A Pluralist Account of Deference and Legitimate Expectations: Pluralism in Action

I will be contributing a chapter, with the working title, “A Pluralist Account of Deference and Legitimate Expectations” to a forthcoming Hart Publishing collection edited by Matthew Groves and Greg Weeks. In the fourth and final section of my draft I apply my pluralist approach to three controversial areas of doctrine: knowledge and reliance, substantive legitimate expectations and enforcement of ultra vires representations. As always, comments are very welcome. Download the draft in its entirety here.

Protecting Legitimate Expectations: Pluralism in Action

Three controversial doctrinal questions form a crucible within which the appropriate interaction of the rule of law, good administration, democracy and separation of powers can be observed. These are, first, the role of knowledge and reliance, second, substantive legitimate expectations and third, expectations based on ultra vires representations. I argue that neither knowledge nor reliance should be a pre-requisite to the enforcement of legitimate expectations. On substantive legitimate expectations, I take the side of the Court of Appeal in Coughlan, arguing in the teeth of strong criticism of that decision that there are indeed good reasons for courts to enforce substantive legitimate expectations, though I also observe that the Court of Appeal probably went too far in one important respect. And as to ultra vires representations, I ally myself with Paul Craig and others in advocating an approach that would permit courts to enforce them in some circumstances.

Knowledge and Reliance

Does an individual need to know of and rely on the official action said to give rise to a legitimate expectation? In respect of knowledge it has been suggested, for instance, that “[a] person cannot lose an expectation that he or she does not hold”.[1] And as to reliance, Binnie J. would have refused to protect substantive legitimate expectations in the absence of a need to demonstrate detriment: “One would normally expect more intrusive forms of relief to be accompanied by more demanding evidentiary requirements”.[2] The objection here seems to be that in the absence of knowledge and reliance there is little or no imperative for judicial intervention because the individual’s dignity and autonomy interests have not been compromised by a broken promise. I suggest, however, that neither knowledge nor reliance should be required.

To begin with, good administration supports a requirement that policies and decisions be public and that they should be departed from only where there is a “reasoned justification”[3] to do so. Personal knowledge of and reliance on a policy or decision is not a pre-requisite:[4] “When a legitimate expectation arises from an agency’s past practice, or non-statutory procedural guidelines, it serves to preclude procedural arbitrariness, not the actual expectation of the individual who may have been unaware of its existence”.[5] Although rule-of-law concerns about the non-performance of personalized promises to individuals are absent, good administration points to the desirability of judicial enforcement even where there has been no knowledge or reliance.

But, of course, the force of knowledge and reliance in a particular case may vary. For instance, an individual whose legitimate expectation is based on “a policy which itself is liable to change” cannot expect the expectation “to survive a policy change”.[6] That a policy may be changed at some point in the future is inherent in the policy’s very existence:[7] it follows inexorably from the decision-maker’s appointed role as the regulator of the portion of the administrative state assigned to it by the legislature that it should be able to change policy to suit changed circumstances. Knowledge of and reliance on a policy does not generate expectations as forceful as those engendered by personalized representations, for the rule of law concerns are less weighty.[8] Indeed, good administration counsels against imposing a high threshold for policy changes, for fear of “ossify[ing] administrative policy”.[9]

A better way of capturing the greater moral force of knowledge and reliance – which comes from the rule of law’s concern for dignity and autonomy – is to place more exacting requirements on the decision-maker who goes back on her word than on the decision-maker who changes a generally applicable policy. Once this point is appreciated, the force of Binnie J.’s reason for rejecting substantive legitimate expectations is greatly reduced: although the evidentiary requirements are lower, the burden on the decision-maker who wishes to change a policy will be lighter than the burden on the decision-maker who wishes to resile from a personalized representation. Less exacting requirements for policy changes protect good administration by holding administrative decision-makers to their published policies until such time as they are revised, reflect the relative lack of rule-of-law imperatives for heightened protection and preserve a higher degree of decisional autonomy for the legislature’s chosen regulator. Conversely, where a personalized promise places a heavier justificatory burden on a decision-maker, this is offset by the more demanding evidentiary requirements placed on the applicant and, in any event, does not fatally compromise the decision-maker’s authority to resile from a promise in appropriate circumstances, as the decision can, in principle, be justified.

In summary, the absence of rule-of-law concerns does not mean legitimate expectations can never be enforced without proof of knowledge and reliance, because the value of good administration may also be engaged and respect for democracy and separation of powers ensures that administrative decision-makers’ freedom to change policy in appropriate circumstances is preserved.

Substantive Legitimate Expectations

Enforcement of procedural legitimate expectations does no violence to democracy or separation of powers: “Judges feel far more comfortable telling public bodies what procedures they should follow rather than what outcomes they should procure”.[10] As Laws L.J. explained in R. (Niazi) v. Secretary of State for the Home Department, in this sort of case “the unfairness or abuse of power which the court will check is not merely to do with how harshly the decision bears upon any individual”.[11]

When it comes to substantive legitimate expectations, however, enforcement by courts is more controversial. At times, courts and commentators have claimed that democracy and separation-of-powers concerns have such force as to defeat substantive legitimate expectations outright:

A significant weakness in the analysis of…academic supporters of substantive legitimate expectations is the failure to acknowledge that the doctrine clearly narrows the freedom of the executive government and, more importantly, the effect that this may have on the relationship between the judicial and executive arms of government.[12]

As an Australian judge put it, if legitimate expectations could be given substantive force, “the notion would become a stalking horse for excesses of judicial power”, liable to set courts “adrift on a featureless sea of pragmatism”.[13]

But this represents a drastic denial of the agency of individuals who respond to and rely on promises made by officials. It would give individualized representations designed to induce action or expectation no effect whatsoever. It would permit governments to make cynical promises. Officials could instrumentalize the citizenry, treating individuals as means rather than ends. Treating reliance by individuals on official representations as entirely irrelevant to the lawful exercise of state power does great violence to the rule of law by making it more difficult for individuals to plan their affairs: “to break a promise is to directly interfere with the liberty of the person or persons who have relied on that promise”.[14]

In addition, permitting separation of powers and democracy to predominate would prevent courts from drawing distinctions between different types of expectation. On this approach, all engagements are equally unenforceable: the engagement made in a general policy document is treated no differently from the solemn promise with “the character of a contract”[15] offered to a citizen. Ruling out substantive legitimate expectations altogether would deprive jurists of the opportunity to hone judicial doctrine so as to make appropriate distinctions between different types of expectation. It discounts completely the “greater moral force” of individualized representations relative to general policies.[16] To treat policies as “a type of promise…distorts the nature of policies and promises”; only the latter “creates a duty or obligation to someone to act in some way”.[17]

There is a better way of harmonizing the values in such a way that the decision-maker’s primacy in policy matters is respected and courts respect the limits of the judicial role. Concerns for democracy and separation of powers need not necessarily exclude the possibility of judicial protection of substantive legitimate expectations. The goal of protecting substantive legitimate expectations “is not to impede executive activity but to reconcile its continuing need to initiate or respond to change with the legitimate interests or expectations of citizens or strangers who have relied, and have been justified in relying, on a current policy or an extant promise”.[18] Indeed, procedural protection will often not be enough: “a hearing will be of no value to the applicant if the authority in making its decision is free to ignore the legitimate expectations it has previously aroused”.[19]

Recognizing substantive legitimate expectations does not place an insurmountable barrier before an administrative decision-maker that wishes to resile from a promise or change a policy; it simply imposes an additional constraint in the form of a reason for the final decision that takes adequate account of the initial promise or policy: “It recognises the primacy of the public authority both in administration and in policy development but it insists, where these functions come into tension, upon the adjudicative role of the court to ensure fairness to the individual”.[20]

Moreover, this additional constraint can be tailored to the strength of the expectation created by official action: “The facts of the case, viewed always in their statutory context, will steer the court to a more or less intrusive quality of review”.[21] Where there has been an individualized representation on which there has been detrimental reliance, the burden on a decision-maker who refuses to give effect to the representation is greater than where an individual seeks to rely on the continuance in force of an official policy that the decision-maker wishes to adapt to new circumstances. So, in Coughlan, where residents of a home for severely disabled patients were promised a ‘home for life’ that the public authority subsequently sought to close, it was appropriate to require “[s]trong reasons…to justify resiling from [the] promise”.[22]

In at least one respect, however, the Court of Appeal in Coughlan went too far by asserting “that it is for the court to decide in an arguable case whether such a judgement, albeit properly arrived at, strikes a proper balance between the public and the private interest”;[23] it adjudged the authority at fault for having “failed to weigh the conflicting interests correctly”.[24] But this is to substitute judicial judgement on the merits for the considered wisdom of the administrative decision-maker identified by the legislature.[25] It is not for courts to determine “fairness of outcome” or the “propriety” of the balancing exercise undertaken by the decision-maker.[26] Separation of powers suggests that the reviewing court has a distinct role, that of checking the legality of executive action. On a pluralist approach, some effect must be given to democracy and separation of powers. A legitimate expectation should not be treated as sacrosanct; rather, the strength of the expectation would require more by way of explanation from the decision-maker seeking to resile from its previous representation.[27] So, “where the representation relied on amounts to an unambiguous promise; where there is detrimental reliance; where the promise is made to an individual or specific group; these are instances where denial of the expectation is likely to be harder to justify…”[28] Conversely, an applicant will face a “steeper climb” in cases where a policy change adopted in the public interest has a general effect on a diffuse group of individuals.[29] In any event, what matters is whether the justification provided by the decision-maker is reasonable or rational, not whether the court would prefer to have struck the balance differently.

Courts should generally be less willing to constrain administrative decision-making in the “macro-political” realm.[30] Generalist judges are poorly placed to pull on the spider web of carefully balanced interests woven by an expert decision-maker as it might unravel in an unfortunate way.[31] Deference has its place – especially in the macro-political realm – but it cannot permit administrative decision-makers to ride roughshod over legitimate expectations they themselves created.

Ultra Vires Representations

The conventional view is that ultra vires representations can never give rise to legitimate expectations: “there can only be a legitimate expectation founded on a lawful representation or practice”.[32] Unlawful representations cannot create enforceable rights, interests or obligations, as the Supreme Court of Canada recently explained in respect of a local authority that had long permitted a company to operate a commercial parking lot in violation of a zoning provision: “The authorization by a municipal employee or elected official of a use that violates a provision of a by-law cannot create rights or oust the applicable standards set out in the by-law”.[33]

Here, the democratic principle, as expressed through statutory language, dominates. In particular, courts have concluded, administrative bodies should not be able to expand the limits of their statutory mandates by making representations that are beyond their powers:[34] “Any doctrine of estoppel in that context would threaten to undermine the doctrine of ultra vires by enabling public authorities to extend their powers both de facto and de jure by making representations beyond power, which they would then be estopped from denying”.[35] It has even been suggested that the rule of law could be threatened by the enforcement of ultra vires representations: “When the Government is unable to enforce the law because the conduct of its agents has given rise to an estoppel, the interest of the citizenry as a whole in obedience to the rule of law is undermined”.[36]

Nonetheless, there are examples of judicial concern about unfairness created by non-enforcement of ultra vires representations. Thus in Stretch v. West Dorset District Council, Peter Gibson L.J. had “considerable sympathy” with the individual defeated by a local authority’s ability to assert the illegality of its own stance (an “unattractive” proposition),[37] and in Rowland v. Environment Agency, May L.J. concluded that there was no legitimate expectation with “undisguised reluctance” at the “unjust” outcome.[38]

However, a blanket rule of non-enforcement of ultra vires representations clashes with the rule of law and good administration values for the protection of legitimate expectations. Great unfairness may be done to an individual who relied on an official representation whose validity he or she had no reason to doubt. Accordingly, it is doubtful that the rule of law is better protected by requiring losses caused by ultra vires representations to fall on individuals.

Good administration may also be imperilled. Requiring individuals to interrogate administrative officials at length about the scope of the latter’s authority, or engaging a legal representative to do so, is not a recipe for the efficient exercise of scarce resources. Indeed, individuals may be dissuaded from dealing honestly – or at all – with administrative decision-makers if they cannot trust the words of officials. And even though one might want to guard against the deliberate expansion of powers accorded by the legislature, causing the loss to fall on an individual is “misdirected”: “In the rare cases of intentional expansion of power it strikes at the wrong person, the innocent representee, rather than the public officials. In the more common case of careless, or inadvertent, extension of power any deterrent effect upon the public officer will be minimal”.[39]

Moreover, the democratic principle is not necessarily compromised by the enforcement of all types of ultra vires representations. An Irish judge has suggested that there might be some cases where there may be “greater scope” to enforce an ultra vires representation which involves “something more marginal than the essence of the [administrative] action which is impugned – for example, its scope of extent rather than its nature…”[40] One might add that the fact that an official would have made the representation in the first place surely indicates – absent proof of malice or bad faith – that the official believed the representation furthered the administrative decision-maker’s statutory mandate. Indeed, in jurisdictions that use a holistic reasonableness test for invalidity rather than the ‘all-or-nothing’ ultra vires principle, a blanket rule against the enforcement of ultra vires representations is particularly inapt, for there is a good chance that a statutory provision will ultimately be interpreted in conformity with the stated views of the administrative decision-maker.[41]

Professor Craig has advocated for a balancing test: “Where the harm to the public would be minimal compared to that of the individual, there is good reason to consider allowing the representation to bind”,[42] by focusing on “more sensitive ways of reconciling the needs of…innocent individuals and the requirements of the public body”.[43] A pluralist approach leads to a similar conclusion. Enforcement of ultra vires representations should not be barred outright. Rather, courts should examine the interplay of rule of law, good administration and democratic considerations to determine whether, on the facts, an ultra vires representation should be enforced, aware that in some cases “[t]he public interest may be better served by honouring [the] undertaking than by breaking it”.[44] That is, by acknowledging “that legality is not absolute, but rather an important aspect of the legal system, the consequences of which are subject to important countervailing considerations”.[45] Due to democracy and separation of powers concerns, courts should not engage in an independent balancing exercise that pays no heed to the decisional autonomy of the administrative decision-maker under review, but as in the area of substantive legitimate expectations should assess the rationality or reasonableness of any failure to give effect to an official representation, unencumbered by a dogmatic refusal to countenance the enforcement of ultra vires representations.


[1] Minister for Immigration and Ethnic Affairs v. Teoh (1995), 128 A.L.R. 353, at p. 383, per McHugh J.

[2] Mount Sinai Hospital Center v. Quebec (Minister of Health and Social Services), [2001] 2 S.C.R. 281, at para. 31.

[3] R. (Nadarajah) v. Secretary of State for the Home Department, [2005] EWCA Civ. 1363, at para. 38, per Laws L.J.

[4] See R. v. Department of Education and Employment, ex parte Begbie, [2000] 1 WLR 1115, at para. 101, per Sedley L.J.: “in cases where the government has made known how it intends to exercise powers which affect the public at large it may be held to its word irrespective of whether the applicant had been relying specifically upon it”. Indeed, in the foundational case of Attorney-General of Hong Kong v. Ng Yuen Shiu, [1983] 2 A.C. 629, the individual concerned had no personal knowledge of the announcement that gave rise to his enforceable legitimate expectation of a procedural protection: see p. 635. See also Heatley v. Tasmanian Racing and Gaming Commission (1977), 14 A.L.R. 519; Keogh v. Criminal Assets Bureau, [2004] 2 I.R. 159; Minister for Immigration and Ethnic Affairs v. Teoh (1995), 128 A.L.R. 353.

[5] Apotex Inc. v. Canada (Attorney General), [2000] 4 F.C.R. 264, at para. 122, per Evans J.A. See similarly, Haoucher v. Minister for Immigration, Local Government and Ethnic Affairs (1990), 19 A.L.R. 577, at p. 590, per Toohey J.

[6] R. v. Minister for Agriculture, Fisheries and Food, ex parte Hamble (Offshore) Fisheries Ltd., [1995] 2 All E.R. 714, at p. 729, per Sedley L.J.

[7] See e.g. Thamotharem v. Canada (Minister of Citizenship and Immigration), [2008] 1 F.C.R. 385.

[8] See also Søren J. Schønberg, Legitimate Expectations in Administrative Law (Oxford University Press, Oxford, 2000), at p. 17: “Reliance is only relevant in so far as it strengthens the Rule of Law justification for protection of legitimate expectations”

[9] Paul Craig, “Legitimate Expectations: A Conceptual Analysis” (1992), 108 Law Quarterly Review 79, at p. 90.

[10] Iain Steele, “Substantive Legitimate Expectations: Striking the Right Balance?” (2005) 121 Law Quarterly Review 300, at p. 300. See similarly, Mark Elliott, “Legitimate Expectations and the Search for Principle”, [2006] Judicial Review 281, at p. 288.

[11] [2008] EWCA Civ. 755, at para. 30.

[12] Matthew Groves, “Substantive Legitimate Expectations in Australian Administrative Law” (2008), 32 Melbourne University Law Review. 470, at p. 492.

[13] Attorney General of New South Wales v. Quin (1990), 93 A.L.R. 1, at p. 27, per Brennan J.

[14] Jack Williams, “Clarity and Ambiguity: A New Approach to the Test of Legitimacy in the Law of Legitimate Expectations” (2010), 30 Legal Studies 633, at p. 641.

[15] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 59.

[16] Iain Steele, “Substantive Legitimate Expectations: Striking the Right Balance?” (2005) 121 Law Quarterly Review 300, at pp. 302-303.

[17] Farrah Ahmed and Adam Perry, “The Coherence of the Doctrine of Legitimate Expectations” (2014), 73 Cambridge Law Journal 61, at p. 66.

[18] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 65. Emphasis added.

[19] Christopher Forsyth, “The Provenance and Protection of Legitimate Expectations” (1988), 47 Cambridge Law Journal 238, at p. 253.

[20] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 70.

[21] R. v. Department of Education and Employment, ex parte Begbie, [2000] 1 WLR 1115, at para. 80, per Laws L.J.

[22] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 86.

[23] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 52. Emphasis added.

[24] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 89. Emphasis added. See similarly R. (Niazi) v. Secretary of State for the Home Department, [2008] EWCA Civ. 755, at para. 35.

[25] Philip Sales and Karen Steyn, “Legitimate Expectations in English Public Law: An Analysis”, [2004] Public Law 564, at p. 590-591; Matthew Groves, “Substantive Legitimate Expectations in Australian Administrative Law” (2008), 32 Melbourne University Law Review. 470, at p. 480.

[26] R. v. North and East Devon Health Authority, ex parte Coughlan, [2001] Q.B. 213, at para. 71. See also the formulation at para. 82.

[27] R. (Bibi) v. London Borough of Newham, [2001] EWCA Civ. 607, at para. 39.

[28] R. (Nadarajah) v. Secretary of State for the Home Department, [2005] EWCA Civ. 1363, at para. 69, per Laws L.J.

[29] R. (Nadarajah) v. Secretary of State for the Home Department, [2005] EWCA Civ. 1363, at para. 69, per Laws L.J.

[30] R. (Nadarajah) v. Secretary of State for the Home Department, [2005] EWCA Civ. 1363, at para. 69. See also Attorney General of New South Wales v. Quin (1990), 93 A.L.R. 1, at pp. 12, per Mason C.J.

[31] See generally Paul Daly, A Theory of Deference in Administrative Law: Basis, Application and Scope (Cambridge University Press, Cambridge, 2012), chapter 3.

[32] Rowland v. Environment Agency, [2003] EWCA Civ. 1885, [2005] Ch. 1, at para. 81, per Peter Gibson L.J.

[33] Immeubles Jacques Robitaille inc. v. Québec (City), [2014] 1 S.C.R. 784, at para. 25, per Wagner J.

[34] See e.g. Re Green Dale Building Co., [1977] I.R. 256.

[35] Minister for Immigration, Local Government and Ethnic Affairs v. Kurtovic (1990), 92 A.L.R. 93, at p. 108, per Gummow J.

[36] Heckler v. Community Health Services, 467 U.S. 51 (1984), at p. 60, per Stevens J.

[37] [1998] 3 E.G.L.R. 62, at p. 67.

[38] [2003] EWCA Civ. 1885, [2005] Ch. 1, at para. 100.

[39] Paul Craig, “Representations by Public Bodies” (1977), 93 Law Quarterly Review 398, at p. 420.

[40] Ashbourne Holdings v. An Bord Pleanàla, [2003] 2 I.R. 114, at p. 137, per Hardiman J.

[41] See e.g. McLean v. British Columbia (Securities Commission), [2013] 3 S.C.R. 895, at para. 38, per Moldaver J.

[42] Paul Craig, Administrative Law, 7th ed. (Sweet and Maxwell, London, 2012), at p. 711.

[43] Paul Craig, “Legitimate Expectations: A Conceptual Analysis” (1992), 108 Law Quarterly Review 79, at p. 89.

[44] R. v. Liverpool Corporation, ex parte Liverpool Taxi Fleet Operators’ Association, [1972] 2 Q.B. 299, at p. 308, per Lord Denning M.R.

[45] Søren J. Schønberg, Legitimate Expectations in Administrative Law (Oxford University Press, Oxford, 2000), at p.104.


This content has been updated on October 23, 2015 at 11:04.