Duty of Procedural Fairness

The common law trigger for the duty of fairness from Cardinal is that a decision is administrative and affects the “rights, privileges or interests of an individual.” [1] This has been applied to a number of fact scenarios and is generally a low bar.[2]  The application of Cardinal to these facts are relatively straightforward.[3] First, LTB decisions are administrative in nature: the LTB is an administrative tribunal born out of the Residential Tenancies Act.[4]  Second, LTB decisions do affect the rights, privileges or interests of both tenants and landlords. For example, LTB may decide whether a landlord’s eviction of a tenant was valid or not, and this decision may affect the living conditions of the tenant or the financial stakes of the landlord.[5]

1. Limitations on Procedural Fairness

For completeness, it is necessary to consider any statutory limitations which may override a duty of fairness.[6] For this, we need to trace the transfer of power. It is important to note that explicit legislation and policy-making can override any duty of procedural fairness.[7] Common law cannot trump statute if it insists on a structural lack of independence.[8] Of course, any statute must be constitutionally compliant and it must be consistent with limitations, such as the Charter or jurisdictional constraints. [9] However, none of these are live issue on the facts.

The LTB is born out of the RTA. There is nothing in the RTA which grants an exception to the duty of fairness, but there are further empowering statutes referenced through the RTA.[10] For example, the chair can establish rule of procedure according to the SPPA, and the SPPA is silent on limits to procedural fairness.[11] Moreover, the power of the RTA is further constrained by the ATAGAA,[12] and ATAGAA is further traced to the PSOA.[13] By and large, these statutes play a role in defining the powers of the chair, but they do not override the duty of procedural fairness.

• Quantum of Procedural Fairness

A useful starting point for determining the quantum of fairness is Baker. The standard of procedural protection depends on what is at stake or how important the decision is to the individual.[14] For the LTB, the stakes are quite high.[15] Furthermore, the procedural rights of the tenants and landlords ought to be robust because the nature of the LTB’s decision is quasi-judicial and final.[16] The rights of the individuals protected by procedural fairness are commonly understood to be the right to be heard (audi alteram partem) and the right to an impartial decision-maker (nemo judex in sua causa).[17] The right to be heard is not a live issue for the LTB, so it is important to limit the focus to the right to an impartial (or unbiased) decision-maker.[18]

The standard for procedural fairness in the right to an impartial decision-maker have two thresholds depending on the facts.[19] The “lower” threshold is that the “the apprehension of bias must be a reasonable one…”[20] This means that there does not have to be actual bias proven in the mind of the decision-maker, but just a reasonable apprehension of bias. The “higher” threshold is whether there is prejudgment or a “closed-mind” incapable of persuasion.[21] The thresholds depend on the function of the administrative body. If the administrative body is adjudicative, then the appropriate test is reasonable apprehension; however, if it is regulatory or based on policy, like a popular election, then the appropriate test is closed-mind.[22] The rationale for a relaxed standard for elected members is that they usually have expressed views on public policy, and this is given more deference.[23] Defining the function of an administrative body is not always straightforward given the “spectrum” of functions.[24]

Sandra was appointed for presumably policy reasons, which suggests the closed-mind test. However, the LTB members engage in quasi-judicial, adjudicative matters, which suggests the reasonable apprehension test. This ambiguity must be explored before arriving at a solution.

Function of the LTB

If we consider Sandra and the LTB members to be not distinct but a single agent of the LTB, then we must consider the problem of overlapping function.[25] Although there is no statutory exception in the RTA which overrides procedural fairness,[26] it is not clear that Sandra’s function in reappointing LTB members generates this problem because Sandra is not involved in the adjudicative process. This point raises the more material issue: the implications of the relationship between Sandra and the LTB members.

Consider first the bright line drawn in Newfoundland Telephone between policy and adjudicative matters.[27] What is remarkable about this case is that it established that the same individual can, in principle, have a policy function at one moment while reasonably appearing to be bias (but not closed-minded), then switch to an adjudicative function and jettison the reasonable appearance of bias – in other words, switching functions clears the slate of past bias. This seems, however, to have a narrow scope in that the switch in function was limited to a single individual. In contrast, Sandra and the LTB members have separate and discrete functions.

1. Sandra’s Bias

It is possible to focus only on Sandra’s reappointment of LTB members and raise the issue of her bias, but this is unlikely to succeed. Recall that Sandra’s appointment is policy-driven and thus the rest for her bias is one of closed-mind. The prejudice on the part of Sandra must be such that attempts to present alternative views are “futile,” and the “appearance of bias will not satisfy the test unless the court concludes that they are the expression of a final opinion on the matter.”[28] That means that, in theory, all of Sandra’s reappointed LTB members can be staunchly committed to seeing gentrification as good and tenants as barriers to certain ends; as long as they have an open mind, they have met their duty of procedural fairness.[29]

On the current facts, all that is establish is that Sandra looks favorably towards gentrification, she reappointed 6 members who are favorable towards landlords, and there is data to buttress the fact that there are favorable remedies for landlords.[30] One possible argument is that, due to her partiality towards gentrification, Sandra cannot be persuaded to appoint any LTB member who is neutral (or more favorable towards tenants).[31] However, on the balance of probabilities, it is most likely not sufficient establish a closed mind; in principle, it is still possible for Sandra to explicitly announce her favoritism towards LTB members who favorable to landlords, and still maintain that she has an open mind.

It is likely that a sufficient pattern can be established on the part of Sandra’s bias towards gentrification, but the problem for this line of analysis is that the closed-mind test is too high a threshold. If we can establish the same bias in the LTB members themselves, we have a lower threshold to meet: a reasonable apprehension of bias.

• Sandra’s Connection to the LTB Members’ Personal Bias

It is clear that the LTB has an adjudicative function.[32] Still, an argument can be mounted on Sandra’s bias, despite her non-adjudicative function.[33] To be sure, the bright line between policy and adjudicative matters in Newfoundland Telephone does not apply to this line of argument because we keep Sandra and the LTB members separate in function. The focus now is the direct (or close) connection between the bias of one appointing a decision-maker and the decision-maker themselves, which gives rise to the question of a reasonable apprehension of bias.[34]

Consider now the possibility of Sandra’s personal bias infecting the downstream decision-making of the LTB members. The argument is that Sandra’s bias connects to the reappointment and composition of the LTB, and this forms a nexus between Sandra’s bias and the bias of the LTB members.[35] The strategy is not analyzing Sandra for bias, rather it should focus on analyzing the LTB members. At this point, however, the mere fact that the composition of the LTB members reflects members who are favorable to landlords is tenuous and likely not sufficient for a reasonable apprehension of bias.[36] The use of this is to buttress the other empirical data to form a cumulative case. The statistic data suggests that Sandra’s biased composition is being realized in the form of personal bias in the LTB members.[37]

The standard of “reasonable apprehension of bias” is not a sharply defined threshold, so the applicability of the empirical statistical data to this test requires further examination.[38] In this case, what we need to establish is that a 37% increase in landlord remedies and a 23% decrease in tenant remedies is sufficient to form a pattern such that a reasonable person would perceive this as bias.[39] In arguing that this data is a pattern that establishes a reasonable apprehension of bias, it is crucial to contextualize this data with our previous point. Without connecting the data to Sandra, the case for bias becomes significantly weaker.[40] Perhaps a reasonable person would say 50% is sufficient for bias, or perhaps 25% – the point is that the exact number is arbitrary, so scrutinizing numbers alone will not strengthen the argument.[41] In order to meet such objections, we must connect the data to a line of explanation originating and ending in bias.[42]

The increased frequency of remedies for landlords and the decrease for tenants reflects a person bias in the LTB members themselves that is traced back to Sandra. This personal bias in the LTB members becomes clearer in virtue of Sandra’s mode of reappointment. Sandra biasedly chose to reappoint LTB members who themselves had the same species of bias, so the composition of the LTB members is directly connected to Sandra’s bias.[43] To reiterate, Sandra’s bias is legally unproblematic given her function, but the same bias in the LTB members is legally problematic given their functions as adjudicators. In this light, the empirical data is a reification of the LTB members’ personal bias.[44] Sandra’s bias is connected to the LTB members’ bias, and the emergent statistical data reflect “a pattern that gives rise to a reasonable apprehension of bias.”[45]

---

[1] Cardinal v. Director of Kent Institution, [1985] 2 S.C.R. 643 at p 653, 24 D.L.R. (4th) 44 [Cardinal].

[2] There is a vast body of case law illustrating a variety of fact scenarios. To list a few: applications for immigration on humanitarian and compassionate grounds (Baker v.  Canada (Minister of Citizenship and Immigration), [1999] 2 S.C.R. 817, 174 D.L.R. (4th) 193 [Baker]); an employer’s dismissal of an employee (Dunsmuir v. New Brunswick, [2008] 1 S.C.R. 190, 2008 SCC 9, 291 D.L.R. (4th) 577.); and professional disciplinary decisions (Green v. Law Society of Manitoba, 2017 SCC 20, [2017] 1 S.C.R. 360, 407 D.L.R. (4th) 573.).

[3] While we are not specifically focusing on a particular decision made by the LTB, we are generalizing the decisions that LTB is empowered to make – more specifically, the power to make decisions insofar as they concern disputes between tenants and landlords. 

[4] Residential Tenancies Act, 2006, S.O. 2006, c. 17 [RTA].

[5] Knight provides further guidance with a three-prong test, which consists of the nature of the decision, the relationship, and the impact of the decision: Knight v. Indian Head School Division No. 19, [1990] 1 S.C.R. 653 at 669-677, 335 D.L.R. (4th) 513 [Knight]. The nature of the LTD is a quasi-judicial institution and the final adjudicator between landlord-tenant disputes. The relationship, or the nature of the decision, is one of adjudication. Finally, the impact is again the interests of the tenants and landlords. The LTD have a duty of fairness towards both landlord and tenant, and there are no statutory limitations on this duty.

[6] Kate Glover Berger, “Chapter 5: The Principles and Practices of Procedural Fairness” in Coleen Flood & Lorne Sossin, Administrative Law in Context (3rd ed) (Toronto: Emond P, 2018) 183-236.

[7] See, for example, Canada (Attorney General) v. Mavi, 2011 SCC 30, [2011] 2 S.C.R. 504.

[8] Ocean Port Hotel Ltd. v. British Columbia (General Manager, Liquor Control and Licensing Branch), [2001] 2 S.C.R. 781, 2001 SCC 52, 204 D.L.R. (4th) 33 [Ocean Port].

[9] Evan Fox-Decent & Alexander Pless, “Chapter 6: The Charter and Administrative Law Part 1: Procedural Fairness” in Coleen Flood & Lorne Sossin, Administrative Law in Context (3rd ed) (Toronto: Emond P, 2018) 237-252. See also Canadian Charter of Rights and Freedoms, s 8, Part 1 of the Constitution Act, 1982, being Schedule B to the Canada Act 1982 (UK), 1982, c 11, s 7.

[10] See, for example, RTA, supra note 4 at ss. 173, 178, 180, or 182.2(2).

[11] See RTA, supra note 4 at s. 176(1)-(2). See also Statutory Powers Procedure Act, RSO2006, c. 17, s. 176(2) [SPPA].

[12] Adjudicative Tribunals Accountability, Governance and Appointments Act, 2009, S.O. 2009, c. 33, Sched. 5 [ATAGAA].

[13] Public Service of Ontario Act, 2006, S.O. 2006, c. 35, Sched. A [PSOA]. The PSOA imposes general rules for public bodies (see ibid at s. 58-59).

[14] Baker, supra note 2 at para 25.

[15] An addition to a landlord’s pecuniary interests, an individual tenant’s place of residence is (arguably) essential and the decision has a significant impact on their wellbeing.

[16] Baker, supra note 2 at para 55.

[17] Van Harten, Gus; Gerald Heckman & David J Mullan. Administrative Law: Cases, Text, and Materials, 7th ed (Toronto: Emond Montgomery, 2015).

[18] I do not argue “institutional” bias here, nor do I think it would have much success on the facts; instead, focus on “personal” bias. The authoritative case for institutional bias is 2747-3174 Quebec: 2747-3174 Québec Inc. v. Quebec (Régie des permis d’alcool), [1996] 3 SCR 919, 140 D.L.R. (4th) 577 [2747-3174 Quebec]. Institutional bias commonly refers to the operating processes or structure being independent. There are established considerations or “indices of independence” which are applied to quasi-judicial bodies like the LTB. For example, security of tenure, fixed renumeration independence, and administrative control: Martineau v. Matsqui Institution, 106 D.L.R. (3d) 385, S.C.J. No. 121, [1980]. Even though the same standard of reasonable apprehension of bias applies, the point is not moot because we develop a cumulative case of bias against the LTB. The general rationale for institutional bias is independence and protection from outside influences. The RTA and its empowering statutes offer some protection on this front and, according to the legislation, Sandra is acting within her capacity. Generally, Canada has relatively weak protection from institutional bias. See Gerald Heckman & Lorne Sossin, “How Do Canadian Administrative Law Protections Measure up to International Human Rights Standards – The Case of Independence” (2005) 50 McGill LJ 193.

[19] I employ the term “lower” to define the point of view of the plaintiff in the sense that she has a less onerous burden of proof in establishing bias; however, this also means that the “standard” for the decision-maker is more stringent.

[20] Quebec (Attorney General) v. Canada (National Energy Board), [1994] 1 S.C.R. 159, 112 D.L.R. (4th) 129.

[21] See Old St. Boniface Residents Assn. Inc. v. Winnipeg (City),[1990] 3 S.C.R. 1170, 75 D.L.R. (4th) 385 [Old St Boniface]. See also Save Richmond Farmland Society v. Richmond (Township), [1990] 3 S.C.R. 1213, 75 D.L.R. (4th) 425 [Richmond].

[22] See Newfoundland Telephone Co. v. Newfoundland (Public Utilities Board), [1992] 1 S.C.R. 623, 89 D.L.R. (4th) 289 [Newfoundland Telephone].

[23] Richmond, supra note 19.

[24] Newfoundland Telephone, supra note 20.

[25] It is possible to arrive at this view because the chair delegates of power to the LTB members pursuant to the PSOA (part 3), as outlined in section 182.3(1) of the RTA. The LTB members only have power insofar as being a part of the same agent as the chair. For an interesting theoretical discussion of agency and law, see Paula J Dalley, “A Theory of Agency Law” (2011) 72:3 U Pitt L Rev 495.

[26] Brosseau v. Alberta Securities Commission, [1989] 1 SCR 301, 57 D.L.R. (4th) 458. Here, Justice L’Heureux-Dubé notes: “If a certain degree of overlapping of functions is authorized by statute, then, to the extent that it is authorized, it will not generally be subject to the doctrine of “reasonable apprehension of bias” per se.”

[27] Supra note 23.

[28] Old St. Boniface, supra note 19 at 1197.

[29] At this point, we might also consider whether Sandra has any conflicts of interest. I do not think this is a fruitful line of argumentation. The test for a conflict of interest is reasonable apprehension of bias. On the facts, the relevant conflict of interest is a financial stake. As required by section 172 of the RTA, the is ostensibly no conflict because she quit her other work and put her investments in a blind trust. Arguably, her limited (and not life or at a set retirement age) may still give rise to a conflict of interest in finances. For example, she may have an interest in gentrifying the city to continue consulting on property developments (her work after 2010 was related to property development projects), which may conflict with her role in appointing LTB members (who may be favorable towards landlords and resultantly promote gentrification and her property development work even after her tenure as the Associate Chair). However, it is doubtful that this would constitute a reasonable apprehension of bias. In this case, the pecuniary interest seems “far too remote.” See Energy Probe v. Atomic Energy Control Board, 15 D.L.R. (4th) 48 at para 31, 28 A.C.W.S. (2d) 310.

[30] I frame the content of the bias specifically in terms of “gentrification,” but a more general label would be “anti-tenant orientation.” It is possible to tease out substantive implications in the differences in these terms, but the argument sketched here uses these terms interchangeably.

[31] Furthermore, her lack of appointments of LTB members who are neutral is evidence of her closed mind. We can support the fact that the 6 reappointed members are more favorable to landlords with the other data (the percent increase in landlord remedies and decrease in tenant remedies).

[32] RTA, supra note 2 at s 1.

[33] Again, the non-adjudicative function implies the closed-mind test and her bias passing this test results in a legally permissible form of bias.

[34] MacBain v. Canada (Canadian Human Rights Commission), 22 D.L.R. (4th) 119, [1985] 1 F.C. 856.

[35] While there might be a line drawn between Sandra’s policy function and the LTB members’ adjudicative function (which correspond to different standards), we should not ignore that the composition of the LTB members is held to the standard of reasonable apprehension of bias.

[36] For a discussion on justification, see Mary Liston, “Chapter 5: ‘Alert, alive and sensitive’: Baker, the Duty to Give Reasons, and the Ethos of Justification in Canadian Public Law” in David Dyzenhaus, The Unity of Public Law (Oxford: Hart P, 2004) 113-142.

[37] Put differently, while the mere fact that the 6 reappointment members were “outliers” in granting remedies to landlords is not alone a sufficient basis for a reasonable apprehension of bias, we can use the statistical data to suggest that the biased is material and emerges problematically in the decision-making process.

[38] Empirical approaches to analyzing judicial or quasi-judicial decision-making is a relatively recent phenomenon. See, for example, Ji Li, “Suing the Leviathan – An Empirical Analysis of the Changing Rate of Administrative Litigation in China” (2013) 10:4 J Empirical Leg Stud 815; Sean Rehaag, “Judicial Review of Refugee Determinations: The Luck of the Draw” (2012) 38:1 Queen’s LJ 1; Robert Danay, “Quantifying Dunsmuir: An Empirical Analysis of the Supreme Court of Canada’s Jurisprudence on Standard of Review” (2016) 66:4 U Toronto LJ 555.

[39]The success of the argument depends both on Sandra’s influence on the composition of the LTB members and the statistical data of the decisions of the LTB members. Again, we should be careful not to conflate the legal standard with a scientific one – while notions of standard deviation, P-values, and the like are important epistemic standard for science, the scientific standards do not correspond to legal standards. For an illuminating discussion on deploying different standards for legal analysis, see Brian Sheppard, “Calculating the Standard Error: Just How Much Should Empirical

Studies Curb Our Enthusiasm for Legal Standards” (2010-2011) 123 Harvard L Rev Forum 92.

[40] For an interesting discussion on contextualizing empirical evidence for legal application, see Dean M Hashimoto, “Justice Brennan’s Use of Scientific and Empirical Evidence in Constitutional and Administrative Law” (1991) 32:4 Boston College L Rev 739.

[41] For example, the reasonable person might accept that bias is not the correct explanation for the numbers, rather it is based on some other variable, like coincidence, landlords bringing stronger claims in light of Sandra, or any number of non-bias related explanations for anti-tenant orientations. There are a number of legitimate criticisms towards empirical analysis in administrative law. For a balanced discussion, see Cary Coglianese, “Empirical Analysis and Administrative Law” [2002] 2002:4 U III L Rev 1111.

[42] For a thorough survey on the application of empirical data to support legal arguments, see Jennifer Robbennolt, Robert M Lawless, & Thomas Ulen, Empirical Methods in Law (New York: Wolters Kluwer Law & Business, 2010).

[43] This point can be clarified with an example. Outside of the courtroom, a judge might have a reasonable apprehension of bias towards people with purple shirts, but we might say that the slate is cleared and the judge’s previous bias ought not to be considered as soon as the judge enters the courtroom and assumes her role. The judge may prudently rid herself of this bias when assuming the role of the judge. In contrast, if we separate the individuals, say, the politician who is also biased towards people with purple shirts then elects a judge because the judge also holds the same bias. It cannot be said that this judge prudently rids themselves of their bias; rather, the judge was chosen because of their biased decision-making.

[44] At this point, one might object to the idea that Sandra’s bias has any relevance to the bias of the LTB members, and instead a hard line must be drawn between Sandra’s bias and the LTB members’ bias. This would take away all the contextualization out of the statistical data on the LTB members (putatively biased) remedies and result in undermining the force of the data. This objection rests on a confusion of Newfoundland Telephone – more specifically, this case should not be read as broadly apply to two separate roles (viz. chair vs. members) with separate functions (viz. policy vs. adjudication). If we understand Sandra and the LTB members to be conceptual separate, then the same “type” of bias (i.e., partiality towards gentrification) is instantiated in separate “tokens” (i.e., first in Sandra in deciding LTB reappointment, but then triggered when the LTB members make their decision). For a precise elaboration on this conceptual distinction, see Charles SS Pierce, “Prolegomeno to an Apology for Pragmaticism” (1906) 16:4 The Monist 492.

[45] Hazelton Lanes Inc. v. 1707590 Ontario Limited, 2014 ONCA 793 at para 71.