When Do Guidelines Bind? An Analytical Framework: Part 1

I have written a paper entitled, “How Binding are Binding Guidelines? An Analytical Framework” to be published later this year in Canadian Public Administration. You can access a pre-publication version here. I will be posting the paper in parts over the coming weeks. Here are the introductory sections:

Guidelines play an important role in contemporary public administration. Since Sir Robert Megarry VC (1944) first identified “quasi-legislation” in the 1940s, the use of guidelines has only expanded. Guidelines are now pervasive, found at every level of public administration (Rawlings 2015).

Despite the propagation of guidelines in recent decades, their legal effects are often difficult to classify. Clearly, guidelines are neither legislation nor delegated or subordinate legislation, but they are nonetheless designed to influence people’s behaviour. Guidelines are a type of “soft law”: “soft” because they do not have the mandatory character of statute or regulation; but “law” because they are intended “to influence or control behaviour” (Creyke and McMillan 2008: 383). Sometimes they have more influence or control than the “actual rules” (Thomas 2022: 73).

Understandably, distinguishing guidelines which have binding force from guidelines which do not have binding force is particularly challenging. Indeed, it might be objected that the very proposition that guidelines can be “binding” is a misnomer. Are “binding” guidelines simply just a “form of law”? (Bell Canada v. Canadian Telephone Employees Association, [2003] 1 SCR 884 (SCC): 37). This objection highlights the fact that nominally “non-binding” guidelines, or “soft law” might, despite the label, be “binding” or “hard law”; they may impermissibly “harden into a rule” (Daly, 2021: 59). If so, however, it is necessary to have criteria at hand to determine when a guideline has crossed the Rubicon and become binding. The challenge cannot be avoided. The need to distinguish binding guidelines from non-binding guidelines arises for several reasons:

  1. Binding guidelines cannot be promulgated without statutory authority. Sometimes decision-makers are explicitly authorized to promulgate binding guidelines (Friends of the Oldman River Society v. Canada (Minister of Transport), [1992] 1 SCR 3 (SCC)), but in other situations their authority is limited to the promulgation of non-binding guidelines. Promulgating binding guidelines would therefore be unlawful, as in Ainsley Financial Corp v. Ontario (Securities Commission),[1994] 121 DLR (4th) 79 (Ont. CA).
  • The application of the principle that a decision-maker may not fetter a discretionary power will differ as between binding and non-binding guidelines. Where a decision-maker has been given a discretionary power – that is, a power to choose between alternatives – the decision-maker must presumptively exercise the power on a case-by-case basis. Guidelines binding a decision-maker on the substance of the decision to be taken will violate the no-fettering principle; non-binding guidelines will not (Thamotharem v. Canada (Minister of Citizenship and Immigration), [2008] 1 FCR 385 (FC)). Guidelines may only bind a decision-maker as to matters of procedure (Canada (Attorney General) v. Mavi, [2011] 2 SCR 504 (SCC): 68).
  • Binding guidelines can be “law” as a matter of Canadian constitutional law. If binding guidelines infringe the rights protected by the Charter of Rights and Freedoms, they will be treated as limitations “prescribed by law” and will have to be Charter—compliant. Furthermore, pursuant to Canada’s supremacy clause – s. 52(1) of the Constitution Act, 1982 – binding guidelines will be “law that is inconsistent with the provisions of the Constitution” and, accordingly, “of no force or effect” to the extent of the inconsistency (Greater Vancouver Transportation Authority v. Canadian Federation of Students — British Columbia Component, [2009] 2 SCR 295 (SCC)).
  • Whether a guideline is binding or not will influence the availability of judicial review. The justiciability of a guideline may depend on its ability to produce legal effects (Martineau et al v. The Matsqui Institution Inmate Disciplinary Board, [1978] 1 SCR 118 (SCC): 129; Ishaq v. Canada (Citizenship and Immigration), 2015 FC 156).i

For these reasons, it is necessary to be able to distinguish binding guidelines from non-binding guidelines. Judges exercising the supervisory jurisdiction, drafters crafting guidelines and administrative decision-makers applying guidelines all need to know when and whether guidelines are binding.

Commentators have noted that this is an important question, which remains without an answer (Keyes 2002; Sossin and Pottie 2005; McHarg 2017). Missing from contemporary public law is “a cohesive framework with which to characterize and respond to soft law instruments” (Sossin and van Wiltenburg, 2021: 626). Accordingly, it is difficult to predict when a soft law instrument will be treated as binding, even when the instrument addresses similar subject-matter. In one high-profile Supreme Court of Canada case, the manual used internally by customs inspectors who systematically confiscated gay pornography they viewed as obscene was described as a mere “internal administrative aid” that was “not law” (Little Sisters Book and Art Emporium v. Canada (Minister of Justice), [2000] 2 SCR 1120 (SCC): 85); but a few years later policies regarding advertisements on public transport were characterized by the same court as “rules that establish the rights of the individuals to whom they apply” and thus “law” subject to review for compliance with the Charter (Greater Vancouver Transportation Authority v. Canadian Federation of Students — British Columbia Component, [2009] 2 SCR 295: 72). In both cases, freedom of expression was at stake and in both cases a soft law instrument was close at hand, but in only one case did the instrument have binding quality: the reasons offered by the judges were for the most part conclusory and certainly did not illuminate the reader as to why the cases were resolved differently.

Questions about bindingness lead to litigation. A useful recent example comes from the Patented Medicines Prices Review Board: new guidelines on the calculation of maximum prices were challenged in part because they were binding (Smart & Biggar, 2020). The litigation was eventually settled, but not without significant sparring between the parties. Analytical uncertainty breeds litigation, and when litigation is ongoing, policy decisions may be held in abeyance. Accordingly, developing an analytical framework for determining when guidelines are binding would help to reduce the possibility of litigation, both by informing governmental actors about how to draft non-binding guidelines and by giving potential challengers information about the likelihood of success of a legal challenge.

Moreover, this uncertainty often influences decision-making within administrative agencies. There is admittedly little or no public information about how soft law instruments are formulated. In my experience regulators often debate questions of bindingness internally and with stakeholders. Sometimes (and possibly quite often), external stakeholders, such as industry representatives, or government departments, are successful in resisting the adoption of new soft law instruments on the basis that the regulator’s proposal is binding and, thus, unlawful. These negotiations take place in the shadow of the law, where the contours of the law are, furthermore, difficult to trace.

Greater analytical clarity would, therefore, be of value to the legal community, both the judges who sometimes have to review guidelines, and (perhaps more importantly) the individuals within government who are asked to draft and design guidelines. Drafters and designers of guidelines could use this analytical framework on a regular basis. For courts, my analytical framework will be useful in the four instances identified above, though in other instances, such as where an individual challenges an administrative decision based in whole or in part on a guideline, the framework is less likely to be relevant.

There is significant debate in the literature about the appropriateness of guidelines. On the one hand, it has been said that they are “crucial to good public administration” (Daly and Fournier 2021: 292), a means of “allowing sufficient discretion to do justice in individual circumstances and requiring sufficient structure that exercises of discretion are broadly predictable is integral” (Weeks 2016: 17). Professor Davis considered that guidelines could confine — through the development of “meaningful standards” — structure — through the development of “plans, policy statements and rules” — and check — through internal “supervision and review” — discretionary powers (Davis 1969: 55). In this way, decision-makers could “move from vague or absent statutory standards to reasonably definite standards, and then, as experience and understanding develop, to guiding principles, and finally, when the subject-matter permits, to precise and detailed rules” (Davis 1969: 219). Most contemporary commentators, similarly, take the view that guidelines have “salutary benefits” (Daly 2021: 58).

On the other hand, however, commentators have noted the absence of “a fair and accountable process of development or meaningful forms of public review” of guidelines (Sossin and Smith 2003: 887; Sossin 2004; Weeks 2018). This sort of “back-door regulation”, it has been suggested, “gives too much discretion to regulators, sets higher compliance standards than are required by law, leads to uncertainty and litigation, causes confusion for small business and consumer organisations, and imposes higher costs on industry” (Creyke and McMillan 2008: 387—388). Indeed, rigid policies lead to sham hearings where individuals have no realistic opportunity to contest the application of the policy in question: rigid policies, therefore, “are disrespectful, disingenuous and wasteful; they also undermine the values of dignity and self-defence” (Perry 2017: 397). In short, “the unregulated use of rules contains considerable potential for unfairness and abuse of power” (McHarg 2017: 298).

This paper will not explore such first-order normative questions, rather it will focus on how one should decide whether a guideline is binding or non-binding, in an attempt to trace the boundaries of permissible use of guidelines. If it is possible to provide an answer to the unresolved question of bindingness, however, the answer should assist in resolving first-order normative questions. The greater the scope for binding guidelines, the greater the normative concerns about the use of such soft law instruments; but if guidelines are only rarely binding, normative concerns about their use are much less pressing. Being able to distinguish binding from non-binding guidelines will therefore bring much needed analytical clarity to normative debates.

To distinguish binding guidelines from non-binding guidelines, I have developed an analytical framework consisting of three indicia of bindingness:

  1. Imperative Language: guidelines which are couched in mandatory terms are more likely to be binding; guidelines which state that they “may” be applied or be considered or are explicitly said to be “non-binding” are less likely to be binding.
  2. Detail and Precision: guidelines which set out detailed and precise expectations are more likely to be binding; guidelines which are less detailed and precise and thus require an exercise of judgment are less likely to be binding.
  3. Intended effects on Third Parties: guidelines which are designed to direct the behaviour of third parties are more likely to be binding, especially when the guidelines are expressly linked to an administrative decision-maker’s enforcement powers; guidelines which are internal in nature and affect only the in-house operations of an administrative decision-maker are less likely to be binding.

The question of bindingness is, invariably, binary. Either a guideline (or a part of a guideline) is binding, or it is not, and various legal consequences will follow. The indicia are designed to answer this binary question. As will be immediately obvious, each of these indicia may pull in different directions in different cases. Furthermore, none of these indicia is determinative on its own. Rather, the relative force of the indicia will help to determine whether a particular guideline is binding or non-binding.

These indicia are directed at the drafting and design of guidelines, not at empirical evidence of how they operate in practice. Put another way, whether language is actually treated as imperative, whether language is actually treated as detailed and precise, and whether language actually has effects on third parties are not questions addressed here. If there is convincing evidence that a guideline does in fact bind decision-makers, then the guideline will be binding. Evidence trumps drafting and design. In some instances, there will be evidence about whether a particular set of guidelines is binding or not. For example, it might be clear from the record in a particular case that a decision-maker applies guidelines “as if they were law” (Toussaint v. Canada (Attorney General), 2010 FC 810 (FC): 56). A detailed empirical study may lead to similar conclusions (Sossin and Pottie 2005). The goal here is to determine whether it can be said, looking at the text of a guideline, that the guideline is binding, essentially creating a bindingness handbook which drafters and designers of guidelines, and judges who review guidelines, can refer to.

In an important recent contribution to the literature, Justice Sossin and Chantelle van Wiltenburg have suggested that courts should look to the “impact” and “influence” of soft law in determining whether judicial intervention is appropriate and on what terms (Sossin & van Wiltenburg, 2021: 650). I take a different tack for three reasons. First, as explained in the Introduction, questions about bindingness often arise in the shadow of the law and affect negotiations within government about the drafting and design of guidelines: a focus on “impact” and “influence” directs attention away from how guidelines are developed internally; whilst useful for courts, who assess matters ex post, the impact/influence focus is less useful for administrative decision-makers who draft and design guidelines without certainty about how they will apply in the real world. Second, in some scenarios a court will be concerned with a purely textual question: whether an individual has a legitimate expectation that a particular procedure will be followed based on a guideline turns on whether the guideline is “clear, unambiguous and unqualified” ((Canada (Attorney General) v. Mavi, [2011] 2 SCR 504 (SCC): 68). Third, there are “forensic limitations” on the ability of the courts to conduct factual inquiries (Moncrief-Spittle v. Regional Facilities Auckland Limited, [2022] NZSC 138: 86). The information a court will have at its disposal is limited: judicial review, whether of a guideline or an individual decision taken by reference to a guideline (or which should have been taken by reference to a guideline) is not a general empirical inquiry into how the guideline actually operated in practice. Where evidence is available and admissible, a court can certainly look to it to inform a decision about whether a guideline is binding. But often evidence will be neither available nor admissible. My analytical framework will be useful in such circumstances and a helpful complement to the impact/influence focus in all other circumstances.

This content has been updated on March 13, 2023 at 22:23.