel Poulin for respondent/ appellant Canadian Human Rights Commission.

Peter A. Gall and Robert W. Grant for respondent Canada Post Corporation.

SOLICITORS OF RECORD

Raven, Cameron, Ballantyne & Yazbeck LLP/s.r.l., Ottawa, for appellant/respondent Public Service Alliance of Canada.

Canadian Human Rights Commission, Ottawa, for respondent/appellant Canadian Human Rights Commission.

Heenan Blaikie LLP, Montréal, for respondent Canada Post Corporation.

The following are the reasons for judgment rendered in English by

Sexton and Ryer JJ.A.:

I. INTRODUCTION

[1] For convenience, these reasons are organized under the following headings:

Paragraph

(b) Element two—Employment in the same establishment..                  39

(ii) Element two—Employment in the same establishment                101

C. Whether the Tribunal applied the incorrect standard of proof             128

[2] The Public Service Alliance of Canada (PSAC) filed a complaint (the complaint) against Canada Post Corporation (CPC) in 1983, alleging discrimination by CPC against “employees in the female-dominated Clerical and Regulatory Group” by paying “employees in the male-dominated Postal Operations Group” more than the Clerical and Regulatory group employees for work of equal value, contrary to section 11 of the Canadian Human Rights Act [R.S.C., 1985, c. H-6].

[3] In 2005, the Canadian Human Rights Tribunal (the Tribunal) finally released a decision upholding the complaint [Public Service Alliance of Canada v. Canada Post Corp., 2005 CHRT 39].

[4] In 2008, the Federal Court allowed an application for judicial review brought by CPC and directed that the complaint be dismissed.

[5] In order for the complaint to be upheld, the Tribunal itself determined, and it was not disputed, that among other things, it was required to make findings that the four elements of a case of wage discrimination had been established. The Tribunal described all four of these elements in its reasons. The third element is a finding that a comparison of the work of the two groups reveals that they were performing work of equal value. Again, it is not disputed that PSAC must establish this on a balance of probabilities.

[6] There are three steps that the Tribunal is required to take when determining whether each necessary element is satisfied. In the first step, the Tribunal must determine whether evidence relating to that element is admissible. In the second step, the Tribunal must determine the weight that should be given to that admissible evidence. This turns on the reliability of the admissible evidence. Finally, in the third step, the Tribunal must determine if that admissible evidence, taking into account its reliability, establishes the element on the appropriate standard of proof. The Tribunal erred in this case by failing to determine if the admissible evidence, taking into account its reliability, established the third element on a balance of probabilities. The Tribunal prematurely concluded its analysis of the third element at the second step after considering admissibility and weight.

[7] Instead of considering whether the third element was satisfied on a balance of probabilities, the Tribunal purported to apply the balance of probabilities standard in deciding that the job information pertaining to the work being compared, an essential component of the work of equal value requirement, was [at paragraph 700] “reasonably reliable, albeit at the ‘lower reasonably reliable’ sub-band level.” The Tribunal equated this to a 50 percent level of certainty. Even if the language used by the Tribunal could somehow be construed as being a finding of work of equal value, which we do not accept, the fact that the Tribunal used a 50 percent level of certainty means that whatever their conclusion was, it was something less than a balance of probabilities, which requires proof in excess of 50 percent.

[8] This is not a case about fundamental jurisprudential pay equity concepts. Rather it is a case in which a tribunal has made a reviewable error by awarding damages without establishing liability. Specifically, liability was not established because the Tribunal, after stipulating that four elements were required to find a case of wage discrimination, only proceeded to find three of those elements.

[9] We further note that the record and reasons of the Tribunal in this case are not adequate to permit an appellate court to properly resolve fundamental jurisprudential pay equity concepts. In addition to not making a finding on the third element, the Tribunal also erred by failing to define what work of equal value is and how the concept applied in this case. Such an explanation is necessary to arrive at a finding of liability and damages. An appellate court should not be put in the position of determining the definition of “work of equal value” and its application to the present case when the decision makers below have not addressed the matter. Hence, the record in this case does not lend itself to the making of authoritative statements on these concepts.

[10] Since the Tribunal has failed to make a finding on the third element of a case of wage discrimination and because both the Tribunal and this Court believe such a finding to be absolutely necessary in order to uphold the complaint, the appeals should be dismissed.

II. PROCEDURAL HISTORY

[11] The three appeals (A-129-08, A-130-08 and A-139-08) before the Court relate to two applications for judicial review of a decision (the Tribunal decision, 2005 CHRT 39) of the Tribunal finding that the respondent, CPC, had engaged in a discriminatory practice, as defined by section 11 of the Canadian Human Rights Act, R.S.C., 1985, c. H-6 (the Act), by paying employees in the male-dominated Postal Operations (PO) group more than em- ployees in the female-dominated Clerical and Regulatory (CR) group. The two applications for judicial review were heard together by Justice Kelen (the applications Judge) in the Court below (Canada Post Corp. v. Public Service Alliance of Canada, 2008 FC 223, [2008] 4 F.C.R. 648).

[12] In A-129-08 and A-139-08, PSAC and the Canadian Human Rights Commission (CHRC) appeal the applications Judge’s decision (the Federal Court decision) that the Tribunal decision be set aside. In A-130-08, PSAC appeals the applications Judge’s dismissal of PSAC’s application for judicial review of the portion of the Tribunal decision that reduced the damages awarded against CPC by 50 percent.

[13] The three appeals were heard together by this Court. These reasons will apply to each of the appeals. A copy of these reasons will be filed as reasons for judgment in the Court file for each of the appeals.

III. RELEVANT STATUTORY PROVISIONS

[14] The statutory provisions that are relevant to the appeals are section 11 and subsections 27(2) [as am. by S.C. 1998, c. 9, s. 20], 49(1) [as am. idem] and 50(3) [as am. idem] of the Act, paragraphs 18.1(3)(b) [as enacted by S.C. 1990, c. 8, s. 5; 2002, c. 8, s. 27] and 52(b) [as am. idem, s. 50] of the Federal Courts Act, R.S.C., 1985, c. F-7 [s. 1 (as am. idem, s. 14)], sections 12 to 15 of the Equal Wages Guidelines, 1986, SOR/86-1082 (the 1986 Guidelines). These provisions are reproduced in the Appendix to these reasons.

IV. BACKGROUND

A. Investigation of the complaint

[15] On August 24, 1983, PSAC filed a complaint with the CHRC which reads as follows:

It is alleged that the Canada Post Corporation as Employer, has violated Section 11 of the Canadian Human Rights Act by paying employees in the male-dominated Postal Operations Group more than employees in the female-dominated Clerical and Regulatory Group for work of equal value. The wage rates of the male-dominated Postal Operations Group exceed those of the female-dominated Clerical and Regulatory Group by as much as 58.9 percent for work of equal value. It is alleged that sex composition of the two groups has resulted in wage discrimination against the Clerical and Regulatory Group, contrary to Section 11.

Corrective Action:

1. That all employees within the CR Group employed by Canada Post Corporation receive wages, as defined in paragraph 11(6) [now section 11(7)] of the Canadian Human Rights Act, equal to the wages of employees within the PO Group performing work of equal value.

2.  That this corrective action be made retroactive to October 16, 1981.

[16] The CR group is made up of clerical and regulatory workers. Typical position titles for workers in the CR group include benefits clerk, accounting clerk and accounts payable clerk. The PO group consists of workers who sort and deliver mail. Typical position titles for workers in the PO group include letter carrier, mail handler and manual sortation clerk.

[17] In essence, the complaint alleges that the employees in the male-dominated PO group were paid higher wages than the employees in the female-dominated CR group who were performing work of equal value to that which was performed by PO group employees. The determination of the equivalence of the value of the work that was performed by both groups requires an assessment of the value of that work having regard to the composite of skill, effort, responsibility and working conditions applicable to that work. Job evaluation is the field of expertise that deals with these types of assessments, which are called evaluations. The process of making these assessments is known as evaluating. Job evaluations are the product of a process in which a methodology, often called a plan, is applied to information about the content of jobs being evaluated.

[18] Prior to the filing of the complaint, PSAC and CPC had been working together with respect to the development of a job evaluation system (System One) that was intended to permit an evaluation of the jobs of all the CPC employees represented by PSAC. Throughout 1984 and part of 1985, CHRC awaited the outcome of those efforts in the hope that System One could be of use with respect to the complaint.

[19] Because of delays with respect to the development of System One, beginning in October 1985 the CHRC pursued its investigation of the complaint more actively. To this end, the CHRC developed a questionnaire (the Job Fact Sheet) to gather data from CR group and PO group employees about the skill, effort, responsibility and working conditions applicable to their jobs.

[20] In the summer of 1986, the job fact sheets were given to somewhere between 246 and 355 CR group employees. The CHRC received 194 completed and usable job fact sheets. To clarify responses to the job fact sheets, the CHRC developed an interview guide (the Interview Guide). Follow-up interviews with the Job Fact Sheet respondents were completed by December 1986 in accordance with the Interview Guide. From April to September 1987, the CHRC evaluated the sample of the 194 CR group employees using System One based on the information in the job fact sheets and the follow-up interviews, notwithstanding that it had not been completed and PSAC advised against its use. Ultimately, these evaluations were not used in the final investigative process.

[21] Contrary to its original intention, the CHRC did not use the job fact sheets and the Interview Guide to collect data from the PO group employees similar to that obtained from the CR group employees. This was partially because CPC questioned the CHRC’s proposed sample size and refused to permit PO group employees to complete the job fact sheets during normal working hours. Additionally, the PO group employees were represented by a union other than PSAC, the Canadian Union of Postal Workers (CUPW), and CUPW refused the CHRC’s request that the PO group employees fill-out the job fact sheets outside normal working hours.

[22] To compensate for the lack of actual information with respect to the work performed by the PO group employees, from July to October 1991, the CHRC created 10 generic job specifications for PO group employees, using information provided by CPC in 1990 and 1991. While the CR group sample included supervisors at the CR-5 level, the generic job specifications did not include the PO supervisors sub-group (PO-SUP) because the CHRC decided that it would be too onerous to fit the wide range of tasks performed by PO-SUP employees into the generic job specifications.

[23] In July 1991, using the 194 CR group responses, the CR group interviews and the PO group generic job specifications, the CHRC began an evaluation using an off-the-shelf plan, the XYZ Hay Plan, for evaluating jobs for the purpose of a pay equity analysis. This plan was selected, at least in part, because System One could only be used to evaluate positions held by employees represented by PSAC, and a number of employees in the PO group were represented by CUPW. As a result, the evaluation generated using this plan did not rely upon the earlier evaluation of the CHRC that utilized System One. In order to make the evaluation more manageable, the CHRC reduced the CR group sample to 93 employees in September 1991.

[24] The CHRC’s evaluation (the CHRC evaluation) was completed in November 1991. This evaluation formed the basis of the CHRC’s Final Investigation Report (the Report), dated January 24, 1992, which concluded that there was a wage difference when comparing the wages and job evaluations of the CR and PO groups, as alleged in the complaint. After considering the Report, the commissioners of the CHRC referred the complaint to the Tribunal for an inquiry on March 16, 1992, pursuant to subsection 49(1) of the Act.

B. The Tribunal inquiry

[25] The Tribunal panel was struck on May 11, 1992 and hearings commenced on November 25, 1992. Written and oral submissions were completed on August 27, 2003. The Chair of the Tribunal retired in June 2004. Additional written submissions were made in August 2004. The Tribunal decision was released on October 7, 2005, over two years after the conclusion of the hearing.

[26] After the Tribunal had begun hearing evidence, it became apparent that there were serious deficiencies in the CHRC evaluation. As a result, PSAC engaged three professional job evaluators (the professional team), Dr. Bernard Ingster, Dr. Martin G. Wolf and Ms. Judith Davidson-Palmer. Dr. Wolf was the spokesperson of the group and the Tribunal qualified him as an expert in Hay-based job evaluation and Hay-based compensation. PSAC asked the professional team to review the CHRC evaluation and to undertake independent evaluations. Ultimately, both the CHRC and PSAC relied exclusively on the professional team’s evaluations to substantiate the complaint.

[27] In May and June of 1993, the professional team conducted its initial evaluation (the Phase 1 evaluation). To conduct this evaluation, the professional team supplemented the information used in the CHRC evaluation with information from its own interviews with CR group employees that it conducted in May 1993. Where a respondent could not be reached, the professional team tried to interview a stand-in. It is not clear how many of the professional team’s interviews were conducted with stand-ins, but the Tribunal found that of a total of 93 possible telephone interviews, 59 were completed.

[28] In September 1994, the professional team attempted to conduct interviews for 97 of the CR group positions that were omitted from the CHRC evaluation and 55 of these interviews were completed. The information obtained in these interviews, combined with information gathered by the CHRC, was used by the professional team in November and December 1994 to evaluate the 97 CR group positions that were omitted from the CHRC evaluation. This second evaluation was called the Phase 2 evaluation.

[29] Based on its Phase 1 and Phase 2 evaluations, members of the professional team prepared two reports. The first report (the Professional Value Report), dated January 1995, was prepared by Dr. Wolf in consultation with Dr. Ingster and Ms. Davidson-Palmer. The Professional Value Report concludes at page 6:

Based on the findings of the total evaluation process in Phases One and Two, the consultants concluded that the rigorous application of the Hay Guide Chart-Profile Method of job evaluation produced substantial evidence that 122 of the 194 incumbents (62.9%) holding CR positions included in this study were in jobs with content greater than one or more of the ten PO jobs covered by these analyses.

In light of further information provided by CPC, the professional team later revised the 62.9 percent figure to 34.2 percent.

[30] The second report (the Professional Wage Gap Report), dated February 1995, was prepared by Dr. Wolf alone. The Professional Wage Gap Report calculates the relationships between the hourly rates of pay of PO group jobs and the value of PO group jobs, as determined by the professional team’s evaluations for 1983, 1989 and 1995, using several different approaches. The Professional Wage Gap Report concludes that, under any of its approaches, there is “a significant gap between the wages paid to CR’s and to PO’s performing work of equal value”.

[31] To support its position that the methods of the CHRC and the professional team were insufficient to substantiate the complaint, CPC called three expert witnesses, Ms. Nadine Winter, Mr. Norman Willis and Mr. P. G. Wallace. The three experts’ critiques of the professional team’s method highlighted the failure of the professional team to follow the industry standard application of the Hay method and the inapplicability of the Hay method to clerical and blue collar positions.

V. Decisions Below

A. Canadian Human Rights Tribunal

[32] In its reasons, the Tribunal addressed the following four fundamental issues:

(a) Does the ability of the CHRC to issue Equal Wage Guidelines that are binding on the Tribunal create a reasonable apprehension of bias?

(b) Can the 1986 Guidelines be applied to the complaint, even though it was filed in 1983?

(c) Can factors other than those identified in the 1986 Guidelines be used to rebut the presumption that when men and women are paid different wages for work of equal value, that difference is based on sex?

(d) Has the complainant established a prima facie case of discrimination under section 11 of the Act on a balance of probabilities?

[33] In view of our disposition of this appeal, it will only be necessary to consider the Tribunal’s analysis with respect to the last issue.

[34] In paragraph 254 of its reasons, the Tribunal determined that each element of section 11 of the Act had to be substantiated on a balance of probabilities in order to substantiate the complaint. In assessing the value of the work that is being compared, the Tribunal found, at paragraph 255 of its reasons, that the criterion in subsection 11(2) of the Act—the composite of the skill, effort and responsibility required in the performance of the work and the conditions under which the work is performed—was required to be used. The Tribunal went on to state, at paragraph 256 of its reasons, that discrimination based on sex will be presumed when a difference in wages has been found to exist between male and female employees, employed in the same establishment, performing work of equal value. At paragraph 257 of its reasons, the Tribunal set forth its determination of the four elements that were required to be proven, on a balance of probabilities, to establish a prima facie case of discrimination as alleged in the complaint:

(1) The complainant occupational group is predominantly of one sex and the comparator occupational group is predominantly of the other sex. In this Complaint, that means the complainant CR’s must be predominantly female and the comparator PO’s must be predominantly male.

(2) The female-dominated occupational group and the male-dominated occupational group being compared are composed of employees who are employed in the same establishment.

(3) The value of the work being compared between the two occupational groups has been assessed reliably on the basis of the composite of the skill, effort, and responsibility required in the performance of the work, and the conditions under which the work is performed. The resulting assessment establishes that the work being compared is of equal value.

(4) A comparison made of the wages being paid to the employees of the two occupational groups for work of equal value demonstrates that there is a difference in wages between the two, the predominantly female occupational group being paid a lesser wage than the predominantly male occupational group. This wage difference is commonly called a “wage gap”.

In these reasons, these four elements are referred to as “element one”, “element two”, “element three” and “element four” respectively.

(a) Element one—The comparator group

[35] Relying on sections 12 and 13 of the 1986 Guidelines, the Tribunal held that the PO group was a male-dominated occupational group and the CR group was a female-dominated occupational group. The Tribunal found that in 1983, just over 80 percent of the 2 316 employees in the CR group were female. At the same time, just over 75 percent of the 50 912 employees in the PO group were male. In 1992, the time of the referral of the complaint to the Tribunal, the CR group was over 83 percent female and the PO group was over 71 percent male.

[36] CPC challenged the selection of the PO group as a comparator on the grounds that the PO group employees should not be viewed as a single group. They opposed the use of a comparator group hand-picked by PSAC and featuring the highest paid group of women working for CPC. Instead, CPC suggested that the PO-4 level of the PO group should be used as the comparator since it was the most representative of the PO group. Since the PO-4 level was 53 percent male and 47 percent female in 1983, it was not male dominated under the 1986 Guidelines and hence CPC argued that the first element of the prima facie case of discrimination was not established. Furthermore, CPC argued that PSAC selected the PO group as a comparator because it was highly paid and that this was inappropriate “cherry picking”.

[37] The Tribunal rejected the argument that the PO group should not be viewed as a single group, because the federal government job classification inherited by CPC from the Post Office Department of the Government of Canada is important in the designation of an “occupational group” under the 1986 Guidelines. The Tribunal also rejected CPC’s suggestion that the complainant was “cherry picking” the comparator group. It noted that the PO group represented approximately 80 percent of the CPC workforce and that by virtue of its size, its selection therefore could not have constituted “cherry picking”. Additionally, the only other possibilities, the General Labour and Trades and General Services groups, represented only a small percentage of CPC employees, and the Tribunal found there was no evidence that their work was at all similar to that performed by employees in the CR group.

[38] At paragraph 283 of the Tribunal decision, the Tribunal stated its conclusion with respect to this element of subsection 11(1) of the Act, as follows:

Accordingly, the Tribunal finds that the complainant, a predominantly female occupational group, and the comparator, a predominantly male occupational group, are appropriately designated under section 11 of the Act and the 1986 Guidelines as representative groups for comparison of work generally performed by women and work generally performed by men. Therefore, the first element necessary to the establishment of a prima facie case under section 11 of the Act has been met. [Emphasis added.]

(b) Element two—Employment in the same establishment

[39] The Tribunal next considered whether the CR and PO groups were both employed in the same establishment and in particular, whether a geographical or functional definition of establishment was applicable. In the Tribunal’s view, employees are in the same geographical establishment where they work in the same building, municipality or district. In contrast, employees are in the same functional establishment where they are subject to a common set of personnel and wage policies. Relying on this Court’s decision in Canada (Human Rights Commission) v. Canadian Airlines International Ltd., 2004 FCA 113, [2004] 3 F.C.R. 663, the Tribunal eschewed a geographical definition of establishment and adopted a functional one.

[40] The Tribunal then determined that the evidence before it demonstrated that the CPC was a well integrated business with considerable corporate level policy direction, leading it to reach its conclusion with respect to this element. Specifically, at paragraphs 353 and 354 respectively of its reasons, the Tribunal stated:

Therefore, the Tribunal finds that all employees of Canada Post have been, as applicable, subject to the various common corporate policy directives issued by the Corporation, including those respecting personnel and wage policies. As a result, the Tribunal finds that, for the purposes of section 11 of the Act, the employee groups representing the complainant and the comparator are employed in the same establishment.

Accordingly, the second element necessary to the establishment of a prima facie case under section 11 of the Act has been met. [Emphasis added.]

(c) Element three—Work of equal value

[41] The Tribunal framed the question with respect to this element as whether the comparison of the work of the complainant group and the comparator group establish that the work being compared is equal in value. Further, the Tribunal stated, at paragraph 355 of its reasons:

To be able to come to a reasonable conclusion concerning the value of the work performed by the complainant and the comparator occupational groups, the evaluation process as a whole must be reliable, on a balance of probabilities.

[42] The Tribunal accepted the importance of undertaking job evaluations with reliable job information and with a reliable job evaluation plan. At paragraph 358 of its reasons, the Tribunal reproduced a portion of a booklet, entitled “Implementing Pay Equity in the Federal Jurisdiction”, that was put into evidence by the CHRC. The Tribunal accepted the booklet as a general guide with respect to the collection and processing of information that should, given an acceptable job evaluation plan and competent evaluators, result in the determination of reliable values of work being assessed and compared. The first sentence of the booklet that was reproduced by the Tribunal, in paragraph 358 of its reasons, reads as follows:

Job evaluation plans are the key to determining what constitutes “work of equal value”.

Later, the following sentence appears:

Because pay equity is premised on the assumption that the worth of different positions across an organization should be compared, use of a single plan to evaluate all jobs is essential.

[43] The Tribunal also noted that the booklet was not developed for use in a litigious context.

[44] The Tribunal then specified, at paragraph 362 of its reasons, the issues that it intended to address with respect to this element of the requirements of subsection 11(1) of the Act:

Consequently, the issues which will be addressed are as follows:

1.  What job evaluation system, or plan, was used to undertake the evaluation of the CR and PO jobs/positions, and how reliable was it?

2.  What process was used and how reliable was it in analyzing the collected job data/information for purposes of assigning values to the CR and PO jobs/positions considered?

3.  What job data/information was collected, and from what sources, and how reliable was it?

4.  What were the resulting values attributed to the various CR and PO jobs/positions, and how reliable were they?

[45] The Tribunal turned its mind to the basis upon which it was required to approach the resolution of these issues, stating, at paragraph 410 of its reasons, that there is support for “a flexible case-by-case approach to the determination of how the concept of equal pay for work of equal value is to be effected”.

[46] At paragraph 411 of its reasons, the Tribunal quoted from the decision of Justice Hugessen in Public Service Alliance of Canada v. Canada (Department of National Defence), [1996] 3 F.C. 789 (C.A.), which places the burden of proof in pay equity disputes at the ordinary civil burden of a balance of probabilities. At paragraph 412, the Tribunal framed the issue before it:

These rulings support a call for a standard of reasonableness, there being no such thing as absolute reliability. The application of such a standard will depend very much on the context of the situation under examination. The issue is, then, given all the circumstances of the case before this Tribunal, is it more likely than not that the job information, from its various sources, the evaluation system and the process employed, and the resulting evaluations are, despite any weaknesses, sufficiently adequate to enable a fair and reasonable conclusion to be reached, as to whether or not, under section 11 of the Act, there were differences in wages for work of equal value, between the complainant and comparator employees concerned?

[47] In this paragraph, which is in the portion of the Tribunal’s reasons dealing with the question of whether the work being compared is equal in value, the Tribunal frames the issue as whether four things, namely, the job information, the evaluation system, the process employed and the resulting evaluations are sufficiently adequate to permit a conclusion to be reached with respect to whether or not there are differences in wages for work of equal value. In particular, the Tribunal focuses on whether these four things are reasonably reliable.

[48] At paragraph 555 of its reasons, the Tribunal reaffirms its focus on “reasonable reliability”, stating:

Each of the elements necessary in testing reasonable reliability should be examined. In other words, the job evaluation system chosen should be reasonably reliable, the process and methodology used in evaluating the relevant jobs/positions should be reasonably reliable, and the job information and its sources should be reasonably reliable. The findings of the Tribunal should be based on the civil standard of a balance of probabilities.

It is of note that the Tribunal did not refer to the fourth item that it referred to in paragraph 412 of its reasons, namely the “resulting evaluations”.

[49] The Tribunal went on to determine that reasonable reliability was present on a balance of probabilities.

[50] With respect to the job evaluation system and the process, the Tribunal, at paragraphs 571 and 593 respectively, stated:

Therefore, the Tribunal finds that, on a balance of probabilities, the Hay Plan, whether using the factor comparison method or other approaches, is, in the hands of competent evaluators as were the members of the Professional Team, a suitable overall job evaluation scheme which will address the issues of this “pay equity” Complaint in a reasonably reliable manner.

…

Therefore, the Tribunal finds that it is more likely than not that the evaluation process which the Professional Team used in its work was reasonably reliable.

[51] The Tribunal acknowledged that the determination of reasonable reliability with respect to the matter of job information was a daunting task and in paragraph 673, framed the question as follows:

But, given the somewhat painful and prolonged circumstances of the case before this Tribunal, was the job information “good enough”, on a balance of probabilities, to generate reasonably reliable job/position values that, in turn, could be used to demonstrate whether there was a wage gap?

[52] To assist in its determination of whether the job information used by the professional team was reasonably reliable, at paragraph 679 of its reasons, the Tribunal referred to a passage from S. M. Waddams, The Law of Damages, loose-leaf ed. (Toronto: Canada Law Book Inc., 2004) at page 13-1:

In Anglo-Canadian law ... the courts have consistently held that if the plaintiff establishes that a loss has probably been suffered, the difficulty of determining the amount of it can never excuse the wrongdoer from paying damages. If the amount is difficult to estimate, the tribunal must simply do its best on the material available, though of course if the plaintiff has not adduced evidence that might have been expected to be adduced if the claim were sound, the omission will tell against the plaintiff. In Ratcliffe v. Evans, Bower L.J. said:

As much certainty and particularity must be insisted on, both in pleading and proof of damage, as is reasonable, having regard to the circumstances and to the nature of the acts themselves by which the damage is done. To insist upon less would be to relax old and intelligible principles. To insist upon more would be the vainest pedantry.

[53] Inspired by this idea of making the most of the evidence before the decision maker, the Tribunal arrived at a “spectrum” of reasonable reliability, with one end of the spectrum being very reliable and the other end being minimally reliable. Using this spectrum, at paragraph 683 of its reasons, the Tribunal asked “[w]hile the job information may not meet the degree of reliability that should normally be sought for a ‘pay equity’ situation, is it ‘adequate’… for this specific situation?” The Tribunal then analysed the information before it and concluded at paragraph 689:

The Tribunal must confess that navigating the job information through the straits of “reasonable reliability” has not been a relaxing passage. Yet, balancing the evidence presented by all parties and expert witnesses, and under the unique circumstances of this case in the realm of proscribed discrimination human rights legislation ... the job information, in the hands of the Professional Team, was more likely than not, “reasonably reliable”, or “adequate” as that Team described it, despite certain imperfections.

[54] Though it already appeared to have ruled that the claimants had met the burden of establishing reliability, the Tribunal then further elaborated on the meaning of “reasonable reliability”. At paragraph 693 of its reasons, it defined “reasonable reliability” as “information that is consistently, moderately dependable or in which moderate confidence can be put.” The Tribunal then stated that reliability generally should be viewed as a band, with no one fixed point always considered “reasonable”. Rather, it posited three sub-bands of reasonable reliability: “upper reasonable reliability,” “mid reasonable reliability” and “lower reasonable reliability” (Tribunal decision, at paragraph 696). According to the Tribunal, all three of these sub-bands meet the standard of “reasonable reliability,” but the upper sub-band is preferred.

[55] Working with these sub-bands, the Tribunal characterized the evidence before it as falling in the lower reasonable reliability sub-band. Thus, the Tribunal was able to reiterate the conclusion that it came to in paragraph 689 of its reasons, stating at paragraph 700 of its reasons:

Hence, … it was more likely than not that the job information utilized by the Professional Team in conducting its job evaluations of the CR and PO positions/jobs pertinent to this case, was reasonably reliable, albeit at the “lower reasonably reliable’ sub-band level”.

[56] The Tribunal’s reasons then progressed under a new heading, “VII. WAGE GAP AND WAGE ADJUSTMENT METHODOLOGY”. At paragraph 701 of its reasons, the Tribunal stated the next questions that it intended to address:

Having found that it is more likely than not, that the “off-the-shelf” Hay Plan being used in the traditional factor comparison methodology, the process followed and the job information utilized by the Professional Team in conducting its CR and PO positions/jobs evaluations were reasonably reliable, the next questions to be addressed are:

How reliable were the resulting job evaluation values attributed by the Professional Team to the CR positions and PO jobs concerned?

Was a “wage gap” demonstrated between the female and male predominant groups performing work of equal value?

[57] Given the first question posed by the Tribunal, it is apparent that, to this point in its reasons, the Tribunal had not reached a conclusion with respect to the third element referred to in paragraph 257 of its reasons, as it did in paragraphs 283 and 354 of its reasons, in relation to the first two elements referred to in paragraph 257 of its reasons.

[58] Stating only that the credibility of the professional team had been established and that Dr. Wolf was an expert in relation to the Hay Plan, the Tribunal, at paragraph 703 of its reasons, reached the following conclusion:

Accordingly, the Tribunal concludes that it is more likely than not that the aforementioned reasonably reliable Hay Plan, process and job information, in the hands of competent evaluators, as were the Professional Team, would result in reasonably reliable job evaluation values being attributed to the work performed by CR and PO employees.

[59] This conclusion was reiterated by the Tribunal at paragraph 798 of its reasons, in virtually identical language.

(d) Element four—Wage gap

[60] The Tribunal determined, in paragraph 801 of its reasons, that the evidence presented to it was sufficient, on a balance of probabilities, to demonstrate a wage gap, thus concluding that the final element required by subsection 11(1) of the Act had been fulfilled.

[61] With respect to the size of the wage gap, the Tribunal accepted the proposal that had been submitted to the CHRC. Then the Tribunal dealt with CPC’s contention that this element could not have been satisfied because of insufficient evidence with respect to the non-wage forms of compensation that were received by the CR and PO groups.

[62] The Tribunal accepted the evidence of Dr. Lee on behalf of the complainants to the effect that the levels of non-wage forms of compensation received by both groups of employees were more or less equivalent. In doing so, the Tribunal noted, at paragraph 926 of its reasons, that Dr. Lee was constrained by his late retention by PSAC, and that “[h]e did the best he could given the situation he faced.”

[63] At paragraph 927 of its reasons, the Tribunal found that Dr. Lee’s report fell in the “lower reasonably reliable” band on its reliability spectrum. As a result, in the Tribunal’s view, PSAC had succeeded in establishing that there was no non-wage compensation that needed to be included in determining whether there was a difference in wages between the CR and PO group employees for the purposes of the analysis under section 11 of the Act.

(e) Remedy

[64] The Tribunal then examined how the remedy ordered should reflect the magnitude of the wage gap. It began by comparing the role of damages in human rights law to the role of damages in tort law, noting that the objective in both is to make the victim whole. In this case, that means “restoring the victim to the position or status he or she would have been in had the substantiated discrimination not occurred” (Tribunal decision at paragraph 934).

[65] Relying on Canada (Attorney General) v. Morgan, [1992] 2 F.C. 401 (C.A.); Chopra v. Department of National Health and Welfare, 2004 CHRT 27, affd sub. Nom. Chopra v. Canada (Attorney General), 2007 FCA 268, [2008] 2 F.C.R. 393 and Singh v. Statistics Canada, [1998] C.H.R.D. No. 7 (QL), the Tribunal held that it could reduce the damages when the magnitude of the damages is uncertain. The Tribunal found that the magnitude of the damages was uncertain in this case because job information and evidence relating to non-wage forms of compensation was only “lower reasonably reliable”.

[66] With this in mind, the Tribunal held that where job information and evidence relating to non-wage forms of compensation is categorized in the “upper reasonable reliability” sub-band, the damages award should reflect 100 percent of the wage gap, where they fall in the “mid reasonable reliability” sub-band, the award should reflect 75 percent of the wage gap, and where they fall in the “lower reasonable reliability” sub-band, the award should be 50 percent or less of the calculated gap. Accordingly, the Tribunal discounted the award to the claimants by 50 percent.

[67] Regarding the time period over which lost wages are to be awarded the Tribunal decided that the compensation period should begin on August 24, 1982, one year before the filing of the complaint and not October 16, 1981, as requested in the complaint. The time period ended on June 2, 2002, when the wage gap was eliminated.

B. Federal Court

[68] The applications Judge heard two applications. In the first, CPC requested judicial review of the decision upholding the complaint. In the second, PSAC requested judicial review of the decision to discount the award of damages by 50 percent. In his reasons, the applications Judge considered five issues:

(1)	whether the Tribunal erred in applying the 1986 Guidelines;
(2)	whether the Tribunal erred in applying an incorrect standard of proof;
(3)	whether the Tribunal erred in determining the comparator group;
(4)	whether the Tribunal erred in holding that once a wage gap has been established, the presumption of discrimination is only rebuttable by factors in the 1986 Guidelines; and
(5)	whether the Tribunal erred in discounting the damages by 50 percent.

(a) Applicability of the 1986 Guidelines

[69] The applications Judge held that the standard of review with respect to the issue of whether the Tribunal erred in retroactively applying the 1986 Guidelines was reasonableness simpliciter. (Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190, had not yet been decided.) In the circumstances, the applications Judge determined that their application by the Tribunal was reasonable and that they were not being applied retroactively.

(b) Standard of proof

[70] The applications Judge also identified reasonableness simpliciter as the standard of review with respect to this issue.

[71] The applications Judge held that the standard of proof that must be met in order to establish a discriminatory practice under subsection 11(1) of the Act is the ordinary civil burden of the balance of probabilities. He found that the Tribunal recognized this as the correct standard of proof but then misapplied that standard.

[72] The applications Judge referred to the four elements that were identified as essential to the establishment of a prima facie case of discrimination under subsection 11(1) of the Act. He stated that the parties before him agreed that the issue of whether the Tribunal applied the proper standard of review focused on the third element. At paragraph 122 of his reasons, the applications Judge further defined his focus, stating:

At the hearing, the parties identified three material facts for the evaluation of the work being compared:

(1) the reliability of the job information from the occupational groups being compared, including the sources from which the job information was collected;

(2) the reliability of the evaluation methodology utilized to undertake the evaluations; and

(3) the reliability of the actual evaluation process undertaken.

After carefully considering the submissions of the parties with respect to these three material facts, the Court will concentrate its standard of proof analysis on the first material fact.

[73] The applications Judge referred to the evidence with respect to the issue of the job information and the Tribunal’s findings with respect to that issue. At paragraphs 131 and 132 he stated:

At paragraph 673, the Tribunal held that there is little doubt the job information used in conducting the evaluations “did not meet the standard that one would normally expect from a joint employer-employee ‘pay equity’ study.” Having said that, the Tribunal continued, asking:

… was the job information “good enough”, on a balance of probabilities, to generate reasonably reliable job/position values that, in turn, could be used to demonstrate whether or not there was a wage gap?

At this point, the Court notes that the Tribunal appears to be about to apply the balance of probabilities as the standard of proof required to establish the essential element of work of equal value. [Emphasis added.]

[74] The applications Judge found that the Tribunal then veered off track by adopting the passage from Professor Waddams’ book, reproduced in paragraph 133 of his reasons, with respect to the principle that when assessing damages, a tribunal must do the best it can with the evidence before it. The applications Judge also found that the Tribunal introduced ambiguity into its application of the standard of proof by its reference to the spectrum analysis.

[75] At paragraph 155 of his reasons, the applications Judge reiterates the obligation of the complainant to prove, on a balance of probabilities, that there were differences in wages for work of equal value between the complainant and the comparator groups. He then stated, at paragraphs 156 to 158 respectively of his reasons:

The Tribunal erred in law in applying a confusing, invented, and novel standard of proof with respect to the reliability of the job information in order to find liability. The Tribunal’s finding that the job information evidence was “reasonably reliable” at the “lower-reasonably reliable sub-band” level is less than a finding that the job information was reliable on the balance of probabilities.

The Court’s conclusion that the Tribunal did not find that the job information was reliable on the balance of probabilities is indirectly confirmed by the Tribunals decision to discount the damages by 50 percent. The Tribunal decided to reduce the damages by 50 percent because the “job information” used to determine the wage gap and the non-wage compensation only met the “lower reasonable reliability” standard on the spectrum of reliability. The Tribunal held, at paragraphs 948-949:

Following the spectrum analysis already completed for the two elements of uncertainty, the Tribunal concludes that a wage gap determination based upon “upper reasonable reliability” evidence should, logically, give rise to a 100% award of lost wages, a determination based upon “mid reasonable reliability” to a 75% award, and a determination based upon “lower reasonable reliability” to an award of 50% or less.

Accordingly, the Tribunal concludes that the finally determined award of lost wages for each eligible CR employee, by whatever methodology, should be discounted by 50% in line with the lower reasonable reliability status of the relevant job information and non-wage forms of compensation.

This finding demonstrates that the Tribunal was so unsure about the reliability of the job information evidence that it only awarded the complainant 50 percent of its damages. In law, the Tribunal cannot decide to award the complainant only 50 percent of its damages where it is unconvinced that the evidence regarding liability was probably reliable. A party cannot be half liable — half liable means that the evidence is less than probable. By reducing the damage award by 50 percent, the Tribunal indirectly confirms that it does not think that the evidence was reliable on the balance of probabilities. At the end of the hearing, if the evidence on liability is evenly balanced, the balance of probabilities has not been tilted in favour of the complainant, and the complaint must be dismissed.

[76] The applications Judge rejected the arguments of PSAC and the CHRC that the Tribunal reached a conclusion that the balance of probabilities threshold had been met with respect to the work of equal value element. In that regard, the applications Judge rejected the assertion that the conclusion stated in paragraph 801 of the Tribunal’s reasons was sufficient to cover this point. At paragraph 160 of his reasons, he refers to the conclusion of the Tribunal at paragraph 703 of its reasons, stating:

For instance, at paragraph 703, the Tribunal identifies the issue before it — i.e., that the material facts are “reasonably reliable”:

Accordingly, the Tribunal concludes that it is more likely than not that the aforementioned reasonably reliable Hay Plan, process and job information, in the hands of competent evaluators, as were the Professional Team, would result in reasonably reliable job evaluation values being attributed to the work performed by CR and PO employees.

In concluding that the material facts must create “reasonably reliable” job values, the Tribunal applies a standard of proof less than reliable on the balance of probabilities. [Emphasis added.]

[77] Finally, the applications Judge noted that the standard of proof that was actually applied by the Tribunal was more akin to a “reasonable basis in the evidence” standard, which is lower than the required balance of probabilities standard.

(c) Comparator Group

[78] The applications Judge then considered whether the Tribunal erred in finding that the PO group was an appropriate comparator group. His conclusion on this issue is set forth in paragraph 207 of his reasons:

While the Tribunal analyzed the evidence about the appropriateness of the PO group as a comparator group, the Court finds the Tribunal unreasonably ignored the factual reality that the largest group of women at Canada Post were the 10 000 women working as “mail sorters” within the PO group, and that these 10 000 women were the best paid unionized employees at Canada Post. The Court finds it unreasonable to choose a comparator group that masked the 10 000 women and, in fact, considered them men for the purposes of section 11. This is contrary to the intent of section 11 and is illogical. Moreover, it is evident that there was no systemic wage discrimination against female employees at Canada Post since the largest group of women within Canada Post were the highest paid of all unionized employees. [Emphasis in original.]

(d) Presumption

[79] The applications Judge subsequently turned to the question of whether the Tribunal erred in holding that only factors in the 1986 Guidelines can be used to rebut the presumption of discrimination based on sex. He concluded that having regard to his conclusions with respect to the issues of the standard of proof and the appropriateness of the comparator group, the issue of the legal presumption did not arise.

(e) Damages

[80] The applications Judge determined that no damages should have been awarded since the complaint had not been established on a balance of probabilities. Finally, the applications Judge lamented the long duration of the proceedings.

(f) Disposition

[81] The applications Judge ordered the complaint to be sent back to the Tribunal with the direction that the complaint be dismissed as not substantiated according to the legal standard of proof.

VI. ISSUES

[82] In A-129-08 and A-139-08, PSAC and the CHRC appeal the applications Judge’s decision to set aside the Tribunal’s decision. In A-130-08, PSAC appeals the applications Judge’s dismissal of PSAC’s application for judicial review of the portion of the decision that reduced the damages awarded against CPC by 50 percent.

[83] The following issues arise in these appeals:

(a) whether the Tribunal failed to make a finding that the third element of a prima facie case of wage discrimination had been established and if so, whether such a failure would vitiate the Tribunal’s decision that such a case of wage discrimination had been made out against CPC;

(b) whether the applications Judge erred in concluding that the Tribunal failed to apply the correct standard of proof with respect to its findings in relation to the elements required to establish a prima facie case of discrimination;

(c) whether the applications Judge erred by showing insufficient deference to the Tribunal when determining whether the PO group was an appropriate comparator group;

(d) whether the applications Judge erred by referring the complaint back to the Tribunal with the direction that it be dismissed; and

(e) whether the Tribunal erred by discounting damages.

We are of the view that this appeal can be disposed of by reference to the first two of these issues. For the reasons that follow, we are in agreement with the disposition of the applications Judge.

VII. ANALYSIS

A. The role of the Court in this appeal

[84] It is now settled that when this Court hears an appeal from a decision of the Federal Court in an application for judicial review of a decision of an administrative tribunal, this Court’s task is to determine whether the reviewing judge correctly identified the standard of review and applied it correctly in reviewing the tribunal’s decision. (See Telfer v. Canada (Revenue Agency), 2009 FCA 23, [2009] 4 C.T.C. 123.) As stated by Rothstein J.A. (as he then was), in Prairie Acid Rain Coalition v. Canada (Minister of Fisheries and Oceans), 2006 FCA 31, [2006] 3 F.C.R. 610, at paragraph 14, “[i]n practical terms, this means that the appellate court itself reviews the tribunal decision on the correct standard of review.” Consequently, we will now review the Tribunal’s decision on the two issues identified above.

B. Whether the Tribunal erred by failing to make a necessary finding

[85] In the hearings of these appeals, the Court requested the parties to make submissions with respect to whether the Tribunal made a finding that the third element of a prima facie case of wage discrimination had been established. It was common ground among the parties, and we agree, that such a failure would be sufficient to vitiate the Tribunal’s decision. Since the applications Judge did not analyse this issue, we will undertake this analysis.

[86] For the reasons that follow, we conclude that the Tribunal failed to make the requisite finding. Specifically, we are of the view that the Tribunal terminated its analysis of the third element after considering the reliability of the evidence related to that element and failed to conclude that such element was established on a balance of probabilities.

(a) Standard of review

[87] The question of whether the failure to make a finding on each of the four elements of a prima facie case of wage discrimination would vitiate a conclusion that such a case has been made out against CPC would, in all likelihood, be a pure legal question that would be reviewed on the standard of correctness. However, that question is not in issue. Instead, the question is whether, upon a fair reading of the Tribunal’s reasons, it can be said that the Tribunal actually failed to make the requisite finding with respect to the third element. Consideration of this question requires the interpretation of certain general legal principles and the applicable provisions of the Act and the 1986 Guidelines, as well as their application to the facts as found by the Tribunal. Thus, the question may be regarded as one of mixed fact and law. As such, it is to be reviewed on the standard of reasonableness, unless the legal component is readily extricable from the factual component. It may well be possible to extricate the legal component with the result that the standard of review of the extricated question would be correctness. However, in the circumstances, we decline to undertake the task of trying to extricate a discrete legal component from the question, with the result that the standard of reasonableness will be applied.

(b) The general approach to analysing elements of a case of wage discrimination

[88] The complainant alleges that CPC violated section 11 of the Act by paying employees in the male-dominated PO group more than employees in the female- dominated CR group for work of equal value.

[89] The basis for the establishment that such a violation has occurred has been well summarized in Ontario Human Rights Commission and O’Malley v. Simpsons-Sears Ltd. et al., [1985] 2 S.C.R. 536, a case in which the Supreme Court of Canada found that a rule requiring all employees to work on Saturdays could be discriminatory under the Ontario Human Rights Code, R.S.O. 1980, c. 340, even if there was no discriminatory intention on the part of the employer. At page 558, McIntyre J. stated:

To begin with, experience has shown that in the resolution of disputes by the employment of the judicial process, the assignment of a burden of proof to one party or the other is an essential element. The burden need not in all cases be heavy—it will vary with particular cases—and it may not apply to one party on all issues in the case; it may shift from one to the other. But as a practical expedient it has been found necessary, in order to insure a clear result in any judicial proceeding, to have available as a ‘tie-breaker’ the concept of the onus of proof. I agree then with the Board of Inquiry that each case will come down to a question of proof, and therefore there must be a clearly-recognized and clearly-assigned burden of proof in these cases as in all civil proceedings. To whom, should it be assigned? Following the well-settled rule in civil cases, the plaintiff bears the burden. He who alleges must prove. Therefore, under the Etobicoke rule as to burden of proof, the showing of a prima facie case of discrimination, I see no reason why it should not apply in cases of adverse effect discrimination. The complainant in proceedings before human rights tribunals must show a prima facie case of discrimination. A prima facie case in this context is one which covers the allegations made and which, if they are believed, is complete and sufficient to justify a verdict in the complainant’s favour in the absence of an answer from the respondent-employer. [Emphasis added.]

[90] As previously indicated, in these appeals, it is common ground that in the circumstances of the complaint, the establishment of a prima facie case of wage discrimination requires the complainant to establish four elements on a balance of probabilities. The description of those elements that is contained in paragraph 257 of the Tribunal’s reasons has not been challenged by any party.

[91] With respect to the establishment of those elements, a portion of the reasons of Lord Hoffmann in In re B (Children) (Fc), [2008] UKHL 35, at paragraph 2 (one of the U.K. decisions referred to by Justice Rothstein in F.H. [infra]) is apposite.

If a legal rule requires a fact to be proved (a “fact in issue”), a judge or jury must decide whether or not it happened. There is no room for a finding that it might have happened. The law operates a binary system in which the only values are 0 and 1. The fact either happened or it did not. If the tribunal is left in doubt, the doubt is resolved by a rule that one party or the other carries the burden of proof. If the party who bears the burden of proof fails to discharge it, a value of 0 is returned and the fact is treated as not having happened. If he does discharge it, a value of 1 is returned and the fact is treated as having happened.

[92] In this context, each of the requisite elements can be regarded as a “fact in issue” that must be proved on a balance of probabilities. Where the Tribunal finds that an element has been proved, such a finding would return a value of one in Lord Hoffmann’s binary system. However, where the requisite degree of proof is not present, the “fact in issue” or element has not been proved and, to return to Lord Hoffmann’s binary system, a value of zero would be returned.

[93] In its fact-finding analysis with respect to the four elements or “facts in issue”, the Tribunal was required to take the following three steps (the “steps required to find a ‘fact in issue’”).

[94] In the first step, the Tribunal must determine whether evidence relating to the particular “fact in issue” is admissible. This depends on whether such evidence meets certain rules at common law, or rules developed by the Tribunal.

[95] In the second step, the Tribunal must determine the weight to be given to the admissible evidence with respect to the particular “fact in issue”. At this point, the reliability of that evidence is central to the determination of the weight it should receive.

[96] Finally, in the third step, the Tribunal must determine whether the overall standard of proof has been met with respect to the “fact in issue”. In civil matters, the standard of proof is the balance of probabilities. In our view, this standard has been definitively settled by the Supreme Court of Canada in F.H. v. McDougall, 2008 SCC 53, [2008] 3 S.C.R. 41, wherein, at paragraphs 40 and 49 respectively, Justice Rothstein stated:

Like the House of Lords, I think it is time to say, once and for all in Canada, that there is only one civil standard of proof at common law and that is proof on a balance of probabilities. Of course, context is all important and a judge should not be unmindful, where appropriate, of inherent probabilities or improbabilities or the seriousness of the allegations or consequences. However, these considerations do not change the standard of proof. I am of the respectful opinion that the alternatives I have listed above should be rejected for the reasons that follow.

…

In the result, I would reaffirm that in civil cases there is only one standard of proof and that is proof on a balance of probabilities. In all civil cases, the trial judge must scrutinize the relevant evidence with care to determine whether it is more likely than not that an alleged event occurred.

Justice Rothstein continued at paragraph 54:

Where the trial judge expressly states the correct standard of proof, it will be presumed that it was applied. Where the trial judge does not express a particular standard of proof, it will also be presumed that the correct standard was applied.

[97] Because all four elements are essential to the establishment of a prima facie case of wage discrimination, a value of one must be returned with respect to each of those elements. In other words, if the requisite level or standard of proof is not met with respect to any element, the complaint must be dismissed as not proved.

(c) The Tribunal’s approach to finding elements of a case of wage discrimination

[98] Before addressing the question of whether the Tribunal failed to make the requisite finding with respect to the third element, we wish to consider the Tribunal’s treatment of the first two elements. Given our disposition of these appeals, it is unnecessary for us to give any consideration to the Tribunal’s treatment of the fourth element.

(i) Element one—The comparator group

[99] The Tribunal considered the evidence that was before it in the context of sections 12 and 13 of the 1986 Guidelines and found that the complainant occupational group, the CR group, was a female-dominated group and the comparator occupational group, the PO group, was a male-dominated occupational group for the purposes of those provisions. Having reached this conclusion, the Tribunal went on to conclude, at paragraph 283 of its reasons, that these two groups

… are appropriately designated under section 11 of the Act and the 1986 Guidelines as representative groups for comparison of work generally performed by women and work generally performed by men. Therefore, the first element necessary to the establishment of a prima facie case under section 11 of the Act has been met. [Emphasis added.]

[100] We emphasize this conclusion, without commenting upon its reasonableness, to illustrate that the Tribunal addressed the requirements of this element and found that they had been met. We also note that using Lord Hoffmann’s binary system, this finding would return a value of one in relation to the establishment of this element.

(ii) Element two—Employment in the same establishment

[101] In a manner similar to that in which the Tribunal dealt with element one, the Tribunal ascertained the requirements of this element, considered the evidence that was tendered in relation to it and made a finding that the employee groups representing the complainant and the comparator were employed in the same establishment. Then, at paragraph 354 of its reasons, the Tribunal stated, “Accordingly, the second element necessary to the establishment of a prima facie case under section 11 of the Act has been met” (emphasis added).

[102] Again, we emphasize the clearly stated conclusion of the Tribunal with respect to this element. Also, we reiterate that under Lord Hoffmann’s binary system, this finding would return a value of one in relation to the establishment of this element.

(iii) Element three—Work of equal value

[103] The Tribunal’s description of elements 3 and 4, as stipulated in paragraph 257 of its reasons, is reproduced for convenience:

(3) The value of the work being compared between the two occupational groups has been assessed reliably on the basis of the composite of the skill, effort, and responsibility required in the performance of the work, and the conditions under which the work is performed. The resulting assessment establishes that the work being compared is of equal value.

(4) A comparison made of the wages being paid to the employees of the two occupational groups for work of equal value demonstrates that there is a difference in wages between the two, the predominantly female occupational group being paid a lesser wage than the predominantly male occupational group. This wage difference is commonly called a “wage gap”. [Emphasis added.]

[104] After reaching its conclusion with respect to element two, the Tribunal addressed the third element, posing the following two questions to itself [at paragraph 354]:

D.	Does the comparison of the work of the Complainant group and the Comparator group establish that the work being compared is equal in value?
	Are the jobs/positions data and the process comparing the work of the Complainant and the Comparator groups reliable?

[105] The first of these questions makes it clear that the Tribunal considered the establishment of this element requires a conclusion that a comparison of the work of the two groups reveals that they were performing work of equal value in the same establishment. However, the Tribunal failed to provide any explanation with respect to the substance of this element or any interpretation of the phrase “work of equal value” for the purposes of the question that it posed to itself.

[106] In addition to this failure, the Tribunal also failed to explain how the phrase “work of equal value” applies in the circumstances of the complaint. Such an explanation would be necessary before the Tribunal could make a determination that the third element had been established.

[107] Without these explanations, the Court is put in the position of having to speculate on what the Tribunal may have thought about the meaning of “work of equal value” or attempting to undertake the analysis itself. In our view, the former alternative is impermissible and the latter is unfeasible having regard to the record before us. That said, we would venture to suggest that a fulsome analysis of the proper interpretation of “work of equal value”, conducted on the basis of an adequate record and the application of the criterion in subsection 11(2) of the Act, would permit the conclusion that in cases involving comparisons of the work of occupational groups, the equality of the compared work could be determined on a relative basis. In other words, it would be possible that the work of the occupational groups could be compared, for the purposes of subsection 11(1) of the Act, even if the value of that work was different in absolute terms. For example, the third element might be said to have been established by a finding that two units of the work done by the complainant occupational group equates in value to one unit of work done by the comparator occupational group.

[108] Notwithstanding its failure to provide any explanation of the requirements of the third element, the Tribunal found that to reach a conclusion with respect to the value of the work performed by the complainant and comparator occupational groups, the Tribunal must have evidence with respect to:

(a) the jobs that are being done in the two groups (the job information);

(b) the plan or methodology that is to be used to examine and evaluate the job information (the evaluation plan); and

(c) the process under which the evaluations are actually undertaken (the evaluation process).

[109] The Tribunal then determined that the evidence with respect to the job information, the evaluation plan and the evaluation process must be reliable.

[110] We take no issue with this determination by the Tribunal. However, we are of the view that the presence of reliable evidence with respect to these three matters does not automatically lead to a finding that the third element has been established. Assessing reliability is only the second of the steps required to find a “fact in issue”. In our view, the presence of such evidence is a necessary precondition to a finding that this element had been established. Indeed, a portion of the booklet upon which the Tribunal placed reliance (see Tribunal reasons, paragraph 358) states, “job evaluation plans are the key to determining what constitutes ‘work of equal value’.”

[111] Thus, once this necessary precondition has been fulfilled, the Tribunal would then be in a position to reach a conclusion that element three had been established. Unfortunately, we have been unable to discern any portion of the Tribunal’s reasons in which a conclusion with respect to this element has been reached.

[112] In our respectful view, the Tribunal became confused, and therefore fell into error, when it conflated the requisite conclusion with respect to the third element—that the work of the two groups that was being compared was of equal value on a relative basis—with the three evidentiary matters that must be present to permit that conclusion to be reached. This confusion is evident in paragraph 362 of the Tribunal’s reasons, wherein it framed four issues that it wished to consider. For ease of reference, paragraph 362 is reproduced:

Consequently, the issues which will be addressed are as follows:

1.	What job evaluation system, or plan, was used to undertake the evaluation of the CR and PO jobs/positions, and how reliable was it?
2.	What process was used and how reliable was it in analyzing the collected job data/information for purposes of assigning values to the CR and PO jobs/positions considered?
3.	What job data/information was collected, and from what sources, and how reliable was it?
4.	What were the resulting values attributed to the various CR and PO jobs/positions, and how reliable were they?

[113] In our view, the determinations in relation to the first three issues specified by the Tribunal—the existence of evidence with respect to the evaluation plan, the evaluation process and the job information and the reliability of such evidence—are essential inputs that must be present before a conclusion can be reached with respect to whether, on a relative basis, the work being compared is of equal value. However, we believe that the “resulting values”, referred to as the fourth issue, are the actual conclusion that the Tribunal is required to make with respect to the third element. It is these values that permit the determination that, on a relative basis, the work of the two groups that is being performed is of equal value. And, as has been stated earlier, the conclusion with respect to this element or “fact in issue” must be established on a balance of probabilities.

[114] A further indication of this conflation on the part of the Tribunal can be found in paragraph 412 of its reasons, in particular, the following sentence:

The issue is, then, given all the circumstances of the case before this Tribunal, is it more likely than not that the job information, from its various sources, the evaluation system and the process employed, and the resulting evaluations are, despite any weaknesses, sufficiently adequate to enable a fair and reasonable conclusion to be reached, as to whether or not, under section 11 of the Act, there were differences in wages for work of equal value, between the complainant and comparator employees concerned? [Emphasis added.]

In our view, the “resulting evaluations” are the conclusion that the Tribunal is required to make with respect to this element and should not be intertwined with findings of reliability. In other words, a finding by the Tribunal that the evidence tendered with respect to the determination of the relative value of work that is being compared is reliable is not the equivalent of a finding by the Tribunal, after considering all of the evidence before it, that, on a relative basis, the work being compared is of equal value, as required by subsection 11(1) of the Act.

[115] At paragraph 555, the Tribunal concludes that it must test, assess or weigh the reasonable reliability of the job information, the evaluation plan and the evaluation process using the “civil standard of a balance of probabilities”. It then goes on to make these determinations with respect to the evaluation plan and the evaluation process stating at paragraphs 571 and 593 respectively:

Therefore, the Tribunal finds that, on a balance of probabilities, the Hay Plan, whether using the factor comparison method or other approaches, is, in the hands of competent evaluators as were the members of the Professional Team, a suitable overall job evaluation scheme which will address the issues of this “pay equity” Complaint in a reasonably reliable manner.

…

Therefore, the Tribunal finds that it is more likely than not that the evaluation process which the Professional Team used in its work was reasonably reliable. [Emphasis added.]

[116] The Tribunal then considered the “daunting task” of testing the reasonable reliability of the job information evidence. Acknowledging the difficulty in assessing this evidence, the Tribunal found it helpful to consider the approach espoused by Professor Waddams (reproduced at paragraph 679 of the Tribunal’s reasons) with respect to proof of damages, namely that the court must make the most of, or do the best it can with, the available evidence.

[117] In addition, the Tribunal determined that, for the purposes of assessing the reasonable reliability of the job information, it was appropriate to consider “three sub-bands of reasonable reliability”. Using this construct, the Tribunal determined that the job information evidence was nonetheless reasonably reliable, stating at paragraph 700:

Hence, the Tribunal found, as stated in paragraph [689], that it was more likely than not that the job information utilized by the Professional Team in conducting its job evaluations of the CR and PO positions/jobs pertinent to this case, was reasonably reliable, albeit at the “lower reasonably reliable” sub-band level.

[118] With respect, we find that the use of the balance of probabilities standard in relation to the assessment of the reliability of evidence with respect to intermediate facts to be unusual and the use of “sub-bands of reasonable reliability” for that purpose to be even more unusual. In our view, the Tribunal’s findings with respect to the reliability of the evidence pertaining to the job information, the evaluation plan and the evaluation process are nothing more than findings that such evidence has some probative value and is worthy of consideration by the Tribunal. However, findings that evidence is reliable or worthy of consideration do not necessarily lead to the conclusion that such evidence has sufficient probative value to establish the third element, the “fact in issue” in respect of which such evidence was tendered. Again, in our view, the Tribunal has failed to reach the third of the steps required to find a “fact in issue” and has terminated its analysis at the second step.

[119] Having made its finding that the evidence with respect to the job information was reasonably reliable, the Tribunal, at paragraph 701 of its reasons, asked itself the following question, “How reliable were the resulting job evaluation values attributed by the Professional Team to the CR positions and PO jobs concerned?”

[120] The Tribunal then answered this question in paragraph 703 of its reasons, stating:

Accordingly, the Tribunal concludes that it is more likely than not that the aforementioned reasonably reliable Hay Plan, process and job information, in the hands of competent evaluators, as were the Professional Team, would result in reasonably reliable job evaluation values being attributed to the work performed by CR and PO employees.

[121] The applications Judge concluded that the Tribunal’s finding of reasonably reliable job values in this paragraph fell short of a finding that the third element had been established on a balance of probabilities. We are in agreement with his conclusion in that regard.

[122] To reiterate, the establishment of the third element requires a finding that the assessment of the value of the work performed by the two groups proves, on a balance of probabilities, that on a relative basis, the work being compared is of equal value. With respect, a finding that the value attributed to the work being compared is reasonably reliable cannot reasonably be said to be a finding that the work that is being compared is of equal value.

[123] We accept that because the Tribunal has correctly stated the standard of proof, it is entitled to the presumption that it has applied the correct standard of proof. However, the presumption cannot reasonably be considered to turn a finding that the value attributed to the work being compared is reasonably reliable into a finding that the work that is being compared is of equal value. At most, the presumption suggests that the Tribunal has found that the value attributed to the work being compared is reasonably reliable on a balance of probabilities, which is sufficient only to complete the second step required to establish a “fact in issue”. Accordingly, we are unable to conclude that paragraph 703 of the Tribunal’s reasons can reasonably be considered to contain a finding that the third element has been established. Certainly, it is not written in the same clear manner as paragraphs 283 and 354 of the Tribunal’s reasons, which stipulate that each of elements one and two has been established.

[124] Similarly, we do not accept that the first sentence of paragraph 801 of the Tribunal’s reasons can reasonably be considered to establish that the third element has been met. That sentence reads as follows:

The Tribunal accepts that the evidence of the Professional Team, both through the viva voce evidence of Dr. Wolf and also through the presentation of the Team’s Reports to the Tribunal, is sufficient, on a balance of probabilities, to demonstrate a wage gap when the work of the predominantly female CR’s was compared with the work of equal value being performed by the predominantly male PO’s at Canada Post.

Fairly interpreted, that sentence addresses no more than the acceptance by the Tribunal that the fourth element, which relates to the wages paid to employees in the two groups, has been established. Furthermore, paragraph 801 is found in section VII of the Tribunal’s reasons entitled “VII. WAGE GAP AND WAGE ADJUSTMENT METHODOLOGY”, which follows the section of the reasons in which a conclusion with respect to the third element would have been expected to have been reached.

[125] It is apparent that the Tribunal was aware of the requirement to make findings with respect to each of the four elements of a prima facie case of wage discrimination. Indeed, in paragraphs 283 and 354 of its reasons, the Tribunal made such findings in relation to the first and second elements in clear and unequivocal terms. The failure to make a clear and unequivocal finding that the third element had been established is as clear to us as are the Tribunal’s findings that the first two elements had been established.

[126] No reasons were given by the Tribunal for its failure to make a finding that the third element had been established. And, in our view, no justifiable, transparent or intelligible reasons could be offered to support that failure. Moreover, given the Tribunal’s awareness of the requirement to make such findings with respect to all four of the elements and the fact that it made two such findings in clear and unequivocal terms, we are hard pressed to conclude that the failure to make a finding that this important element was established was due to inadvertence or that such a finding should be considered to be implicit in its reasons.

[127] In the result, we are of the view that the Tribunal cannot reasonably be considered to have made a finding that the third element of a prima facie case of wage discrimination has been established. It follows, in our view, that in the absence of such a finding, it is a sufficient basis upon which to dismiss the complaint. To return to Lord Hoffmann’s binary system, the failure of the Tribunal to find that this element had been established leads to a value of zero being returned.

C. Whether the Tribunal applied the incorrect standard of proof

[128] The applications Judge found that the Tribunal correctly stated that the standard of proof with respect to the four elements is the balance of probabilities. No party takes issue with this finding. In addition, having regard to the presumption referred to by Rothstein J. in F.H., the Tribunal is presumed to have applied the correct standard of proof. The issue then becomes whether it can be said that this presumption has been rebutted having regard to the reasons of the Tribunal read as a whole.

[129] The applications Judge addressed this issue in the context of the third element and that will be our focus as well.

(a) Standard of review

[130] The question of whether the Tribunal applied the correct standard of proof is a question of mixed fact and law that contains no readily extricable legal issue. Accordingly, the standard of review of this question is reasonableness.

(b) Did the Tribunal apply the correct standard of proof?

[131] Earlier in these reasons, we concluded that the Tribunal erred to the extent that it held that the establishment of the third element would automatically result from findings that reasonably reliable evidence had been adduced with respect to the job evaluations, the evaluation plan and the evaluation process.

[132] To the extent that the Tribunal’s view has any validity, each of the three evidentiary matters would have to be characterized as an essential element or “fact in issue”, which would have to be established on a balance of probabilities, rather than an intermediate fact. On that basis, the complaint would still be doomed to fail. In our view, the findings that the Tribunal made with respect to each of these matters, in paragraphs 571, 593 and 700 of its reasons, to the effect that these matters were “reasonably reliable”, constitute findings that fall short of the requisite standard of proof on a balance of probabilities.

[133] Having made a finding, at paragraph 700 of its reasons, that the evidence with respect to the job information was reasonably reliable, the Tribunal, at paragraph 701 of its reasons, asked itself the following question, “How reliable were the resulting job evaluation values attributed by the Professional Team to the CR positions and PO jobs concerned?”

[134] The Tribunal then answered this question, at paragraph 703 of its reasons, stating:

Accordingly, the Tribunal concludes that it is more likely than not that the aforementioned reasonably reliable Hay Plan, process and job information, in the hands of competent evaluators, as were the Professional Team, would result in reasonably reliable job evaluation values being attributed to the work performed by CR and PO employees.

[135] The applications Judge concluded that the Tribunal’s finding of reasonably reliable job values in this paragraph fell short of a finding that the third element had been established on a balance of probabilities. We are in agreement with his conclusion in that regard.

[136] In addition, we note that the Tribunal found, at paragraphs 699 and 941 of its reasons, that the job information used in evaluating the CR positions and the PO positions was “lower reasonably reliable”. This led the Tribunal to conclude that there was a “significant degree of uncertainty” in the job information. The Tribunal then made the following comments in paragraphs 943, 944, 948 and 949:

Taking into account these elements of uncertainty which affect the very crucial aspect of determining the extent of the wage gap, it is, in the Tribunal’s view, more likely than not that if the job information and the non-wage benefits had been “upper reasonably reliable,” the resulting wage gap would have more accurately reflected reality. In other words, the greater the reliability of the job information and the non-wage benefits, the greater the accuracy of the wage gap determination. This determination is seminal to the extent of the award of damages.

Recognizing these elements of uncertainty in the state of the job information and non-wage benefits documentation, the Tribunal finds that it cannot accept the full extent of the wage gap as claimed by the Alliance and endorsed by the Commission.

…

Following the spectrum analysis already completed for the two elements of uncertainty, [job information, paragraph 941, and non-wage compensation, paragraph 942] the Tribunal concludes that a wage gap determination based upon “upper reasonable reliability” evidence should, logically, give rise to a 100% award of lost wages, a determination based upon “mid reasonable reliability” to a 75% award, and a determination based upon “lower reasonable reliability” to an award of 50% or less.

Accordingly, the Tribunal concludes that the finally determined award of lost wages for each eligible CR employee, by whatever methodology, should be discounted by 50% in line with the lower reasonable reliability status of the relevant job information and non-wage forms of compensation.

[137] Thus, at paragraph 949 of its reasons, the Tribunal reduced the award of lost wages by 50 percent so as to be “in line with the lower reasonable reliability status of the relevant job information and non-wage forms of compensation.” In doing so, it is our view that the Tribunal has thus equated the “lower reasonable reliability status” of the job information to 50 percent certainty. In our view, this conclusion of the Tribunal clearly demonstrates that a standard of proof lower than the balance of probabilities was applied by the Tribunal with respect to the third element.

[138] In our view, the presumption that the Tribunal applied the correct standard of proof has been amply rebutted. This is apparent for a number of reasons. First, the balance of probability standard requires the establishment of a “fact in issue”. In the binary formulation, the finding is either zero or one. The necessary finding is not “reasonably reliably one”, or “almost one” or “closer to one than zero”.

[139] Second, the Tribunal’s reliance on Professor Waddams’ urging that in the assessment of damages, the assessing body must do the best it can with the evidence that it has, indicates to us that the Tribunal had concerns with the evidentiary record before it. It is not necessary for us to consider whether it may have been appropriate for the Tribunal to rely upon Professor Waddams’ approach when attempting to determine the reliability of evidence with respect to job information, the evaluation plan and the evaluation process where those matters constitute intermediate facts upon which findings of “facts in issue” are based. However, in our view, it is not acceptable for the Tribunal to rely upon that approach when making findings with respect to those matters, where they constitute “facts in issue”, which must be established on a balance of probabilities.

[140] Third, the Tribunal referred a number of times (see paragraphs 573, 574, 581, 673 and 683 of the Tribunal’s reasons) to the particular circumstances of the case being difficult or unusual or litigious, as if to justify some sort of relaxation of the long-standing rules with respect to the burden of proof and the standard of proof in civil matters. This theme is amplified by the adoption of the “bands” and “sub-bands” of acceptability or reasonable reliability to which the Tribunal resorted.

[141] All of these justifications by the Tribunal demonstrate to us that it failed to make the requisite finding that the third element had been established on a balance of probabilities. In our view, they clearly rebut the presumption that such a finding was made. As such, we agree with the disposition of the applications Judge on this issue and conclude that it is a sufficient basis upon which to dismiss the appeal.

VIII. CONCLUSION AND DISPOSITION

[142] We have concluded that the Tribunal has made two errors, each of which is sufficient to vitiate the decision of the Tribunal. With respect to the first issue, which was not directly considered by the applications Judge, we have concluded that the Tribunal cannot reasonably be considered to have made a finding that the third element of a prima facie case of discrimination has been established. As a result, the absence of this essential finding makes it impossible for the complaint to be upheld.

[143] With respect to the second issue, we are in agreement with the applications Judge that the findings made by the Tribunal in relation to the third element of a prima facie case of wage discrimination fall short of proof of the level required by the balance of probabilities standard. As such, it was unreasonable for the Tribunal to uphold the complaint when the requisite level of proof of this essential element was not present.

[144] For the reasons previously given, we agree with the applications Judge that the level of proof with respect to at least one of the four elements failed to exceed 50 percent and therefore failed to attain the required level of proof of greater than 50 percent. As a result, we are of the view that the applications Judge was correct when he determined that the matter should be remitted to the Tribunal with a direction that the complaint should be dismissed as not having been substantiated. We are also of the view that it would be of no use to remit the matter to the Tribunal for reconsideration, given our conclusion that the Tribunal’s findings with respect to each of the four elements failed to meet the necessary level of proof on a balance of probabilities.

[145] Like the applications Judge, we also note the exceptional amount of time and resources consumed by this case. The length of the Tribunal hearing alone was 11 years, and it has now been 26 years since the complaint was filed. As the applications Judge noted at paragraph 274, “A legal hearing without discipline and timelines both delays and denies justice.” However, this exceptional consumption of time and resources does not influence our choice of remedy.

[146] For the foregoing reasons, we would dismiss the appeals without costs.

APPENDIX

RELEVANT STATUTORY PROVISIONS

Canadian Human Rights Act, R.S.C., 1985, c. H-6

Equal wages	11. (1) It is a discriminatory practice for an employer to establish or maintain differences in wages between male and female employees employed in the same establishment who are performing work of equal value.
Assessment of value of work	(2) In assessing the value of work performed by employees employed in the same establishment, the criterion to be applied is the composite of the skill, effort and responsibility required in the performance of the work and the conditions under which the work is performed.
Separate establishments	(3) Separate establishments established or maintained by an employer solely or principally for the purpose of establishing or maintaining differences in wages between male and female employees shall be deemed for the purposes of this section to be the same establishment.
Different wages based on prescribed reasonable factors	(4) Notwithstanding subsection (1), it is not a discriminatory practice to pay to male and female employees different wages if the difference is based on a factor prescribed by guidelines, issued by the Canadian Human Rights Commission pursuant to subsection 27(2), to be a reasonable factor that justifies the difference.
Idem	(5) For greater certainty, sex does not constitute a reasonable factor justifying a difference in wages.
No reduction of wages	(6) An employer shall not reduce wages in order to eliminate a discriminatory practice described in this section.
Definition of “wages”	(7) For the purposes of this section, “wages” means any form of remuneration payable for work performed by an individual and includes (a) salaries, commissions, vacation pay, dismissal wages and bonuses; (b) reasonable value for board, rent, housing and lodging; (c) payments in kind; (d) employer contributions to pension funds or plans, long-term disability plans and all forms of health insurance plans; and (e) any other advantage received directly or indirectly from the individual’s employer. … 27. …
Guidelines	(2) The Commission may, on application or on its own initiative, by order, issue a guideline setting out the extent to which and the manner in which, in the opinion of the Commission, any provision of this Act applies in a class of cases described in the guideline. …
Request for inquiry	49. (1) At any stage after the filing of a complaint, the Commission may request the Chairperson of the Tribunal to institute an inquiry into the complaint if the Commission is satisfied that, having regard to all the circumstances of the complaint, an inquiry is warranted. … 50. …
Additional powers	(3) In relation to a hearing of the inquiry, the member or panel may (a) in the same manner and to the same extent as a superior court of record, summon and enforce the attendance of witnesses and compel them to give oral or written evidence on oath and to produce any documents and things that the member or panel considers necessary for the full hearing and consideration of the complaint; (b) administer oaths; (c) subject to subsections (4) and (5), receive and accept any evidence and other information, whether on oath or by affidavit or otherwise, that the member or panel sees fit, whether or not that evidence or information is or would be admissible in a court of law; (d) lengthen or shorten any time limit established by the rules of procedure; and (e) decide any procedural or evidentiary question arising during the hearing.

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

	18.1 …
Powers of Federal Court	(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 refusd 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.
	…
Powers of Federal Court	52. The Federal Court of Appeal may … (b) in the case of an appeal from the Federal Court, (i) dismiss the appeal or give the judgment and award the process or other proceedings that the Federal Court should have given or awarded, (ii) in its discretion, order a new trial if the ends of justice seem to require it, or (iii) make a declaration as to the conclusions that the Federal Court should have reached on the issues decided by it and refer the matter back for a continuance of the trial on the issues that remain to be determined in light of that declaration …

Equal Wages Guidelines, 1986, SOR/86-1082

12. Where a complaint alleging different wages is filed by or on behalf of an identifiable occupational group, the group must be predominantly of one sex and the group to which the comparison is made must be predominantly of the other sex.

13. For the purpose of section 12, an occupational group is composed predominantly of one sex where the number of members of that sex constituted, for the year immediately preceding the day on which the complaint is filed, at least

(a) 70 per cent of the occupational group, if the group has less than 100 members;

(b) 60 per cent of the occupational group, if the group has from 100 to 500 members; and

(c) 55 per cent of the occupational group, if the group has more than 500 members.

14. Where a comparison is made between the occupational group that filed a complaint alleging a difference in wages and other occupational groups, those other groups are deemed to be one group.

15. (1) Where a complaint alleging a difference in wages between an occupational group and any other occupational group is filed and a direct comparison of the value of the work performed and the wages received by employees of the occupational groups cannot be made, for the purposes of section 11 of the Act, the work performed and the wages received by the employees of each occupational group may be compared indirectly.

(2) For the purposes of comparing wages received by employees of the occupational groups referred to in subsection (1), the wage curve of the other occupational group referred to in that subsection shall be used to establish the difference in wages, if any, between the employees of the occupational group on behalf of which the complaint is made and the other occupational group.

* * *

The following are the reasons for judgment rendered in English by

Evans J.A. (dissenting):

A. INTRODUCTION

[147] This case concerns a pay equity claim filed 27 years ago. Its investigation and adjudication must have involved the expenditure of vast quantities of money and time, both public and private. For members of the appellant, the Public Service Alliance of Canada (PSAC), the resolution of a complaint of long-standing gender discrimination in the workplace is at stake. For Canada Post Corporation (CPC), the financial implications of the decision by the Canadian Human Rights Tribunal (Tribunal) are no doubt considerable.

[148] This appeal raises three important questions for the implementation of the legal principle that an employer may not pay men and women differently for work of equal value. First, in a pay equity claim by a predominantly female occupational group, may the Tribunal select a predominantly male comparator group that includes a significant number of relatively well-paid women? Second, when the Tribunal states that it is applying the civil standard of proof, what weight must be given to the presumption that this is the standard that it in fact applied? Third, after the jobs of members of the complainant and comparator groups have been evaluated, what findings must be made before their wages can be compared in order to determine if they are being paid differently for performing work of equal value?

[149] I have had the benefit of reading the reasons of my colleagues, Sexton and Ryer JJ.A. I regret that I am unable to agree that the appeals of PSAC and the Canadian Human Rights Commission (CHRC) in Court files A-129-08 and A-139-08 should be dismissed. In my opinion, the Tribunal made no error warranting judicial intervention when it upheld the pay equity claim by PSAC on behalf of the predominantly female Clerical and Regulatory (CR) occupational group of employees of CPC in respect of the period August 24, 1982 to June 2, 2002.

[150] On June 2, 2002, CPC implemented a new job evaluation plan and awarded a 15 percent wage increase to the CR group, while limiting other groups to wage increases of approximately 1.5 percent. PSAC regarded the new job evaluation plan and the 15 percent wage increase as prospectively removing any violation of section 11 of the Act. However, CPC denies that the CR group had previously been paid less than the comparator group in the pay equity claim, the Postal Operations occupational group (PO), for performing work of equal value.

[151] In my respectful opinion, when the reasons of the Tribunal are read holistically, and against the background of the expert evidence on which it relied, it found on a balance of probabilities that members of the CR group were paid less than members of the PO group for work of equal value.

[152] In order to establish a breach of section 11 of the Canadian Human Rights Act, R.S.C., 1985, c. H-6 (Act) in a group pay equity claim, the Tribunal must find on a balance of probabilities that the jobs performed by members of the groups have been properly evaluated. In my view, the Tribunal was entitled to infer this from its conclusions that the evidence on the nature of the jobs was reasonably reliable, that the Hay method evaluation plan was a reasonably reliable tool for evaluating the job data (the methodology issue), and that it had been applied in a reasonably reliable manner by the evaluators (the process issue).

[153] After the jobs performed by members of the two groups have been evaluated, the values attributed to the various positions are compared. If a substantial portion of the positions in the complainant group have a value equal to or greater than one or more of the positions in the comparator group, their wages can be compared in order to determine if the complainant group is being paid less than the comparator group for performing work of equal value.

[154] In 1993, PSAC retained three professional job evaluators, Dr. Wolf, Dr. Ingster and Ms Davidson-Palmer, referred to as the professional team, to provide an expert review of the evaluations undertaken by CHRC in 1991 and, on the basis of an independent evaluation of the jobs, to identify any difference in the wages paid to members of the two groups for performing work of equal value.

[155] Having evaluated the CR and PO positions by the Hay method, the professional team reported that a substantial portion of the CR positions were at least equal in value to one or more of the PO positions, or, in other words, fell within the PO value line. Hence, the wages of the two groups could properly be compared to determine if the complainant group had established the existence of a wage gap for work of equal value in breach of section 11. The Tribunal (at paragraphs 799 and 801) accepted the professional team’s conclusions.

[156] Like my colleagues, I would dismiss PSAC’s appeal in Court File No. A-130-08 from Justice Kelen’s dismissal of its application for judicial review of the amount of compensation awarded to it by the Tribunal. However, because I would allow PSAC’s appeal on the breach of section 11, I have had to consider the appeal on the amount of compensation on its merits. I am not persuaded that the Tribunal’s award was unreasonable.

[157] On the other hand, I need not determine if, as PSAC and CHRC allege, the applications Judge erred in law when he set aside the Tribunal’s decision and did not remit the matter for redetermination, on the ground that, in the Judge’s view, the evidence did not establish a breach of section 11 on a balance of probabilities.

B. CONTEXTUAL BACKGROUND

[158] I gratefully adopt my colleagues’ description of the factual background to this litigation and the history of the judicial proceedings leading to this appeal. I would add only the following by way of context to my reasons.

[159] First, this is yet another example of the marathon litigation that has plagued the resolution of pay equity claims at the federal level. The timelines of the proceedings are so extraordinary that they bear repeating.

• August 24, 1983: PSAC files a pay equity claim under section 11 of the Act on behalf of the CR group employed by CPC, a Crown corporation created in 1981 to take over the functions previously performed by the Post Office;

• March 16, 1992: CHRC refers PSAC’s complaint to the Tribunal for adjudication;

• November 25, 1992 to August 2003: Tribunal hearing, comprising 410 hearing days spread over more than 10 years;

• October 7, 2005: Tribunal renders its decision;

• February 21, 2008: Federal Court sets aside Tribunal’s decision;

• February 22, 2010: Federal Court of Appeal dismisses appeal.

[160] Second, PSAC and CPC were unable to work together on a joint union-management study to produce an agreed evaluation of the work performed by members of the CR group and the occupational group that it identified for wage comparison purposes, the PO group. As a result, the information available to the Tribunal about both the nature of the work performed by members of these groups and the non-monetary components of their wages had some significant limitations.

[161] Responsibility for these deficiencies is attributable to, among other things, the fact that: the parties were operating in litigation mode from relatively early in the process; the CHRC did not require CPC to produce all relevant documents; and the underlying rivalry between, on the one hand, PSAC, the bargaining agent for the CRs and for two small, predominantly male occupational groups at Canada Post, the General Labour and Trades, and General Services groups, and, on the other, the Canadian Union of Postal Workers, the bargaining agent for most of the POs.

[162] Third, courts never intervene lightly in the administrative process, both out of deference to the expertise of specialized tribunals and in recognition of the limitations of reviewing courts’ perspectives on the problem before the agency. Further, setting an administrative decision aside, whether or not the matter is remitted to the tribunal for rehearing, inevitably results in a waste of resources.

[163] These considerations are particularly apt in this case: the subject-matter of these proceedings is complex and has some highly technical aspects, and, as already noted, the public and private resources already spent on this dispute must be enormous. In the course of the 1016 paragraphs of its reasons, the Tribunal is not always as clear as it might have been, as it struggled to come to terms with the mass of technical detail, evidence, and analysis before it. As already noted, the Tribunal’s task was particularly challenging because the parties failed to produce a joint study on the values of the jobs and the total wages paid over the period of the complaint. However, perfection is not the standard and, when read in light of the evidence before it and the nature of its statutory task, the Tribunal’s reasons adequately explain the bases of its decision.

[164] In my opinion, the Tribunal’s reasons make it clear that it understood the relevant law and approached the complex evidential issues before it in a careful and thoughtful manner. Its reasons sufficiently demonstrate “justification, transparency and intelligibility within the decision-making process”: Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 S.C.R. 190 (Dunsmuir), at paragraph 47. The Court in Dunsmuir (at paragraph 48) also endorsed the view of Professor Dyzenhaus (“The Politics of Deference: Judicial Review and Democracy”, in M. Taggart, ed., The Province of Administrative Law (Oxford: Hart Publishing, 1997), 279, at page 286) that, when reviewing a decision on the reasonableness standard, a court must pay “respectful attention to the reasons offered or which could be offered in support of a decision” (emphasis added). The underlined words avoid an unduly formalistic approach to judicial review. Thus, to the extent that the Tribunal does not fully explain aspects of its decision, the Court may consult evidence referred to by the Tribunal in order to flesh out its reasons. However, I do not regard the Court in Dunsmuir as inviting a reviewing court to usurp the tribunal’s responsibility for justifying its decisions.

[165] The resolution of pay equity claims involves a mix of art, science, human rights, and labour relations. It can be difficult to fit multi-disciplinary inquiries of this nature within a legal framework: social scientists and management consultants do not always express themselves in the same terms as lawyers, on questions of evidence and proof, for example.

[166] Fourth, the underlying purpose of section 11 of the Act is to eliminate the financial consequences of systemic gender discrimination in the labour market resulting from occupational segregation. However, with the benefit of hindsight, it now seems to have been a mistake for Parliament to have entrusted pay equity to the complaint-driven, adversarial, human rights process of the Canadian Human Rights Act.

[167] There is now much to learn from the experience of provincial pay equity regimes, which seem not to have been plagued with the same problems of protracted litigation as the federal scheme. In the interests of all, a new design is urgently needed to implement the principle of pay equity in the federal sphere. For criticisms of the present arrangements, and recommendations for reform, see the Final Report of the Pay Equity Task Force, Pay Equity: A New Approach to a Fundamental Right (Ottawa: Public Works and Government Services Canada, 2004).

C. ISSUES AND ANALYSIS

[168] The standard of review applicable to the various issues in dispute in this case is common ground: correctness governs the Tribunal’s choice of standard of proof, and reasonableness (appropriately contextualised in its application) the Tribunal’s findings of fact and discretionary decisions, including its choice of comparator group and remedy. As my colleagues point out, the task of this Court on appeal is to decide if the applications Judge selected the appropriate standard of review and applied it correctly.

[169] The three principal areas of inquiry explored by counsel in this appeal concern the Tribunal’s choice of the PO group as the comparator, the standard of proof that it applied to reach its conclusion that members of the CR group were paid lower wages than members of the PO group for performing work of equal value, and the remedy that it awarded. More particularly, argument focussed on whether the Tribunal applied the balance of probabilities standard of proof when accepting the professional team’s evaluation of the jobs performed by the complainant and comparator groups, and the adequacy of the evidence on which the Tribunal based its finding of a wage gap.

[170] In addition, counsel for CPC argued that the Equal Wages Guidelines, 1986 (Guidelines) did not apply to the complaint from 1983 until the Guidelines were issued, because to do so would give them retroactive effect. For substantially the reasons given by the applications Judge, I agree that this argument cannot succeed. I need only add the following.

[171] First, when the Supreme Court of Canada stated that CHRC’s power to issue guidelines could not validly be exercised retroactively, it gave as an example a guideline issued while a matter was being prosecuted before a Tribunal: Bell Canada v. Canadian Telephone Employees Association, 2003 SCC 36, [2003] 1 S.C.R. 884, at paragraph 47. However, that is not our case: the Guidelines were issued while CHRC was in the relatively early stages of its investigation of PSAC’s complaint, and long before it was referred to the Tribunal.

[172] Second, there is no evidence that the issuance of the Guidelines altered the basis of CHRC’s investigation. For the most part, the Guidelines seem simply to have codified existing CHRC policy on, among other things, the percentages required for occupational groups of various sizes to be treated as predominantly of one sex.

[173] Third, the Guidelines amended the law by specifying the manner in which a breach of section 11 of the Act is established, not by changing the definition of discriminatory conduct. They did not remove CPC’s vested rights.

ISSUE 1: Was the choice of the comparator group unreasonable because it included a substantial number of well-paid women?

[174] PSAC proposed to compare the complainant CR group, which numbered about 2 300 employees over the period of this dispute, with the PO group, which comprised approximately 40 000 employees and constituted 80 percent of CPC’s workforce. This large occupational group, which contains employees engaged in internal and external postal work, as well as in supervisory duties, dates from the time when, as a department of the federal government, the Post Office was responsible for the mail. All its employees who dealt with mail were grouped together within the larger institution of the federal public service. Other Post Office employees, including the CRs, whose work was not peculiar to the Post Office, were in occupational groups found also in other departments of the federal government. These occupational groups were continued by the employer after 1981, when the functions of the Post Office were transferred to CPC, a newly created Crown corporation.

[175] CPC argued in this Court that the Tribunal had unreasonably exercised its discretion to select a comparator group when it included mail sorters (PO-4, internal level), a level of internal workers within the larger PO group. Employees in the PO-4 internal level, numbering about 20 000, or 40 percent of the total PO group, were relatively well paid, and included approximately 10 000 women.

[176] Mail sorting was traditionally “women’s work”. CPC said that PSAC had “cherry-picked” its proposed comparator group in order to include relatively well-paid jobs and thus artificially to create, or widen, a wage gap. Further, the presence of a large number of relatively well-paid female employees demonstrated that there was no systemic gender discrimination at CPC. Counsel also argued that PSAC had deliberately omitted from the comparator group two smaller, predominantly male occupa- tional groups, which it represented, and whose members were relatively low-paid.

[177] There is little law on the selection of a comparator group in a pay equity claim. Two legal requirements must be met: the comparator must be an occupational group (section 12 of the Guidelines) and must be predominantly of the opposite sex from that of the complainant group. Section 13 of the Guidelines defines when a group is predominantly of one sex for the purpose of the Guidelines. If a group has less than 100 members, 70 percent of them must be of one sex; if there are between 100 and 500 members, 60 percent suffices; if the group has more than 500 members, then 55 percent is enough.

[178] Apart from the requirement that a comparator group must constitute an occupational group, and be predominantly of the opposite sex from the complainant group, as defined in section 13, there are no statutory criteria that must be considered in the selection of a comparator. The choice is left to the discretion of CHRC and the Tribunal.

[179] It was suggested in argument that the PO group was not an occupational group because its members performed different kinds of work: internal (mostly mail sorting), external (mostly mail delivery), and supervisory.

[180] However, even if this is right, it does not take matters much further. Section 14 of the Guidelines provides that when a complainant occupational group compares itself to more than one other occupational group, those groups are to be treated as a single group. Accordingly, if the PO group cannot be a comparator group for the purpose of a pay equity complaint, the three separate occupational groups (internal, external, and supervisory employees) that it comprises are to be treated as one. It is undisputed that both the PO group as a whole and each of the three sub-groups meet the requirement of being predominantly male.

[181] The question to be decided is whether the Tribunal abused its statutory discretion in its selection of the POs as the comparator group. A reviewing court should approach this issue with great caution. In determining whether the choice of comparator group in this case was unreasonable, the Court must consider both the Tribunal’s reasons and the outcome: Dunsmuir, at paragraph 47.

[182] The application of the unreasonableness standard requires a consideration of context: Canada (Citizenship and Immigration) v. Khosa, 2009 SCC 12, [2009] 1 S.C.R. 339, at paragraph 59. In the present case, the context includes the fact that the Tribunal’s statutory discretion is broad: it is not subject to any express constraints. Further, the Supreme Court of Canada has held that the function of selecting a suitable comparator group lies at the heart of the expertise of CHRC and the Tribunal: Canada (Human Rights Commission) v. Canadian Airlines International Ltd., 2006 SCC 1, [2006] 1 S.C.R. 3, at paragraph 42. In these circumstances, a high degree of judicial deference is owed to the Tribunal’s exercise of discretion.

[183] The PO-4 internal level (the mail sorters) within the internal workers group comprised approximately equal numbers of men and women, and was therefore predominantly neither male nor female. However, this does not in itself disqualify members of the PO-4 internal level from being included as part of a larger comparator group, that is, either the internal workers or the PO group as a whole.

[184] Rather, CPC argued, since the PO-4 internal level comprised 40 percent of the PO occupational group, and its members were relatively highly paid, it was unreasonable for the Tribunal to include the PO-4s in the comparator group because their presence would artificially create or widen an apparent wage gap between men and women. Further, the inclusion of the PO-4 internal level would be contrary to the purpose of section 11, namely, the elimination of systemic gender discrimination. How, it was asked, could there be systemic discrimination against CPC’s female employees when so many are well paid?

[185] The 1986 Guidelines explicitly recognize that a predominantly male comparator group may contain a minority of women. The female members of a comparator group are not thereby “masked” or treated as males. The assumption of the Guidelines is that a female minority in an occupational group may receive higher wages because of the male predominance. Conversely, a male minority may be disadvantaged by being part of a predominantly female occupational group.

[186] In upholding the PO occupational group as the comparator, as proposed by PSAC and endorsed by CHRC, the Tribunal noted (at paragraph 281) that it comprised approximately 80 percent of CPC’s total workforce. Larger occupational groups provide a more reliable basis than smaller groups for determining the existence of a wage gap between men and women performing work of equal value. Moreover, the Tribunal stated, the duties of some members of the PO group were similar to those of some members of the CR group. The work performed by some members of the groups was also similar in terms of skill, effort, responsibility and working conditions.

[187] Having thus found good reasons for selecting the POs as the comparator group, the Tribunal rejected CPC’s argument that PSAC had “cherry-picked” the PO group in order to skew the comparison by selecting a relatively highly paid group of employees. In my opinion, the Tribunal’s reasons provide a rational basis for its exercise of discretion and that judicial intervention is not warranted.

[188] Nor am I persuaded that the inclusion of the PO-4 internal level within the sub-group of internal workers vitiated the Tribunal’s choice of comparator on the ground that the presence of a substantial number of relatively well-paid women in CPC’s workforce effectively undermined the CR group’s complaint of systemic gender discrimination. I do not agree that, because the PO group includes a substantial number of well-paid women (the PO-4s), the selection of the POs as the comparator is contrary to the purpose of section 11, namely the elimination of systemic gender discrimination in the labour market.

[189] First, the Guidelines specifically contemplate the presence of women within a male-dominated comparator group, and vice versa. The presence of well-paid women in the PO group is not being “masked”. And, as already noted, both the PO group as a whole and the PO internal sub-group are predominantly male as defined by the Guidelines.

[190] Second, the fact that some women at CPC were relatively well paid does not necessarily preclude the existence of systemic gender discrimination elsewhere in the corporation. Systemic gender discrimination means that work performed by women tends to be undervalued, not that this is necessarily the case in every situation. The fact that the PO-4 internal level has become gender-neutral, so that mail sorting has lost its character as “women’s work”, and is performed within a predominantly male occupational group, may well explain why women in the PO-4 internal level are relatively well paid.

[191] Third, counsel referred us to no principle that requires the removal of some members of an occupational group from the comparator group. While the Guidelines contemplate the use of more than one occupational group as a comparator, they do not suggest that part of an occupational group may be used.

[192] CPC has also argued that the PO-4s should themselves constitute the comparator. There are several difficulties with this argument. First, although the PO-4 internal level comprised approximately 80 percent of the internal workers sub-group, and over 40 percent of the PO group as a whole, it is not an employer-designated occupational group, but merely one level within the sub-group of internal workers in the PO group. “Levels” connote wage differentials within an occupational group. Since wages for one level are set in relation to others, the wages of one level cannot be considered in isolation from those of the rest of the occupational group which, as already noted, was in this case predominantly male.

[193] Second, the PO-4 internal level comprised roughly equal numbers of men and women, and was thus “gender neutral”. It therefore could not be a comparator, because it was not predominantly of the opposite sex from the predominantly female CR group.

[194] The Tribunal also rejected the argument that the General Labour and Trades, and General Services occupational groups, which were represented in collective bargaining by PSAC, were more appropriate comparators. These groups represented only a small percentage of CPC employees and did not perform work similar to that of any members of the CR group.

[195] For these reasons, I am not persuaded that the Tribunal’s choice of comparator group was unreasonable or contrary to the purpose of the Act. In my respectful view, the applications Judge erred in law in concluding that the Tribunal had committed reviewable error in the exercise of its broad discretion over the selection of a comparator group.

ISSUE 2: Did the Tribunal apply the correct standard of proof when finding that members of the complainant and comparator groups were performing work of equal value?

(i) Facts in issue

[196] When determining whether members of a complainant group are paid less than those of the comparator group for performing work of equal value, a Tribunal must make two findings about their jobs.

[197] First, the value of the jobs performed by the members of the groups must be assessed. Subsection 11(2) of the Act prescribes that the value of work must be assessed on the basis of a composite of the skill, effort, and responsibility required, and the conditions under which the work is performed. The Tribunal must weigh the evidence before it on these aspects of the work and determine whether it is sufficiently cogent to enable the Tribunal to conclude on a balance of probabilities that the jobs had been properly evaluated.

[198] Because it may be impractical to collect the necessary data for all the jobs performed by members of the groups, it is sufficient to evaluate the work performed by representative samples of the groups. The Tribunal will normally have before it job evaluations submitted on behalf of the parties as a result of a joint union-management study. In this case, however, job evaluations were submitted on behalf of the complainants alone. The Tribunal relied on the reports of the professional team retained by PSAC and on the viva voce evidence of the “spokesperson” of the team, Dr. Wolf, in adopting the team’s evaluation of the work performed by members of the groups.

[199] Second, on the basis of an evaluation of the work, the Tribunal must then decide if enough members of the complainant group were performing work of at least equal value to that of members of the comparator group to enable it to determine if there was a gender-based wage gap in breach of section 11. As already indicated, the evidence of the professional team was that if a substantial portion of the CR positions were at least equal in value to one or more of the PO positions, the wages of the two groups could be compared to determine if they were being paid differently for performing work of equal value.

[200] In a report submitted in 1995 (PSAC-29), the professional team found that 62.9 percent of the jobs of the CR group were of equal or greater value than the least valuable job in the PO group. However, in its report of June 2000 (PSAC-180), the professional team reviewed its earlier evaluations in the light of evidence subsequently produced by CPC, mainly related to the PO positions.

[201] In this latter report, the professional team found that, while the new information made relatively little difference to most evaluations, it did significantly affect the value previously attributed to two PO generic jobs: the value of one job, relief mail services courier, was revised down, while the value of the other, counter clerk, was raised. As a result of the increase in the evaluation of the generic job of counter clerk, the number of CR positions falling within the PO value line was reduced by nearly a half.

[202] Nonetheless, the professional team obviously regarded 34 percent as a “substantial portion” of members of the complainant group who come within the comparator group’s value line so as to enable the wages of the two groups to be compared for the purpose of determining whether the CR group was being paid less than the PO group for performing work of equal value. CPC did not argue that 34 percent was too small a number for this determination to be made.

[203] Of course, if the jobs are not reliably evaluated in accordance with the statutory criteria, it cannot be established on a balance of probabilities that members of the complainant group were paid less than the comparator group for work of equal value.

[204] However, I should also add this. Historically, women in predominantly female occupations have generally been paid less than men for work of equal value, including in the federal public service. Hence, a finding that the CR group was paid less than the PO group for performing work of equal value would hardly be a surprise, especially since clerical work has traditionally been “women’s work” and women comprised more than 80 percent of the CR group. Indeed, the surprise would have been a finding that CPC did not fit the historic pattern of undervaluing “women’s work”. Nonetheless, this does not relieve PSAC and CHRC from having to adduce evidence to prove to the Tribunal on a balance of probabilities that CPC was in breach of section 11.

(ii) Standard of proof

[205] The relevant law on this issue is clear and not in dispute in this appeal. Complainants before the Canadian Human Rights Tribunal have the burden of proving that the respondent has prima facie discriminated against them contrary to the Act: see, for example, Public Service Alliance of Canada v. Canada (Department of National Defence), [1996] 3 F.C. 789 (C.A.) (Department of National Defence), at paragraph 33. Absent some special legislation, a balance of probabilities is the standard of proof applicable to civil proceedings in Canada: F.H. v. McDougall, 2008 SCC 53, [2008] 3 S.C.R. 41 (McDougall). “Civil proceedings” include proceedings before human rights tribunals: Department of National Defence, at paragraph 33.

[206] After noting that there was some judicial authority for the proposition that the civil standard of proof varies according to the seriousness of the outcome for the parties and the importance of the interests at stake, Justice Rothstein said in McDougall (at paragraph 44):

In my view, the only practical way in which to reach a factual conclusion in a civil case is to decide whether it is more likely than not that the event occurred.

In addition, he noted (at paragraph 54):

Where the trial judge expressly states the correct standard of proof, it will be presumed that it was being applied. Where the trial judge does not express a particular standard of proof, it will also be presumed that the correct standard was applied.

I take it that, like the standard of proof itself, this presumption applies to decisions of the Canadian Human Rights Tribunal.

[207] Whether the Tribunal in the present case committed a reviewable error in its application of the balance of probabilities standard of proof to the material before it is, of course, a different question. I need only say at this point that, when it comes to fact-finding, especially on difficult technical issues such as those involved in this pay equity dispute, the Tribunal is operating at the heart of its specialized jurisdiction, and its findings of fact are owed a high degree of deference, as the wording of paragraph 18.1(4)(d) [as enacted by S.C. 1990, c. 8, s. 5; 2002, c. 8, s. 27] of the Federal Courts Act indicates: Canada (Citizenship and Immigration) v. Khosa, 2009 SCC 12, [2009] 1 S.C.R. 339, at paragraph 46.

[208] Facts in issue are facts that are legally necessary for a plaintiff to win its case. They must be proved on a balance of probabilities. In this case, the facts in issue are the value of the jobs and, if the distribution of the values of the work performed by the complainant and comparator groups permits, the wages paid for work of equal value. However, establishing the value or, more accurately perhaps, the relative value of work, is not a purely scientific exercise, admitting of a uniquely correct answer. It calls for the exercise of judgment; not all evaluators would necessarily adopt the same methodology for assessing work, or place the same value on given jobs. Those assessing the value of work must be afforded a margin of appreciation in applying the appropriate methodology to the job data.

[209] Facts in issue should be distinguished from the evidence or intermediate facts on which findings of the facts in issue are based. It is unnecessary and, in my view, unhelpful for an adjudicator to introduce the notion of a balance of probabilities when weighing items of evidence to determine their probative value. “Balance of probabilities” is best reserved as the standard to be used by a fact-finder when determining whether, when all the evidence is weighed, a fact in issue has been proved.

[210] However, this is not to say that the weight attached to evidence is unrelated to the question of whether a fact in issue has been proved on a balance of probabilities. An adjudicator cannot conclude that a fact in issue has been proved on a balance of probabilities if the only evidence is unreliable. Conversely, I cannot imagine that an adjudicator would describe evidence as reliable unless it was more likely than not to be true, or would describe evidence as “reasonably reliable” that she thought was no more likely to be correct than to be wrong.

[211] In the present case, the Tribunal had before it three kinds of evidence from which to determine if the professional team had accurately assessed the value of the work performed by the CR and PO groups. First, it considered job information on the nature of the work and the wages paid for that work. Second, it considered whether the Hay method evaluation plan used by the professional team was an appropriate methodology for assessing the value of the work performed by the CR and PO groups by reference to the statutory criteria of skill, effort, responsibility, and working conditions. Third, it considered whether the evaluators had adopted a proper process in applying the methodology to the data.

[212] Having found that the only three items of evidence on which it could assess the professional team’s evaluation of the jobs were reasonably reliable, the Tribunal could conclude on a balance of probabilities that the jobs had been properly evaluated.

[213] In the course of its reasons, the Tribunal did at times refer to a balance of probabilities or its equivalent, “more likely than not”, when assessing the reliability of items of evidence. To ask, as it did, whether it is more likely than not that certain evidence was reasonably reliable may be redundant. It does not in my view, however, amount to an error of law by demonstrating that the Tribunal deviated from the task that it had set itself: to assess the reliability of each of these items of evidence and to ask whether, taken as a whole, they established on a balance of probabilities that the professional team had properly evaluated the work.

(iii) Tribunal’s reasons

[214] I turn now to the reasons of the Tribunal to determine if it applied the balance of probabilities standard of proof when making findings of the facts in issue. The first fact in issue is the proper evaluation of the jobs. If the Tribunal was satisfied of this on a balance of probabilities, and if a substantial portion of the CR jobs were at least as valuable as the lowest valued PO job, it could then determine whether the complainants were being paid less than the comparator group for performing work of equal value contrary to section 11. In its memorandum of fact and law, CPC did not challenge that this was an appropriate basis for being able to compare the wages of the two groups in order to determine if the CR group had been paid less than the PO group for performing work of equal value.

[215] In its overview of the legal principles governing a human rights complaint, the Tribunal correctly stated (at paragraph 69) that a prima facie case of discrimination must be established on “the civil standard, a balance of probabilities.” Turning later to the question of whether there was a prima facie case of discrimination contrary to section 11, the Tribunal provides the reader with a road map of its task (at paragraph 257):

Therefore, when addressing section 11 in the context of the Complaint before this Tribunal, each of the following elements must be proven, on a balance of probabilities. The elements are taken from section 11 of the Act and from the guidance which is offered concerning the particularizing of the section through guidelines promulgated by the Commission pursuant to its mandate under section 27 of the Act.

(1) The complainant occupational group is predominantly of one sex and the comparator occupational group is predominantly of the other sex. In this Complaint, that means the complainant CR’s must be predominantly female and the comparator PO’s must be predominantly male.

(2)The female-dominated occupational group and the male-dominated occupational group being compared are composed of employees who are employed in the same establishment.

(3) The value of the work being compared between the two occupational groups has been assessed reliably on the basis of the composite of the skill, effort, and responsibility required in the performance of the work, and the conditions under which the work is performed. The resulting assessment establishes that the work being compared is of equal value.

(4) A comparison made of the wages being paid to the employees of the two occupational groups for work of equal value demonstrates that there is a difference in wages between the two, the predominantly female occupational group being paid a lesser wage than the predominantly male occupational group. This wage difference is commonly called a “wage gap”. [Emphasis added.]

[216] In my view, the Tribunal correctly identified in this paragraph the facts in issue and the applicable standard of proof. I shall focus on element three because this is where my colleagues say that the Tribunal erred. Nor can I fault the Tribunal’s statement that the complainants must prove on a balance of probabilities that the value of the work had been “reliably” assessed.

[217] The Tribunal’s statement that the assessments establish that the work performed by members of the groups being compared is of equal value is also correct. Since a substantial portion of CR jobs were more valuable than the lowest valued PO job, it was possible to compare the wages paid to the PO and CR groups to determine if they were being paid differently for performing work of equal value.

[218] Later in its reasons, the Tribunal repeats that it has identified a balance of probabilities as the standard to be applied to the proof of facts in issue, this time to the proof of a wage gap (at paragraphs 801 and 803):

The Tribunal accepts that the evidence of the Professional Team, both through the viva voce evidence of Dr. Wolf and also through the presentation of the Team’s Reports to the Tribunal, is sufficient, on a balance of probabilities, to demonstrate a wage gap when the work of the predominantly female CR’s was compared with the work of equal value being performed by the predominantly male PO’s at Canada Post.

…

Having accepted that there is a wage gap, and, consequently, there is proof, on a balance of probabilities, that there has been systemic discrimination in this “pay equity” complaint, the next step is to select the most appropriate wage adjustment methodology to use to calculate an award of lost wages and to eliminate the gap. [Emphasis added.]

[219] In my view, these passages amply demonstrate that the Tribunal has identified a balance of probabilities as the standard of proof of the facts in issue. It is therefore entitled to the benefit of the McDougall presumption that this is the standard that it in fact applied. The question to be decided, therefore, is whether other aspects of the Tribunal’s reasons are so wayward as to rebut the presumption and to lead to the conclusion that, contrary to its clear assertion to the contrary, the Tribunal in fact applied some lower standard.

[220] Without going through the Tribunal’s reasons in undue detail, I shall refer to paragraphs that seem to have caused most concern as to whether the Tribunal applied the balance of probabilities standard to the facts in issue, namely, whether the wage comparison of the CR and PO groups related to work of equal value (at paragraph 412):

These rulings [in the three cases cited above] support a call for a standard of reasonableness, there being no such thing as absolute reliability. The application of such a standard will depend very much on the context of the situation under examination. The issue is, then, given all the circumstances of the case before this Tribunal, is it more likely than not that the job information, from its various sources, the evaluation system and the process employed, and the resulting evaluations are, despite any weaknesses, sufficiently adequate to enable a fair and reasonable conclusion to be reached, as to whether or not, under section 11 of the Act, there were differences in wages for work of equal value, between the complainant and comparator employees concerned?

[221] The difficulty with this paragraph, my colleagues say, is that by focussing on the reliability of the job information, and the methodology and process used to evaluate the jobs, the Tribunal deviated from its task of deciding whether it had been established on a balance of probabilities that the CR group was being paid less than the PO group for performing work of equal value.

[222] In particular, it can be argued that the Tribunal in this paragraph was diluting the standard of proof when it asked whether it is “more likely than not” that the material is “sufficiently adequate” to enable a “fair and equitable conclusion” to be reached on whether there were wage differences for work of equal value. I do not agree.

[223] In my opinion, this paragraph is not sufficient to rebut the presumption that the Tribunal applied the standard of proof that it stated it was applying. At this stage of its reasons, the Tribunal is merely directing itself on its task of weighing the sufficiency of the evidence in order to reach a “fair and equitable conclusion” on whether there were differences in wages for work of equal value. It was not formulating the standard of proof.

[224] Indeed, in the previous paragraph, the Tribunal had quoted a passage from the reasons of Hugessen J.A. writing for the Court in Department of National Defence, at paragraph 33, where he reiterated that, in proceedings before the Canadian Human Rights Tribunal, a balance of probabilities is the standard of proof , a standard, he noted, which is “a long way from certainty”. In my opinion, it is very unlikely that, in writing in paragraph 412 that the evidence must be adequate to enable a fair and equitable conclusion to be reached on whether there had been a breach of section 11, the Tribunal intended to contradict the statement that it had just quoted on the standard of proof that it must apply.

[225] The words “fair and reasonable conclusion” in paragraph 412 of the Tribunal’s reasons have their origin in the reasons in an earlier pay equity decision, Canada (Public Service Alliance) v. Canada (Treasury Board), 1996 CanLII 1874 (C.H.R.T.) (Treasury Board), which is quoted at paragraph 409 by the Tribunal in the present case. The Tribunal opined in Treasury Board (at paragraph 187) that, since perfect gender neutrality is probably unattainable and pay equity is not susceptible to precise measurement, “one should therefore be satisfied with reasonably accurate results based on what is, according to one’s good sense, a fair and equitable resolution” (emphasis added) of a wage gap between men and women performing work of equal value.

[226] When read in context (including a discussion by the Tribunal of a balance of probabilities as the governing standard of proof), the reference to “a fair and equitable conclusion” in the present case is more akin to a statement of the general goal of those implementing pay equity legislation than to an articulation of the narrower legal question of the applicable standard of proof. This does not, in my opinion, establish that the Tribunal had lost sight of its ultimate task, namely, deciding on a balance of probabilities whether there had been a breach of section 11. In any event, who could disagree that the Tribunal’s aim should be to strive to reach “a fair and reasonable conclusion” to a dispute?

[227] My colleagues also rely on the following paragraphs of the Tribunal’s reasons as indicative of its failure to apply the correct standard of proof to the facts in issue (at paragraph 703):

Accordingly, the Tribunal concludes that it is more likely than not that the aforementioned reasonably reliable Hay Plan, process and job information, in the hands of competent evaluators, as were the Professional Team, would result in reasonably reliable job evaluation values being attributed to the work performed by CR and PO employees.

The Tribunal repeats this conclusion in the following at paragraph 798:

The Tribunal has already concluded that it is more likely than not that the reasonably reliable Hay Plan, process and job information, in the hands of the competent Professional Team, would result in reasonably reliable job evaluation values being attributed to the work performed by CR and PO employees (paragraph [703]). In determining the value of the work performed by those employees, the Professional Team applied the composite of the skill, effort and responsibility required in the performance of the work, and the conditions under which the work was performed, all in line with the requirements of subsection 11(2) of the Act.

[228] It is said that the fact that the evidence before the Tribunal was such as to produce “reasonably reliable job evaluations” is not the same as concluding that on a balance of probabilities the work being compared was of equal value. However, if the evaluation of the jobs was “reasonably reliable” and a substantial portion of the CR group was performing work at least equal in value to the least valuable PO job, I cannot see what else needs to be proved, or what finding made, in order to establish that the wage comparison related to work of equal value. As already noted, the Tribunal accepted the evidence of the professional team that “a significant portion of the CR positions were of a value equal to or greater than that of the PO jobs”: paragraph 799.

[229] The Tribunal’s creation of sub-bands of the reasonable reliability of the evidence on which the job values were assessed has also raised a question as to whether the Tribunal reduced the standard of proof below that of a balance of probabilities. While the elaboration of these “sub-bands” may have been unnecessary, it indicates that the Tribunal was well aware of the limitations of the evidence, and weighed it with great care. As already noted, problems with the evidence resulted, in large part, from the failure of the parties to produce a joint pay equity study evaluating the work performed and determining the wages paid, CHRC’s failure to exercise its powers to require CPC to produce information, and the adversarial context in which the exercise was conducted.

[230] In my opinion, the Tribunal eschewed “reliable” as the standard for evaluating the evidence, and the assessment of the value of the work because it equated “reliable” with “absolute correctness”, the standard proposed by Ms Winter, one of CPC’s witnesses, and properly rejected by the Tribunal. It concluded that, for all practical purposes, such a standard was unattainable, and opted instead for “reasonably reliable” as a standard connoting less than certainty or correctness—a standard which, it rightly said, is not demanded by a balance of probabilities.

[231] The Tribunal does not spell out explicitly what it understands by “reasonably reliable”. However, evidence, or a finding of a fact in issue, can surely only be called “reasonably reliable” if it is more likely than not to be true, regardless of the point on the “reliability spectrum” that particular evidence or the evaluation of a job may be located. “Low-level reasonable reliability” is still “reasonable reliability”. While the Tribunal would clearly have preferred the evidence in a pay equity case to meet an upper “sub-band” of reasonable reliability, it was also of the view that evidence that met only the lower sub-band was still reasonably reliable. It said (at paragraph 698): “Thus, while all three sub-bands meet the test of ‘reasonable reliability’”. The Tribunal made the same point at paragraph 700.

[232] Equally instructive as to how the Tribunal viewed the relationship between “a balance of probabilities” and “lower sub-band reasonable reliability” is its discussion (at paragraphs 919 and 927–930) of the value of the non-monetary components of the wages of the CR and PO groups. While the Tribunal regarded the report of an expert as having only “lower sub-band reasonable reliability”, it nonetheless concluded that the report demonstrated on a balance of probabilities an equivalence between the value of the non-monetary components of the wages of the two groups.

[233] Similarly, having found the evidence, methodology, and process to be reasonably reliable, the Tribunal could infer that on a balance of probabilities the jobs had been properly evaluated. Because a substantial portion of the CR jobs fell within the PO value line, a determination could then be made, again on a balance of probabilities, as to whether the CRs were being paid less than the POs for performing work of equal value contrary to section 11.

[234] Nor am I satisfied that it can be inferred from the Tribunal’s reduction of the monetary award to 50 percent of the wages lost according to the wage gap identified by the professional team that the Tribunal must have believed that the evidence fell short of a balance of probabilities. In my view, it is equally plausible that the Tribunal was satisfied on a balance of probabilities that the CR group was paid less for work of equal value, but was not satisfied that, given the limitations of the evidence and the dispute over the methodology appropriate for measuring the wage gap, it should accept the accuracy of the professional team’s measurement of the wage gap. The 50 percent reduction is better seen, in my opinion, as merely a “rounding down” figure. That the Tribunal chose a reduction of 50 percent rather than, say, 49 percent, seems to me inconsequential as far as the standard of proof being applied is concerned.

[235] CPC also says that another indication that the Tribunal reduced the standard of proof required to evaluate the work, a fact in issue, is its reference to the passage in S. M. Waddams, The Law of Damages, loose-leaf ed. (Toronto: Canada Law Book, 2004) at paragraph 13-30, where the author states that when the amount of a loss is difficult to assess, “the tribunal must simply do its best on the material available”. I do not share this view. The Tribunal says (at paragraph 680) only that the passage in question “may be analogous to what the Tribunal considers to be the spectrum of reasonable reliability” [emphasis added]. The Tribunal was not, in my opinion, thoughtlessly transposing comments on evidential difficulties respecting the calculation of damages to proof of liability. Rather, the Tribunal’s point was simply that its adoption of a “reasonableness” standard of reliability of the evaluations was appropriate, not Ms. Winters’ insistence that nothing less than correctness would suffice.

[236] To summarize, I am not persuaded that CPC has rebutted the presumption that the Tribunal applied the standard of proof, a balance of probabilities, which it clearly identified as the applicable standard. In my view, having found that the professional team had evaluated the jobs reasonably reliably, and having accepted the professional team’s evidence respecting the necessary degree of “overlap” between the job value lines of the CR and PO groups, the Tribunal made the necessary findings of the fact, and concluded on a balance of probabilities that the wages compared were with respect to work of equal value.

ISSUE 3: Did the Tribunal commit a reviewable error in finding as a fact that the CR group was being paid less than the PO group for performing work of equal value?

(i) Overview

[237] Having allowed CPC’s application for judicial review on the issues considered above, the applications Judge did not have to decide whether, if the Tribunal had applied a balance of probabilities standard of proof, it would have committed a reviewable error by concluding that the standard had been met. Hence, PSAC does not address this issue in its memorandum of fact and law.

[238] At the hearing of the appeal, however, both parties dealt at some length with the probative value of the evidence on which the Tribunal based its decision. In view of my colleagues’ conclusion that the appeal must be dismissed on the ground that the Tribunal did not apply the correct standard of proof, I shall endeavour to deal relatively briefly with whether the Tribunal committed reviewable error in its application of the civil standard of proof to the evidence.

[239] CPC submits in its memorandum of fact and law (at paragraph 118) that the Tribunal erred in law in concluding that the evidence satisfied a balance of probabilities standard of proof. I do not agree. Whether a standard of proof has been met is essentially a question of fact, on which the Tribunal is entitled to a high degree of deference. Reviewing findings of fact for unreasonableness precludes the Court from making independent findings of fact, reweighing the evidence, or preferring what it thinks is the better evidence. As long as there was evidence on which the Tribunal could reasonably base its conclusion, the Court’s inquiry is at an end.

[240] Three aspects of the evidence before the Tribunal in this case assist in contextualizing the application of the unreasonableness standard.

[241] First, much of the extensive evidence, both oral and written, regarding the evaluation of the jobs and the measurement of the wage gap is highly technical, controversial, and difficult to assess because it was not of the quality normally seen in pay equity cases where there has been a joint union-management study.

[242] Second, in the course of the more than 400 days of hearings, and the more than two years that the Tribunal took to examine the evidence and produce its reasons, the Tribunal would have acquired an understanding, which no reviewing court can hope to match, of the beguilingly simple principle of equal pay for work of equal value and the dauntingly difficult task of implementing it in the present case.

[243] Third, the Tribunal made important findings of credibility which permeate its factual conclusions, setting out (at paragraph 419) seven criteria it used to “examin[e] the evidence of the expert witnesses” in a very systematic manner.

[244] Thus, the Tribunal was impressed by the evidence of PSAC’s professional team and, in particular, that of its “spokesperson”, Dr. Martin Wolf, who had extensive experience in job evaluations in many different employment settings, including office clerical and blue collar work. He made no bones about the problematic features of aspects of the evidence, especially the job data. For instance, he agreed that the information was not complete and had been gathered at different times. For this reason, Dr. Wolf said, he had adopted a very rigorous approach to his evaluation of the jobs and, when in doubt had erred on the side of valuing a PO job generously, and a CR position conservatively. As a result, he said (at paragraph 487), the professional team’s evaluation of the jobs

… certainly at least meets, and in my opinion probably exceeds, the typical commercial standard, if you will, what consultants from Hay or other consulting firms are doing for their clients.

[245] However, after frankly acknowledging the evidentiary limitations, Dr. Wolf concluded that, in his opinion as an experienced job evaluator, the data were adequate to enable him to provide a professional assessment of the relative values of the jobs in question. Indeed, he testified that the quantity of the information to which he had access exceeded what would normally be available in such an exercise.

[246] The Tribunal found Dr. Wolf to be highly credible, even though, as CPC noted, the professional team had no “hands-on” knowledge of postal work. In contrast, the Tribunal was relatively unimpressed by CPC’s expert witnesses. One, Ms Winter, it found to be rigid and unduly definitive in her opinions; she also seemed to the Tribunal unnecessarily adversarial. It discounted her evidence. The Tribunal noted that the other two experts, Mr. Wallace and Mr. Willis, had not seen all the relevant documents and had not themselves attempted to evaluate the jobs with the data available.

[247] The Tribunal’s findings of the credibility of the various expert witnesses go a long way to explaining why it adopted much of the professional team’s analysis, and not that of CPC’s experts, in reaching its conclusion that CPC was in breach of section 11.

[248] As the Tribunal candidly stated, evaluating the evidence, and the conflicting views of it that the experts provided, presented a considerable challenge. However, it is not the role of the Court conducting a judicial review to probe deeply into the evidence or to revisit the Tribunal’s findings of credibility. It must merely ensure that there was a reasonable basis in the evidence for the Tribunal’s findings.

[249] One final feature of the evidence should be mentioned in order to appreciate the nature of the Tribunal’s task. CPC elected, as was its right, not to adduce before the Tribunal evidence of its own on the value of the jobs. Rather, its experts mostly confined themselves to challenging the work of the professional team and of the other expert witnesses retained by PSAC and CHRC. The Tribunal was thus offered no alternative version of the facts to consider. The only question for it to decide was whether it was satisfied on a balance of probabilities that the professional team had established the value of the jobs and accurately measured any difference in the wages paid to members of the two groups for performing work of equal value.

(ii) Methodology

[250] Dr. Wolf pithily described what all job evaluation plans measure: what you know, what you do, and what you have to put up with. Both PSAC and CPC questioned the suitability of the Hay method evaluation plan as a tool for evaluating the work of the two groups, principally on the ground that, because it tended to put too little weight on the working conditions factor, it was not an appropriate plan for evaluating blue collar or clerical work. There was also debate over the appropriateness of the factor comparison approach to the Hay method used by the professional team.

[251] The evidence before the Tribunal on the methodology issue was that the Hay method evaluation plan was the most widely used job evaluation tool and in its earlier years had been used to evaluate blue collar work. Dr. Wolf stated that he had used it extensively in many different work settings and, in the hands of experienced evaluators, it could be appropriately used in a clerical or blue collar context, such as here, especially with a “strengthened working conditions factor” (at paragraph 563).

[252] Noting (at paragraph 566) Dr. Wolf’s extensive experience with and knowledge of the Hay job evaluation method and of its development over time, the Tribunal concluded (at paragraph 571):

… on a balance of probabilities, the Hay Plan, whether using the factor comparison method or other approaches, is, in the hands of competent evaluators as were the members of the Professional Team, a suitable overall job evaluation scheme which will address the issues of this “pay equity” Complaint in a reasonably reliable manner.

[253] In my view, the Tribunal’s finding that the selected methodology was reasonably reliable was not unreasonable in view of the evidence before it.

(iii) Process

[254] CPC was also critical of aspects of the professional team’s process, that is, its application of the Hay method evaluation plan to the material. In particular, CPC expressed concerns about the lack of an adequate “audit trail” that would enable the team’s work to be monitored, the professional team’s lack of direct experience with postal operations, and a certain lack of discipline in the team’s review of the evaluation results.

[255] The Tribunal concluded that, although not of the quality normally expected in job evaluations produced by a joint union-management study, the process was nonetheless reasonably reliable. First, one member of the Team, Ms Davidson-Palmer, had at one time worked in CPC management and organization development, and therefore had some prior knowledge of the Corporation. Second, the unusual circumstances facing the professional team, especially the litigious environment in which the evaluations had to be conducted and the short period of time available to the team, required it to adjust its normal process. Third, for reasons given earlier, it found the professional team more credible than CPC’s experts. Fourth, in accordance with standard practice, the professional team worked as a unit and made its decisions on the value of jobs either unanimously or by consensus.

[256] On the basis of the material before it, and in light of the reasons that it gave, the Tribunal’s conclusion that the process was reasonably reliable cannot, in my opinion, be characterized as unreasonable.

(iv) Job information

[257] CPC argued that the evidence on which the professional team based its evaluation of the work of the CR and PO groups was so flawed that it could not reasonably support the Tribunal’s conclusion that it was reasonably reliable. Accordingly, it said, the Tribunal’s decision that, on a balance of probabilities, the jobs had been properly evaluated cannot be sustained.

[258] The aspects of the evidence which particularly concerned CPC were: the use of the job fact sheets for gathering information about the CR group; the techniques used for sampling the CR positions; the comparison of 10 generic PO jobs and the 194 actual CR positions; and the fact that the job data for the two groups were not all gathered at the same time.

[259] Two preliminary points bear repeating before I turn to the specific issues raised by CPC. First, it is not the role of this Court to retry the facts; CPC had put to the Tribunal the points outlined above, but the Tribunal had not accepted them. An applicant for judicial review who argues that an administrative tribunal erred in its findings of fact has a heavy burden to discharge: it must establish that there was no evidence on which the tribunal could reasonably base a finding of material fact. Second, a reviewing court should not second guess a tribunal’s reasoned findings of credibility. The credibility findings made by the Tribunal in this case respecting the expert witnesses that it heard in the course of this mammoth hearing are an important part of the basis of its findings of fact.

(a) Job fact sheets

[260] In CHRC’s investigation of PSAC’s pay equity complaint during the years 1984 to 1992, it developed a “Job Fact Sheet”, a questionnaire designed to collect information from members of the PO and CR groups about the nature of their positions. It is common ground that these job fact sheets did not meet professionally accepted standards for evaluating jobs. For example, instead of simply asking employees for information about their jobs, they also asked respondents to evaluate their jobs. Further, the information was intended for analysis by a method other than Hay.

[261] When it became apparent that there were problems with the data collected through the job fact sheets, PSAC retained the professional team to supplement the data and to re-evaluate the jobs. Dr. Wolf agreed that the quality of the job fact sheets was “abominable” and not suitable for evaluating the jobs, largely because employees were asked to evaluate, as well as to describe, their jobs. Accordingly, the professional team conducted interviews with 114 members of the CR group. It re-evaluated the CR and PO positions on the basis of both the data collected through the job facts sheets (excluding the employees’ self-evaluation of their jobs), and the additional information obtained through the interviews and supplied by CPC.

[262] While not glossing over problems with respect to the accuracy, consistency and completeness of the data, Dr. Wolf testified that, based on his extensive experience with the Hay method evaluation plan, he was of the view that the data were adequate to enable the professional team to evaluate the jobs in question.

[263] The Tribunal concluded that, although not of the same quality as the data typically generated by a joint union-management pay equity study, the job information in the present case was nonetheless “reasonably reliable”, but only at the “lower sub-band”. For reasons already considered, the Tribunal regarded Dr. Wolf as a highly credible expert, but was much less impressed with CPC’s experts. The Tribunal noted that the professional team had been prepared to adapt to the deficiencies in the data by, for example discarding job information which it regarded as unreliable. The Tribunal also noted that the professional team’s reconsideration of the evaluations in 2000 in light of the additional information supplied by CPC had, with two exceptions in the PO jobs, little impact on the results that it had reached earlier.

[264] In my opinion, it was reasonably open to the Tribunal on the evidence before it to conclude that the information collected through the job fact sheets, as supplemented by interviews and analysed by experienced professionals, was adequate to enable the work of the CR and PO groups to be evaluated in a reasonably reliable manner.

(b) Sampling techniques

[265] CHRC did not seek job information from all of the approximately 2 300 members of the CR occupational group, nor from all of the much larger PO group. Instead, it sent questionnaires to about 400 members of the CRs and received responses from 194 or 45 percent, and gathered information on the 10 PO generic jobs.

[266] CPC’s witness, Dr. David Bellhouse, an expert in statistics with a specialization in survey sampling, was critical of CHRC’s sampling techniques, pointing out that the sample had originally been drawn by a CHRC officer who lacked relevant expertise, and was not supervised by a suitably qualified person. In Dr. Bellhouse’s opinion, both the design of the CR sample and the low response rate were likely to introduce biases into the results. Mr. Willis was generally supportive of Dr. Bellhouse on this issue.

[267] CHRC called as a witness, Dr. John Kervin, a sociologist with an expertise in data collection and the use and analysis of statistics in the context of industrial relations, including gender bias and pay equity. He testified that, in his opinion, there was no basis for concluding that the sampling was flawed in design or response, or that the results were biased; he found the CR sample to be sufficiently representative for its purpose. He further stated that Dr. Bellhouse had overlooked the pay equity context in which the data were being collected and the qualitative nature of some of the analysis of the jobs. Instead, Dr. Kervin said, Dr. Bellhouse had approached the question strictly from the perspective of a statistician who was seeking scientific accuracy, without regard to the “art” aspect of pay equity inquiries.

[268] After considering at some length the conflicting evidence given by these two experts, who brought to bear somewhat different perspectives on the issues, the Tribunal concluded that Dr. Kervin’s evidence was more germane to the issues before it. It was entitled to accept his evidence.

(c) Comparing PO “jobs” and CR “positions”

[269] It was argued that any comparison of the work of the two groups was invalidated because of differences in what was being compared. In particular, CPC said, job information collected about actual CR positions could not be compared with information collected about the 10 generic or composite PO jobs which, the Tribunal stated (at paragraph 472), were “an amalgam of functions for 10 commonly held job[s]” in the PO group.

[270] CHRC had proceeded on the basis of generic jobs because members of the PO group had not completed the job fact sheets. CPC would not allow them to complete the questionnaires on CPC’s time and their union refused to allow them to be completed, without remuneration, outside work hours. These generic jobs were evaluated through the use of job descriptions and job profiles, some of which may have been outdated or incomplete.

[271] Dr. Bellhouse gave his opinion that because the CR group had been evaluated on the basis of actual positions and a description of the work done by incumbents, and the PO group had been evaluated on the basis of a selection of job titles, a proper comparison of the value of their work could not be made.

[272] Dr. Kervin disagreed. He testified that, while not a random sample, the 10 generic PO jobs were likely to provide a reasonably accurate basis for the purpose of making a pay equity comparison. He also stated that the fact that the information for the PO group was based on job titles, rather than positions was not a significant problem, especially since the job specifications used for the PO jobs were similar in some respects to the CR job fact sheets. Dr. Kervin regarded the disparity between the PO jobs and the CR positions as “a difference in the unit of analysis and not as a difference in measurement” (at paragraph 470), and one that was easily remedied.

[273] Despite this disagreement between the experts, the Tribunal never makes a finding of which view it accepts and why. While the Tribunal stated that it prefers the evidence of Dr. Kervin on the sampling issue, largely, it would seem, because his expertise was more directly relevant to a pay equity context, the Tribunal does not reach a similar conclusion on the “jobs v. positions” issue.

[274] Perhaps the Tribunal comes closest to addressing this issue when it refers (for example, at paragraph 660) to the enrichment of the professional team’s understanding of the content of the PO jobs after it received the additional information from CPC, as outlined in its report of June 2000. Otherwise, one must infer from the Tribunal’s finding that the job information was reasonably reliable that it adopted Dr. Kervin’s view on the “jobs v. positions” issue as well as on sampling.

(d) Timing

[275] CPC argued that the fact that the job information pertaining to the CR and PO groups was not collected at the same time undermined its reliability in a pay equity context. The nature of jobs may change over time, as a result, for example, of the introduction of new technology which may have an impact on the skill required for the job, as well as the working conditions. Mr. Willis testified that a valid comparison should be based on contemporaneous information about the work being performed.

[276] This issue seems not to have been addressed expressly and specifically before the Tribunal by PSAC or CHRC. However, they may have done so indirectly by attacking the credibility of CPC’s experts, Mr. Willis and Mr. Wallace, on the ground that they had neither read all the relevant documentation nor attempted to evaluate the jobs on the basis of the information available.

[277] Nor does the Tribunal itself make a finding on the timing issue. Rather, it seems to have rolled it up in its overall acceptance of the professional team’s analysis, and its relatively less favourable view of CPC’s experts.

(e) Conclusion

[278] To conclude this review of the Tribunal’s finding that the professional team’s evaluation of the jobs of the complainant and comparator groups was reasonably reliable, I acknowledge that the Tribunal does not always explain as fully as it might why it accepted one view of the evidence rather than another, especially on the issue of the timing of the collection of data and the difficulty of comparing PO jobs and CR positions. However, deficiencies in the Tribunal’s reasons in this regard do not, in my opinion, render its decision unreasonable on the ground that it is not sufficiently transparent.

[279] It can be inferred from the Tribunal’s careful and full explanation of the conflicting views of the experts that it understood the issues and appreciated the strengths and weaknesses of each. To the extent that the Tribunal does not make a definitive and reasoned finding on one or more of the issues considered above, the Tribunal can be taken to have adopted the view of the relevant expert and the underlying reasoning. I have already emphasized the importance of the Tribunal’s findings of credibility of the parties’ principal expert witnesses.

[280] Nor am I persuaded that the Tribunal’s decision to uphold PSAC’s pay equity claim is vitiated by its findings of fact. There was, in my opinion, a reasonable basis in the evidence, when viewed overall, to support its conclusions.

(v) Wage gap

[281] Having concluded that the wages of the CRs and the composite or “generic” PO positions could be compared because a substantial portion of the CR positions fell within the PO value line, the Tribunal proceeded to determine if there was a wage gap between the two groups for performing work of equal value. Subsection 11(7) of the Act defines “wages” broadly, so as to include both monetary and non-monetary elements. Thus, after listing specific benefits that are to be included as “wages”, subsection 11(7) contains in paragraph (c) the following “catch-all” provision:

	11. …
Definition of “wages”	(7) For the purpose of this section, “wages” means any form of remuneration payable for work performed by an individual and includes … (c) payments in kind;

[282] The Tribunal thus had to determine three issues with respect to the wage gap: (i) the amount of the monetary component of the wages; (ii) the value of the non-monetary components of the wages; and (iii) the methodology for identifying and measuring any wage gap with respect to work of equal value. Since there appears to have been relatively little dispute on the first issue, I shall focus on the other two.

(a) Non-monetary components of wages

[283] Valuing non-monetary items, such as benefits, raises some difficult technical issues. As with the evaluation of the jobs, difficulties in determining the value of the non-monetary elements of the “wages” also stem from evidentiary problems caused by the absence of a joint study by the parties.

[284] In 1995, PSAC retained Dr. Don Lee, an expert in contract analysis and non-wage compensation valuation, to compare the non-monetary components of the wages from 1983, when the complaint was filed, to 1995. Dr. Lee based his report on a review of 14 collective agreements covering this 12-year period, as well as on a number of employee benefit plans of the federal government that had not been incorporated into these collective agreements. He was able also to conduct a detailed valuation analysis of benefits for 1995, including the extent to which employees had actually used the benefits.

[285] However, in calculating the non-monetary value of benefits, Dr. Lee did not include job security for either group, or the uniform and protective clothing allowances for the POs. He concluded that what differences there were between the value of the benefits of the two groups were either non-existent or minor, and were not significant for pay equity purposes. He discounted as insignificant differences of less than 0.1 percent of wages.

[286] Dr. Lee was unable to conduct such a detailed analysis for the years 1983 to 1994, but simply examined the terms of the collective agreements for those years. He concluded from this examination that any differences in the value of the benefits provided to employees were minor and temporary. He did not think it necessary for this purpose to attempt to obtain from CPC a complete file for each employee. He also stated that, when in doubt, he had overstated the value of the differences favouring the CR group’s non-monetary benefits and understated those of the PO group.

[287] Dr. Lee’s report was challenged by Mr. Robert Bass, an expert in costing compensation, retained by CPC for this purpose. Mr. Bass identified flaws in Dr. Lee’s methodology, which he considered “fatal”. In particular, he argued: the analysis should have been based on benefits provided in 1983, not 1995; it was an error to ignore individual differences in non-monetary benefits of less than 0.1 percent of wages, because several such differences could be cumulatively significant; and the generous job security provisions in collective agreements were sufficiently important that they should have been valued. However, Mr. Bass did not offer his view on the value of the non-monetary element of the employees’ wages, in either 1995 or earlier.

[288] The Tribunal carefully considered these criticisms. First, it tended to agree that, as a matter of theory, it would have been better to use 1983 as the baseline; however, some of the relevant evidence was apparently not available. Second, it noted that Mr. Bass had provided no evidence to indicate what impact the inclusion of a non-monetary difference in benefits of less than 0.1 percent of wages would have had and therefore gave little weight to this aspect of his report. Third, on the basis of the credibility of the two experts and the “nebulous nature of costing job security” (at paragraph 910), the Tribunal concluded that there was not likely to be any significant difference in the value of job security between the POs and the CRs.

[289] The Tribunal found (at paragraph 918) that Dr. Lee’s report showed on a balance of probabilities that the value of the non-monetary benefits for the two groups were equivalent and were “tied, in a negotiated pattern, to the value of the wages paid to the two groups.” The absence of the data needed to make more precise calculations, the Tribunal concluded (at paragraph 919), did not mean that Dr. Lee had failed to establish on a balance of probabilities that the non-monetary component of the wages of the two groups was equivalent. The Tribunal accepted Dr. Lee’s conclusion that, on a balance of probabilities, the value of benefits to which the CR and PO groups were entitled over the period covered by the complaint was generally equivalent. However, the Tribunal concluded that Dr. Lee’s report was “lower band reasonably reliable”.

[290] In my opinion, in view of the findings that Dr. Lee was able to make on the basis of the information available to him, it was not unreasonable for the Tribunal to conclude that if additional data for the years before 1995 were available, it would not reveal that the benefits were significantly more valuable for one group than the other.

(b) Methodologies

[291] The Tribunal was presented by the parties with an array of methodologies for determining the existence and extent of any wage gap. Suffice it to say that PSAC and CPC proposed methodologies that seemed likely to be most favourable to their respective positions. CHRC’s preferred approach seems likely to produce a result between the two extremes. It proposed grouping the sample 194 CR positions into jobs with similar characteristics, which, it said, would make it easier to compare with the “composite” or generic PO positions, which included internal and external, but not supervisory, operational functions.

[292] After describing the rationales provided by CPC for its proposed methodology, the Tribunal opted for CHRC’s, on the ground that it was appropriate in a pay equity context to emphasize the content of the work performed, rather than the definition of the position occupied by the employee. However, it did not accept as conclusive the monetary values provided by the parties, holding that access to individual employee records, in consultation with CPC, was required in order to reach a final conclusion.

[293] In my view, the choice of an appropriate methodology for determining the existence and extent of a wage gap is within the discretion of the Tribunal, and is reviewable for unreasonableness. Given both its technical aspects and the absence of statutory criteria, the Tribunal’s selection is entitled to a high degree of deference. I am not persuaded that only the methodology proposed by CPC can reasonably be said to be consistent with the objectives of section 11 of the Act or that the methodology proposed by CHRC, and adopted by the Tribunal, was unreasonable.

ISSUE 4: Did the Tribunal err in law in awarding PSAC compensation in the amount of half of the CR group’s lost wages according to the identified wage gap?

[294] PSAC applied for judicial review of the Tribunal’s decision to compensate the complainants by awarding them half the amount of the wages that they had lost according to the wage gap indicated by the measuring methodology proposed by CHRC and accepted by the Tribunal. PSAC says that the decision is not supported by the evidence before the Tribunal. The Federal Court dismissed the application for judicial review for mootness, since it allowed CPC’s application for judicial review and set aside the Tribunal’s decision that CPC had been in breach of section 11 of the Act.

[295] The remedial powers of the Tribunal relevant to this appeal are contained in the Canadian Human Rights Act, paragraph 53(2)(c) [as am. by S.C. 1998, c. 9, s. 27], which provides as follows:

53. …

Complaint substantiated

(2) If at the conclusion of the inquiry the member or panel finds that the complaint is substantiated, the member or panel may, subject to section 54, make an order against the person found to be engaging or to have engaged in the discriminatory practice and include in the order any of the following terms that the member or panel considers appropriate: … (c) that the person compensate the victim for any or all of the wages that the victim was deprived of and for any expenses incurred by the victim as a result of the discriminatory practice;

It is common ground that this provision governs the award of compensation for a breach of section 11.

[296] The Tribunal has considerable statutory discretion in fashioning an appropriate remedy. Thus, subsection 53(2) provides that if the complaint is substantiated, the panel or member hearing the matter “may ... make an order … and include in the order” [emphasis added] any of the listed terms that the panel or member “… considers appropriate”. Paragraph (c) provides that the panel may order the person found to have committed a discriminatory practice to “compensate the victim for any or all of the wages” [emphasis added] lost as a result of the discriminatory practice.

[297] Like other discretionary decisions, the Tribunal’s award of compensation is reviewable on a standard of unreasonableness: Dunsmuir, at paragraph 53. While the Tribunal’s reasons provide the principal basis for a reviewing court to determine whether an exercise of administrative discretion is unreasonable, the court may also consider the reasonableness of the outcome: Dunsmuir, at paragraph 47.

[298] The Tribunal held that compensation should be awarded for wages lost between August 24, 1982 (that is, one year after CPC was created and one year before PSAC filed its complaint with CHRC) and June 2, 2002, when the wage increase awarded to the CRs, by CPC, and the implementation of a new job evaluation plan, eliminated any wage gap between the CR group and the PO group.

[299] The Tribunal stated that the objective of an award of compensation under paragraph 53(2)(c) of the Act is to make whole the victims of discrimination. However, it also noted that courts had reduced damages awards in order to take into account uncertainties in determining the precise amount of loss. While there were no uncertainties about future events that could affect the amount of wages already lost by the CR group, the Tribunal came back to its finding that the evaluation of the jobs and the non-monetary component of the wages had met only the “lower sub-band” of reasonable reliability. On this basis, it reduced by 50 percent the amount represented by the wage gap identified by CHRC.

[300] PSAC argues that the Tribunal’s reduction of the compensation was unreasonable. First, it submits, the same data and the same methodology proved both the existence and the extent of a wage gap. Having accepted that the evidence established a wage gap, the Tribunal could not logically find that it did not also establish the extent of the gap. Second, if the Tribunal could factor in uncertainties in the evidence when determining the amount of compensation payable, it had no basis for concluding that the evidence over-estimated, rather than under-estimated, the extent of the actual wage gap. Counsel noted that Dr. Wolf had testified that the professional team had taken the limitations in the evidence into account when evaluating the jobs: when in doubt, they had evaluated a PO position up and a CR position down, and had thus underestimated the extent of the wage gap.

[301] I do not agree. Specialized tribunals are owed a particularly high degree of deference in their exercise of a broad statutory discretion to fashion an appropriate remedy. The Tribunal directed itself correctly in law when it stated that an award of compensation should aim to make the victims whole. However, it was, in my view, also open to the Tribunal to extend by analogy principles used to take into account future uncertainties to uncertainties about the past, and on this basis to reduce the amount of compensation. Indeed, this was done in somewhat similar circumstances where it was uncertain whether a person would have obtained a job if he had not been denied it because of the unlawful discriminatory conduct of the employer: Canada (Attorney General) v. Morgan, [1992] 2 F.C. 401 (C.A.), at page 412.

[302] Nor was it unreasonable for the Tribunal to conclude that, while the evidence was good enough to establish the existence of a wage gap, it was not good enough to measure it precisely. PSAC had the burden of proving on a balance of probabilities both the existence and the extent of any wage gap. Accordingly, if the Tribunal was not satisfied that PSAC had discharged its evidential burden by proving the amount of the wages lost on a balance of probabilities, it could reasonably award less than the amount indicated by the evidence that PSAC had adduced.

[303] The following sentence from the passage in Professor Waddams’ text, The Law of Damages (at paragraph 13-30), is particularly apt in this context:

If the amount [of a loss] is difficult to estimate, the tribunal must simply do its best on the material available, though of course if the plaintiff has not adduced evidence that might have been expected to be adduced if the claim were sound, the omission will tell against the plaintiff.

As I have already noted, neither PSAC nor CHRC was without some responsibility for the state of the evidence.

[304] For these reasons, I am not persuaded that the Tribunal’s award of compensation should be set aside as unreasonable.

D. CONCLUSIONS

[305] For all these reasons, I would allow the appeals of PSAC and CHRC in A-129-08 and A-139-08, set aside the decision of the Federal Court except on costs, and dismiss CPC’s application for judicial review. The Federal Court awarded no costs and I would not disturb that finding. I would award PSAC its costs in the appeal. CHRC has not sought costs and none is awarded. I would dismiss PSAC’s appeal in A-130-08 with costs.