This post is a working paper. It contains some ideas I am developing for a longer piece which will appear in an edited collection on substantive judicial review. This working paper, which was first published on the UK Constitutional Law Blog, can also be downloaded from SSRN.

To observe that substantive judicial review, and the notions of proportionality and deference in particular, constitute well-trodden ground would be to engage in reckless understatement. And that, in turn, might suggest that there is nothing more that can usefully be said about these matters. Yet the debate in this area of public law remains vibrant—and for good reason. Like the controversy about the foundations of judicial review in which many public lawyers engaged energetically over a decade ago, the controversy about substantive review is ultimately a manifestation of underlying disagreements concerning the nature, status and interaction of fundamental constitutional principles, including the rule of law, the separation of powers and the sovereignty of Parliament. It is hardly surprising, then, that questions about the intensity of review and (what amounts to the reverse side of the same coin) deference remain under active discussion long after the debate was ignited by the entry into force of the Human Rights Act 1998.


Although the debate about substantive review is in fact a reflection of deeper theoretical and normative controversies, one of its hallmarks is actually an emphasis on doctrine. Much has been said, for instance, about what form the doctrinal contours of substantive review should take now that proportionality is relatively well-established in domestic law. Should the reasonableness and proportionality doctrines co-exist, or can—and should—we make do only with the latter? As is well known, the late Michael Taggart (to the surprise of some) argued strongly in favour of the retention of Wednesbury alongside proportionality, advocating a “bifurcated” public law in order to accommodate such doctrinal diversity. Considerable attention has been paid too to the question whether deference—to the extent that it is appropriate at all—should be characterized as a doctrine in its own right, or whether it (or the thinking that animates it) is better conceived of as something that internally shapes the application of the proportionality doctrine. And proportionality itself is the site of doctrinal debate, in relation both to its structure and the nature and relationship between the various sub-tests—rational connection, necessity, fair balance, and so on—of which it consists.

There is nothing inherently wrong with the fact that such matters have been (and continue to be) debated. Indeed, I have contributed to some of these debates myself, and I have certainly not undergone a Damascene conversion that makes me doubt the value of doctrine in administrative law (or generally). Doctrinal clarity is highly desirable, and debate which serves to clarify doctrine is therefore to be welcomed. There is, however, a risk that an undue focus upon doctrine may ultimately prove to be a distraction, with the doctrinal trees serving to obscure the normative wood. A related risk is that over-emphasizing doctrine may yield a legal architecture whose rigidity is unwarranted. The bifurcation thesis is arguably a case in point. It envisages a bright-line distinction between two realms of public law which are respectively the doctrinal preserves of proportionality and reasonableness. The former (on this scheme) is concerned with the protection of individual rights, the latter with public wrongs. The former is a structured form of review, whilst the latter is unstructured. The former constitutes an intensive form of review, the latter a relaxed one. And the former safeguards norms that are important in a way that those in play in the latter context are not. The difficulty is that these doctrinal distinctions map onto the normative terrain of public law only in a highly approximate way. They capture something of what is going on, but the bluntness of such tools diminishes their usefulness.


A better starting point, I think, is another insight offered by Taggart, according to which public law is increasingly about the enforcement of a “culture of justification”. In fact, public law has always, at some level, been about justification. Whatever one’s stance (if any) in relation to the debate about the foundations of judicial review, it is hard to deny that the ultra vires doctrine institutionalizes the justificatory demands of the rule-of-law principle of legality. On that view, the legitimacy of executive action is determined by reference to the empowering statute, action undertaken beyond the bounds set by the legislation being legally unjustified and so unlawful. It follows that the ultra vires doctrine has always called for exercises of administrative authority to be justified by reference to positive law.

What has changed, however, is that the set of criteria by reference to which governmental action must be justified is generally taken to be increasingly rich. No longer is it sufficient that executive action respects only the boundaries that can be traced in any direct way to the explicit or implicit terms of the relevant legislation: to withstand judicial scrutiny, it must also conform to the demands of the legitimate expectation principle, fundamental constitutional rights, and so on. These changes, in turn, track the transition from thinner to thicker prevailing understandings of the rule of law (and help to explain why the ultra vires doctrine was increasingly perceived as an inadequate foundation for judicial review). The constant in all of this is the requirement that exercises of administrative power be justifiable by reference to law: but what the law requires is taken to have become increasingly demanding.

Against this background, several questions need to be posed about the notion of justification. Most obviously, it is meaningless to ask whether a given decision or policy is justified unless we articulate the standards against which the measure is to be evaluated. At the very least (leaving third-source considerations to one side for present purposes) there will be a need to demonstrate positive legal authority. But if particularly valuable norms, such as the rights or legitimate expectations of the individual, are impinged upon then the sufficiency of any justification may fall to be assessed against additional, more demanding criteria. It is also important to be clear about two further matters. First, when we say that a decision must be justified by reference to a given benchmark, what does that really mean? It presumably means something more exact (and, normally, something less) than that the court has to be satisfied that it too would have proceeded in the way that the administrator did. The nature of justification is, in this sense, tied up with considerations about the standard of review—which, in turn, relates back to an assessment of the normative significance of the value impugned by the decision and hence the appropriate scale of the decision-maker’s justificatory burden. Second, even once the issue of the standard of justification, or review, has been settled, questions will arise about whether that standard has been met—which, in turn, triggers questions about the court’s role in evaluating the quality of any justifications offered by the decision-maker.

It is in the interaction of these sets of considerations—encompassing the normative values that give life to substantive review as well as the nature of justification and allied questions concerning the proper judicial role—that we find the seeds of a mature approach to substantive judicial review. If—in contrast to arguments that place doctrine centre-stage—this sounds less than neat, then that is because a degree of messiness is unavoidable in this sphere. What substantive review amounts to in any given case is a question that cannot satisfactorily be answered by resort to any neat scheme of categorization.

My argument is that it is necessary to move beyond a doctrinal focus which results (depending upon one’s preferences) in either a bifurcated approach or one wedded to a specific doctrine (e.g. proportionality), and to concentrate instead on calibrating substantive review by reference to the normative and institutional considerations which ought properly to shape it. I readily acknowledge that this approach may sometimes—perhaps often—produce outcomes that do not, at least superficially, differ radically from the position that would obtain according to the conventional wisdom. I am certainly not suggesting that the courts invariably, or even often, get it wrong. The difference, however, is one of emphasis. The aim is to secure a framework that is better equipped to enable the courts to get it right, and one that is shaped from the bottom up by the relevant normative and institutional factors. Doctrine should be the servant of such considerations, not a procrustean bed into which they have to be shoehorned.

This approach calls for a distinction to be drawn between two types of “deference”. The meaning of that term, and the proper place (if any) of the concept, are of course contested. One reason for this is that deference, as it has come to be deployed within the substantive-review debate, has assumed a somewhat chameleonic character. This has become an obstacle to understanding.  I have argued elsewhere that an adequately structured approach to proportionality review demands proper disaggregation not only of that doctrine’s various limbs but also of the distinct bases—most notably institutional competence and democratic legitimacy—on which deference may be appropriate. It is, however, necessary to go further, by recognizing not merely the different grounds upon which deference might be warranted, but also the different senses in which deference may interact with the idea of justification. To put the matter more concretely, in any substantive-review case, two quite different things necessarily have to happen (albeit that the court may not always confront this distinction in a methodical and explicit way). And those two things raise two distinct sets of questions concerning deference.

Starting-point deference and the burden of justification

First, the court will have to determine what should constitute the operative standard of justification in the particular circumstances of the case. What, in other words, should be the justificatory burden under which the decision-maker is placed, and which will have to be discharged if the decision is to be found by the reviewing court to be lawful? The proportionality versus rationality debate captures something of this matter—but only in a rudimentary fashion. On the face of it, asking whether a decision is proportionate is different from—and subjects the decision-maker to a more demanding justificatory burden than—asking whether it is merely rational. However, just as proportionality and rationality are distinguishable, so are distinctions—justificatory gradations—concealed within those concepts. The “sub-Wednesbury” and “super-Wednesbury” notions have long been evidence of this in relation to the rationality doctrine, as also is the more recently-developed “cogent reasons” concept. The point is obvious too in relation to proportionality, which can be (and is) deployed in more and less demanding ways. For instance, the requirement that a given measure be a “necessary” means of advancing a legitimate aim does not always mean what it says, in that the necessity criterion does not invariably rule out every option save that which is the least restrictive of the compromised right or other norm.

The upshot is that courts have at their disposal a broad range of justificatory standards. And the fact that particular labels—strict proportionality, cogent reasons, super-Wednesbury, and so on—can be attached to those standards does not alter the fact that they exist as multiple points on a spectrum that necessarily shade into one another, as opposed to cleanly-demarcated silos. In terms of how this point maps on to doctrinal considerations, there is no particular objection to labels such as “reasonableness” and “proportionality” provided that they are acknowledged merely to be broad indicators of the particular justificatory standard being applied in the given case. They can only, however, reflect a conclusion that has been drawn about the appropriate standard of review based upon an assessment of the specific features of the case itself: and they can only, for the reasons given above, describe in very approximate terms what is really going on. While, therefore, I am not arguing for (or against) a wholesale rejection of the conventional nomenclature of substantive review, it is necessary to bear in mind that its casual use may serve to obscure more than it illuminates.

An obvious difficulty with the approach sketches above is that it risks collapsing into multiplicity of single instances, thereby depriving the law of any tangible structure or predictability. The solution, however, lies not in a retreat into rigid categorization—far less bifurcation—but in an attempt at calibration which exposes and harnesses the relationship between underlying normative considerations and administrative law’s doctrinal superstructure. The key factor which drives, and which ought to drive, considerations about the operative standard of justification is the normative significance of the value impacted by the impugned decision: the more significant the norm at stake, the more closely it warrants judicial examination. However, once we recognize the insufficiency of bald distinctions between (for example) rights and non-rights cases, it becomes necessary to engage with the underlying questions concerning the significance of different norms in a more thoroughgoing manner.

This, in turn, calls for an explicit ordering of norms by reference to which the operative standard of justification falls to be set. Such an ordering is already implicit—albeit in a sometimes-crude form—in the courts’ jurisprudence, “rights” cases, for instance, generally resulting in the imposition of a heavier burden of justification than (to adopt Taggart’s terminology) “wrongs” cases. For the reasons sketched above, however, this kind of ordering is insufficiently granular. Not all rights cases exert the same degree of normative pull, some rights—as well as particular manifestations of a single right—being more normatively compelling than others. Equally, it does not follow that cases that are not about rights are all as unimportant as one another (thus calling for a uniformly low standard of justification) or, for that matter, necessarily less important (as a category) than rights cases (as a category).

Another way of looking at this is by reference to the other—that is, the deference—side of the substantive-review coin. Subject to the by-now inevitable warnings about the difficulties associated with the notion (or at least the language) or deference, it is clear that some approaches to review—some standards of justification—are necessarily more deferential than others. For instance, asking whether a decision is barely rational is inherently more deferential—in the sense of being less demanding, and therefore less likely to result in the decision-maker’s view being overturned—than asking whether it is strictly necessary and proportionate. In this way, some approaches to review have a form of deference hard wired into them in a way that others do not.

Any substantive-review case must therefore begin with the court determining what, if any, level of starting-point deference should be exhibited. Thinking about the matter in this way may be helpful for three reasons. First, making a decision about (this form of) deference—or, the flip-side of the same coin, setting the operative standard of justification—must, of practical necessity, be confronted by the reviewing court at the outset of its inquiry. Whether the impugned measure is justified cannot sensibly be decided unless, in the first place, the operative standard of justification has been determined. Second, questions about the standard of justification, or starting-point deference, fall to be examined by reference to a relatively abstract conception of the nature and importance of the norm that is compromised by the impugned decision, rather than by reference to case-specific considerations pertaining to the court-administrator relationship or the interaction of the impugned norm and the specific measure that conflicts with it. Third, characterizing this matter as starting-point deference serves to distinguish it from a second form of deference that is considered below.

Deference in adjudication

Earlier in this post, I said that in any substantive-review case, the court must undertake two distinct tasks, the first of which—determining the operative standard of justification—engages what I have called starting-point deference. The court’s second—and logically subsequent—task is to decide whether the burden of justification imposed upon the decision-maker at the first stage has been discharged. Has the measure being challenged by the claimant been shown to be rational, or supported by cogent reasons, or strictly necessary, or proportionate in the sense of striking a fair balance between the interests of the right-holder and society, or justified by reference to whatever other standard the court deems appropriate in the circumstances of the particular case?

When the court rolls up its sleeves and begins to confront questions of this nature, considerations about the intensity of review—and deference—necessarily shift from the abstract to the particular. The setting of the burden of justification (or of the starting-point level of deference) is undertaken on the basis of an all-other-things-being-equal assessment of the nature and importance of the compromised norm. The reality, however, is that all other things are often not equal, such that the court’s analysis of whether the operative standard of justification has been met may need to be moderated by reference to considerations of adjudicative deference. Starting-point deference is then about determining the onerousness of the decision-maker’s justificatory burden; adjudicative deference, in contrast, is relevant when the court is determining whether that burden has actually been discharged.

For instance, courts are often called upon to determine whether it is necessary to restrict a right to a given extent in order to realize a competing policy aim—an assessment that may involve the making of challenging predictions about the relative expediency of other, less-restrictive potential measures. The court may exhibit adjudicative deference—in recognition of the fact that it may be less well placed than the decision-maker to perform this sort of assessment—by ascribing a degree of respect the decision-maker’s view as to the relative expediency of the various possible measures that were open to it. (By extension, the court should also be prepared to ascribe appropriate weight to the view of other expert parties or interveners, but should only ascribe weight to anyone’s expert view in circumstances where such expertise has actually been brought to bear on the problem at hand.)

The court may also have to decide whether a measure, even if it is necessary, strikes a fair balance between the interests of the right-holder and those of society more generally. This question reduces, at least to some extent, to a value judgment, the acceptability of the balance struck between two incommensurable variables being impossible to determine unless those variables are first invested with values that are inherently contestable. In the light of this, adjudicative deference that ascribes weight to the decision-maker’s view—in recognition, where relevant, of its democratic credentials—may well be warranted.

I recognize that these claims about the appropriateness of adjudicative deference (particularly on democratic grounds) are themselves controversial, and—having examined these matters at length in another piece—I need not address them in depth here. In any event, the thrust of my present argument is principally concerned not with the interstices of adjudicative deference, but with the point that it is distinct—and should be distinguished—from starting-point deference. The former can be thought of as the pragmatic counterpart to the abstraction of the latter: as something which shapes the on-the-ground assessment of whether the normatively-warranted burden of justification has been discharged. Viewed thus, the nature and intensity of substantive review in any given case is a function of the interoperation of the notions of starting-point and adjudicative deference, rather than something that can be described in the preconceived doctrinal terms of Wednesbury or proportionality.

An obvious objection to this analysis is that it openly acknowledges that the practical reality of substantive judicial review may fall short of that which is called for by the normative pull of the right or value lying at the centre of the dispute, as adjudicative deference blunts the scrutiny which the court, at the outset, has found to be appropriate. From another perspective, however, this approach helps to expose—and demands that we confront—the interaction of the array of factors that have the potential to influence substantive review. It does not, in any event, follow that the relationship between starting-point and adjudicative deference is necessarily a one-way street, according to which the latter can serve only to blunt the scrutiny that is called for by the former. For instance, circumstances might arise in which the normative force of the impugned value is regarded as sufficiently compelling to reduce the purchase of adjudicative-deference factors such as institutional competence and democratic legitimacy. Indeed, precisely this notion is evident in the distinction between absolute and qualified rights, the normative pull of the former being so great as to undercut the scope for adjudicative deference. But, as with other bright-line distinctions, the qualified/absolute rights division is merely a manifestation of underlying considerations that are capable of exerting subtler effects. The purpose of distinguishing between starting-point and adjudicative deference is not, then, to erect a new bright-line doctrinal distinction in the place of those that I have already criticized. Rather, the purpose of the distinction (such as it is) is to facilitate more sophisticated—and transparent—engagement with the interlocking considerations that, whether acknowledged or not, are inevitably in play when questions of substantive review are examined.

What this amounts to, therefore, is an attempt to avoid the difficulties (as I see them) that inhere in both the bifurcation thesis and the view which holds that Wednesbury should be jettisoned now that proportionality is firmly a part of the public-law landscape. Neither approach can adequately capture the complexity of what is—or ought to be—going on in cases of this sort. (It has, of course, been pointed out that proportionality’s advance could be accommodated by conceiving of it in almost-infinitely flexible terms—which, to an extent, is correct. But to construct proportionality in such a way would render it so open-textured as to be largely meaningless. In reality, this thesis clings formally to the doctrine-led model whilst largely emptying the doctrine of content; as a result, it begins to collapse into precisely the kind of approach I am advocating.)

I am not, however, arguing that there is no place for doctrine in this area: I am, after all, proposing not one but two doctrines (as they might fairly be described) of deference. The tools of starting-point and adjudicative deference, however, are intended to enable the calibration of substantive review in a manner that is properly sensitive to the constitutional, institutional and normative considerations that underlie the debate in this area. Such an approach is preferable to one which seeks to effect either the doctrine-led bifurcation of public law or its colonization by a single (but potentially empty) doctrine of substantive review such as proportionality. This is not, then, about ridding this branch of administrative law of doctrine—it is about putting doctrine in its proper place.

Posted by Mark Elliott

Mark Elliott is Professor of Public Law at the University of Cambridge, a Fellow of St Catharine's College, Cambridge, and Legal Adviser to the House of Lords Constitution Committee. All views on this blog are expressed in a purely personal capacity.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s