EXAMINATION candidates sitting mathematics papers may earn credit for their working, as well as for arriving at the correct answer. If the same were true of judges deciding cases, then the Supreme Court would be awarded less than full marks for its decision in R. (Gallaher Group Ltd.) v Competition and Markets Authority [2018] UKSC 25, [2018] 2 W.L.R. 1583. While it arrived at a defensible conclusion, its reasoning leaves something to be desired.

Following an investigation by the defendant's predecessor, the Office of Fair Trading (OFT), into alleged price-fixing in the tobacco market, several parties entered into “early resolution agreements” (ERAs) acknowledging (without prejudice to the possibility of subsequent appeal) infractions of competition law in return for lighter penalties. One such party, TM Retail Group Ltd. (TMR), had been assured that, even if it did not itself appeal, it would receive the benefit of any other party's successful appeal that undermined the OFT's decision in its case. Its financial penalty was duly repaid following a relevant appellate decision. However, the OFT decided not to extend similar lenience to the claimant: like TMR, it had entered into an ERA and had declined to launch its own appeal; but, in contrast to TMR, the claimant had received no assurance from the OFT about benefitting from others’ successful appeals.

Challenging the OFT's decision by way of judicial review, the claimant failed at first instance but succeeded in the Court of Appeal ([2016] EWCA Civ 719, [2016] Bus. L.R. 1200), where it was held that the OFT's decision engaged (inter alia) the principle of equal treatment at common law and lacked objective justification. The Supreme Court disagreed. Rationality, it said, was the acid test of legality in these circumstances – and the defendant's decision was not irrational. That conclusion is perfectly tenable, there being an intelligible basis for treating the claimant differently from a party who was in receipt of an assurance. However, it is incumbent upon apex courts not only to arrive at sound conclusions but to do so in a way that safeguards the coherence of the law. Judged thus, the Supreme Court fell short in Gallaher as it sought to grapple with important questions about the configuration of administrative law doctrine – and, in particular, the level of abstraction at which it can most usefully be conceived.

In falling back on the Wednesbury reasonableness, or rationality, test, Lord Carnwath (with whom the other four Justices agreed) and Lord Sumption manifested a strong preference for utilising a limited range of well-established doctrinal tools. For instance, Lord Sumption (by way of agreement with Lord Carnwath) said that “it is important not unnecessarily to multiply categories” because doing so would tend to “undermine the coherence of the law by generating a mass of disparate special rules” (at [50]). That the unnecessary multiplication of categories is to be avoided is self-evident. But the hard questions, as always, lie at the margin. When is the creation or acknowledgement of a new category helpful – because, for instance, it bolsters legal transparency or predictability? And when does it render the law undesirably Byzantine? In seeking to tread that difficult line, the Court enjoyed some, but only limited, success.

The success lies in the bucket of cold water that Lord Carnwath poured over notions such as “abuse of power” and “conspicuous unfairness” (in the substantive, as distinct from the procedural, sense of unfairness), the sloppy deployment of which has muddied the doctrinal waters of administrative law. A prime example is R. (Rashid) v Secretary of State for the Home Department [2005] EWCA Civ 744, [2005] Imm. A.R. 608, in which the legitimate expectation doctrine was contorted so as to apply even though the claimant was unaware of the administrative policy upon which the expectation was purportedly founded. It was plain that no legitimate expectation could really have arisen in such circumstances – it being impossible to expect that of which one is ignorant – yet such problems were papered over by the court, on the ground that an “abuse of power” had occurred.

A similar approach might have been adopted in Gallaher. The case did not fit the legitimate expectation mould, given that the claimant had received no assurance and knew nothing of that made to another party. Yet a court minded to assist might have brushed such doctrinal inconveniences away on the ground that the differential treatment represented an “abuse of power” or was “conspicuously unfair”, so malleable are those notions. The Supreme Court, however, wisely chose not to go down that path. “Abuse of power” and “conspicuous unfairness”, said Lord Carnwath, cannot properly be regarded as free-standing principles of administrative law – and such language “adds nothing to the ordinary principles of judicial review” (at [42]). He is surely correct. The concepts are simply too diffuse to serve as heads of review, and are so manipulable as to render the true reasons for judicial intervention (or forbearance) unacceptably opaque. Meanwhile, the use of such notions to stretch – and, as in Rashid, distort – established heads of review is to be deprecated.

The defensibility of Lord Carnwath's rejection of “abuse of power” and “conspicuous unfairness” as administrative law principles might appear to demonstrate the wisdom of Lord Sumption's less-is-more philosophy, according to which the multiplication of categories of unlawfulness is to be avoided, and well-established heads of review stuck to. There is doubtless something to be said for that view. But it can be taken too far, as it was in Gallaher. The central issue at stake in such cases turns upon the values of consistency and equal treatment, whereby like cases ought to be treated alike, absent a good reason for differential treatment. Those values have recently secured increasing judicial recognition, including at Supreme Court level. For instance, in Mandalia v Secretary of State for the Home Department [2015] UKSC 59, [2015] 1 W.L.R. 4546, at [29], Lord Wilson referred to “a principle, no doubt related to the doctrine of legitimate expectation but free-standing”, according to which (and here Lord Wilson quoted Laws L.J. in R. (Nadarajah) v Secretary of State for the Home Department [2005] EWCA Civ 1363, at [68]) “public bodies ought to deal straightforwardly and consistently with the public”.

Against this background, it is perhaps surprising that in Gallaher Lord Carnwath concluded – without referring to Mandalia – that “equal treatment” is not “a distinct principle of administrative law” (at [24]) and that, when issues of consistency (which, he allowed, is a generally desirable objective) arise, they are best thought of only “as aspects of rationality” (at [26]). It was thus by reference to such “ordinary principles of judicial review” as rationality that the case fell to be decided (at [41]). Unlike Lord Carnwath, Lord Sumption was prepared to acknowledge that there is “a common law principle of equality”, but went on to qualify this, saying – in line with Lord Carnwath – that the principle is “usually no more than a particular application of the ordinary requirement of rationality imposed on public authorities” (at [50]).

Such dicta lead us to some fundamental questions about administrative law's doctrinal configuration. In particular, what, if anything, turns on whether there is an administrative law “principle” of consistent, or equal, treatment? That Lords Carnwath and Sumption were able to disagree about whether such a principle exists but were unanimous that consistent/equal treatment is relevant to questions of rationality suggests that there are conceptual difficulties in play here. So do the internal inconsistencies within the administrative law worldview found in the leading judgment of Lord Carnwath, for whom legitimate expectation is an “ordinary principle of judicial review”, but for whom consistent/equal treatment occupies some lesser status as a factor that is merely pertinent to the application of a true “principle” of review (namely, rationality). The implication is that courts must confront issues of equality/consistency through the prism of rationality, whereas legitimate expectation amounts to a doctrinal lens in its own right.

This view does not, however, withstand scrutiny. The existence of an assurance that gives rise to a legitimate expectation does not in and of itself establish the unlawfulness of administrative conduct that cuts across the assurance. Rather, it triggers judicial evaluation – by reference to the Wednesbury rationality or proportionality doctrine, as appropriate – of the quality of any justification offered for the expectation's frustration. No one is suggesting that the value of consistency ought to receive more judicial protection than that afforded to a legitimate expectation, it being a given that inconsistent treatment should be regarded as lawful provided that it is adequately justified. And, to be fair, Lord Carnwath does not appear to think that consistency should receive any less protection, given his acknowledgement that it can feed into rationality review. It might therefore be retorted that any criticism of Gallaher amounts to little more than taxonomical fetishism: does it really matter whether, like legitimate expectation, consistent/equal treatment is regarded as an administrative law “principle” in its own right (in the sense of being a separate ground of review)?

For three reasons, it does matter. First, recognising consistent/equal treatment as a free-standing principle would be fitting recognition of its normative importance. Secondly, it would facilitate the development of a suitable doctrinal superstructure, as occurred with legitimate expectation. This would enable courts to approach in a more systematic and predictable manner such issues as the circumstances in which a policy or practice should be regarded as sufficiently clear and settled as to render deviations from it prima facie “inconsistent”. Thirdly, carving out a principle of consistent/equal treatment from the generality of rationality review would place less weight on the latter and, as a result, would aid legal transparency. It is now more than 30 years since Jowell and Lester [1987] P.L. 368, 372, highlighted the dangers of allowing the normative reasons for judicial intervention to remain obscured within the folds of the Wednesbury doctrine's “ample cloak”. It is high time that the principle of consistency stopped lurking therein.