Proportionality and intensity of judicial review under EU law: clear as mud?

Christopher Brown

On 17 June 2011 the Court of Appeal gave judgment in an interesting case concerning the legality of a measure prohibiting the use of cigarette vending machines, commonly seen in pubs and bars: R (Sinclair Collis and Another) v Secretary of State for Health [2011] EWCA Civ 437.  It raised the now frequently encountered doctrine of proportionality under EU law. It was common ground that the measure, contained in secondary legislation, was caught by Article 34 TFEU (ex Article 28 EC) on the free movement of goods – most if not all machines are imported from continental Europe – and that the Government’s pleaded justification, namely that the measure pursued the objective of protecting public health, was prima facie a good one (see Article 36 TFEU).  The debate concerned the issue of whether the ban was proportionate.

For those who have not read the judgment, it may come as a surprise to hear that it is some 76 pages long and contains long reasoned judgments from all 3 members of the Court. Why so? The simple answer is that they all approached the matter differently. The majority – the Master of the Rolls (MR) and Arden LJ – came to the same conclusion (just about) but by different means, whilst Laws LJ issued a long dissenting judgment.

The facts are not easy to summarise shortly. The Government started to look at what if anything to do about vending machines back in 2008. Its primary (arguably sole) concern was that they were being used by the under 18s, which was something which the Government naturally wished to prevent as part of its attempt to combat underage smoking. Three options were considered: maintain the status quo; regulate machines by mandating the use of “age restriction mechanisms” (ARMs) such as electronic ID card age verification; or prohibit such machines outright. Initially, and indeed until a late stage in the legislative process, the Government was in favour of regulation which, if unsuccessful, might lead to a ban. The Bill introduced by the Government would have given the Secretary of State the power either to prohibit or to regulate. However, an amendment to the Bill was moved by a backbench MP such that the SoS would only have the power to prohibit cigarette vending machines.  That amendment was adopted and made its way into the Health Act 2009Secondary legislation was adopted soon after the Act entered into force providing for the prohibition of such machines.  The focus was on the under 18s, although the Government also pointed to potential benefits to adults, on the basis that cigarettes would be slightly more difficult to acquire.

The essence of the appellants’ challenge was that the prohibition fell foul of the proportionality principle. It was submitted that there were less restrictive ways of achieving the policy goal of the protection of human health, namely by statutory regulation or even voluntary action on the part of the industry with a threat of prohibition if results were not forthcoming. The appellants pointed to the fact that the SoS had initially supported the idea of statutory regulation as an appropriate means of combating under-age smoking.

All three judges relied on two particular cases: the well-known judgment of the Court of Justice in FEDESA [1990] ECR I-4023 and the Court of Appeal’s judgment in Eastside Cheese [1999] 3 CLMR 123, but they all drew different conclusions from them.  Given their importance, it is worth quoting the salient passages from each:

In FEDESA, the Court of Justice held as follows at para 13:

“The Court has consistently held that the principle of proportionality is one of the general principles of Community law. By virtue of that principle, the lawfulness of the prohibition of an economic activity is subject to the condition that the prohibitory measures are appropriate and necessary in order to achieve the objectives legitimately pursued by the legislation in question; when there is a choice between several appropriate measures recourse must be had to the least onerous, and the disadvantages caused must not be disproportionate to the aims pursued.”

Eastside Cheese concerned an emergency control order, made following a case of E. coli infection, which prohibited the carrying on of any commercial operation in relation to cheese originating from a particular dairy.  The issue was the order’s justification under Article 36 TFEU.  Lord Bingham CJ said this at paras 41-48:

“41 The principle of proportionality is one of the basic principles of Community law. It has been expressed by the European Court of Justice in [FEDESA]… Because the principle is so general (and may affect a range of issues from the validity of primary legislation such as the Shops Act 1950 to much narrower points such as the quantum of penalties for customs infringements) it must be related to the particular situation in which it is invoked…

43. However the test is formulated, it is clear that in the application of Article 36 the maintenance of public health must be regarded as a very important objective and must carry great weight in the balancing exercise. In De Peijper [1976] ECR 613, 635 (paragraph 15) the Court of Justice said that health and the life of humans rank first among the interests protected by Article 36, and it is for member states to decide (within the limits imposed by the Treaty) what degree of protection to provide…

45. In principle the decision on proportionality has to be taken by the national court which is seised of an issue on Article 36, subject of course to any possible reference to the Court of Justice … But in the case of a legislative measure the national court must not simply accept the view of the national legislature or confine itself to deciding whether what the legislature has enacted is reasonable.

46. Nevertheless it is clear that the national legislature has a considerable margin of appreciation, especially in legislating on matters which raise complex economic issues connected with the Community’s fundamental policies.

48. …The margin of appreciation for a decision-maker (which includes, in this context, a national legislature) may be broad or narrow. The margin is broadest when the national court is concerned with primary legislation enacted by its own legislature in an area where a general policy of the Community must be given effect in the particular economic and social circumstances of the member state in question. The margin narrows gradually rather than abruptly with changes in the character of the decision-maker and the scope of what has to be decided (not, as the secretary of state submits, only with the latter)…”

Laws LJ gave the first judgment. He noted that the doctrine of proportionality applies differentially depending on the case, the engine of the differences being the scope of the margin of discretion or appreciation accorded to the decision-maker (in Arden LJ’s words, the intensity of review). He said that two factors in particular affected the scope of such margin – the identity of the decision-maker and the subject-matter of the decision. As for the former, his view was that acts of the primary legislator attracted a broader margin and acts of the secondary legislator a narrower one; as for the latter factor, general policies of the EU and the promotion of a benefit of great general importance, such as the protection of health, attracted a broad margin of appreciation (§23).

Importantly for his judgment, Laws LJ considered that the real decision-maker here was the Secretary of State. This augured in favour of a stricter approach to proportionality. He took issue with the High Court’s conclusion that the ban was to be treated as if contained in primary legislation, saying that that had “constitutional implications” (§28). The fact that Parliament had given the SoS the power to prohibit vending machines could not be construed as meaning that a ban should be put into effect.  The relevant provision of the Act did nothing more than confer a discretion (§30). The legislators’ subjective intention could not condition the rule-maker’s discretion to make regulations pursuant to the Act or broaden the margin of appreciation to be accorded to the rule-maker.

Laws LJ came to the conclusion that the ban was disproportionate, although he rejected a submission by the appellants that the Act itself, properly construed, permitted the SoS to ban vending machines which were not fitted with ARM.  He held (at §76) that:

“[The Act] on what I would hold is its true construction does no more than empower the Secretary of State, at his discretion, to impose the ban. The bare conferment of the power cannot offend TFEU Article 36 or ECHR A1P1. Accordingly the burden of the judicial review must fall on the Regulations, under which the Secretary of State had to decide whether to use the power conferred. Faced with s.3A as enacted he might have decided to do so albeit after considering ARM, if he concluded for sound reasons that ARM would be insufficiently effective to achieve the legitimate aim in hand and only the ban would do. If on the other hand his conclusion was that ARM would suffice and the ban would be disproportionate, no doubt the ban would not be imposed. However he could not impose ARM; and in those circumstances the government might ask Parliament to reconsider the legislation. But the effective decision for the purpose of these proceedings would be that of the Secretary of State, whatever the realpolitik. He would enjoy the margin of appreciation appropriate to him. And he cannot, as I have said, find solace in the fact that the enabling statute on its proper construction did not allow him (or the Minister) to impose ARM as a policy choice.”

And at §79 Laws LJ said:

“What has, I think, made this case elusive is that while s.3A as enacted did not empower the Secretary of State to impose ARM, on the view I take he was nevertheless obliged to consider ARM’s merits before exercising the power to impose a ban. If he did not, the imposition of the ban would be (and in my judgment is) disproportionate. I reiterate: the Minister cannot find solace in the fact that the enabling statute on its proper construction did not allow her to impose ARM as a policy choice. The point is important because, as it seems to me, any other view involves treating the 2009 Act as conferring on the Secretary of State a discretionary power which he might exercise free of the discipline and constraint of proportionality. But there is nothing in the statute, and certainly not in the general law, to justify such a position. This approach moreover preserves the constitutional divide between the power of Parliament and the power of the Secretary of State.”

Laws LJ’s approach may be constitutionally pure, but it does have a whiff of the unreal about it: Parliament had spoken in clear terms, notwithstanding the SoS’s express preference for regulation in the first instance. The judge’s approach also begs the question as to what the SoS would be permitted to do if Parliament said – as might readily be expected – that it would not reconsider the Act.  Would that lead to EU law stalemate? Or would a future decision of the SoS to ban machines then be more immune to challenge?

Arden LJ took a rather different view: whilst ordinarily the secondary legislator attracts a higher level of scrutiny, that was not inevitably the case once other factors were taken into account; there was no “bright line rule” (§137).  She considered that Parliament had made positive decisions both to give the SoS the power to ban vending machines and to rule out the possibility of statutory regulation by the SoS. Parliament had also acted “in the confident expectation” that the SoS would not leave the matter to be regulated by a voluntary code (§168). The intensity of review of decisions of Parliament was that of “manifest error”; the concept of “least intrusive means” was not applicable or applied with a much lower level of intensity.  The same intensity of review applied to decisions of the SoS authorised by Parliament.  For Arden LJ, Parliament and the SoS were “engaged on a common enterprise” here (§151), and so this case was rather different from other cases, such as Eastman Cheese, where the SoS’s role was independent of Parliament: in such cases, the margin of discretion was narrower.  In her view (§174):

“If the powers of the Secretary of State are to be adjudged by some different standard from that applying to Parliament in those areas where their powers overlap, the measure of discretion given by European Union law to the national legislature is undermined and made useless, and the decision of Parliament is also undermined.”

She concluded that no manifest error had been made by either Parliament in conferring the power to ban vending machines or by the SoS in exercising it. In particular, the SoS’s decision to impose a ban was not on its face manifestly inappropriate; the SoS did not have to go further and show that (in this case) a voluntary code would not be a better means of achieving a reduction in underage smoking.  Instead, the burden of adducing evidence that a less intrusive means of achieving the same goal passed to the appellants; and the appellants had not discharged the burden of proving that there were other equally suitable ways of achieving the same legitimate aim pursued by the SoS. Indeed, the evidence on the effectiveness of the installation of ARM in combating underage purchases was weak (§165).

The MR, giving the final judgment, confessed to having changed his mind more than once before settling in favour of the Government. He did not, however, agree with Arden LJ’s views that Parliament and the SoS were acting in “partnership” (§212) and that the margin of appreciation was the same for acts of both of them in this case.  For him, the breadth of the margin in relation to any decision depended on the circumstances of the case and in particular on the identity of the decision-maker, the nature of the decision, the reasons for it and its effect (§200).  He said (§213):

 “While the decision to impose the ban must, therefore, be assessed on the basis that it was made by the executive arm of the Government of a single member state, when deciding on the margin of appreciation to be afforded to the Government in relation to the Regulations, it seems to me that (i) the involvement of the legislature is a factor favouring a broader margin than would be appropriate to a purely executive decision, and (ii) the fact that the Regulations are concerned with health and the prevention of premature death supports a relatively broad margin, as does (iii) the fact that the Regulations are based, in part, on complex assessments of the public interest.

 However, I do not consider that the margin is as broad as it would be if the decision had been that of the democratically elected national legislature. Even less do I consider that the margin is as broad as it would be if the decision had applied across all EU member states.”

In other words, the MR’s approach falls somewhere between those of Laws LJ and Arden LJ: the margin of appreciation in this case is broader than it would be if the decision were ‘purely executive’, but not as broad as that accorded to decisions of the democratically elected legislature.

Contrary to Arden LJ, he acknowledged that there was an obligation to opt for the least restrictive way of achieving a particular objective, but (§203):

“that factor should not be applied by a court in such a way as to usurp the role of the primary decision-maker. So, where there is an alternative possible measure, there may be a difference of view as to which measure would be less onerous, and, unless the view of the Member State’s government that its measure is the more appropriate is manifestly wrong, the court should not substitute its own view for that of the government.

Whilst the MR could see some force in the appellants’ criticisms of the ban, the justification was not “so weak or illogical” as to justify the Court interfering (§237). Nor could the decision to impose the ban be required to be supported by a net cost-benefit figure in a quantitative analysis – such a mechanistic approach was not part of the case law or supported by principle or common sense.  So, although such an analysis had been performed by the Government, and although the Court was sympathetic to criticisms of it, that was by-the-by.  After referring to various possible disadvantages of a voluntary code over a ban, the MR concluded that it was not irrational for the SoS to have decided to proceed with a ban, even in the light of the possibility of a voluntary code (§248).  And although there was some force in the criticism that the SoS had not directly considered a voluntary code as an alternative, the Court had to “concentrative on the substantive merits rather than on the procedural aspects” (§250).  In his view:

“It would be taking the law further than it has been taken by the Court of Justice if we were to hold that a Government measure infringed proportionality simply because another, less onerous, alternative was not considered, in circumstances where it is apparent that the Government reasonably took the view that that alternative would significantly fall short of the measure in terms of achieving the aim sought to be achieved.”


The judgments display a clear divergence of views as to how to assess the proportionality of measures where the executive is exercising powers conferred on it by the legislature. Should the executive be given greater leeway in such circumstances? This distinction is not one drawn by the case law of the Court of Justice, at any rate explicitly, but has frequently been relied on by the domestic courts.

In any event, the fact that there may be a certain margin of manoeuvre for the decision-maker where complex policy decisions are concerned does not obviate the need to consider whether less restrictive ways of achieving the same legitimate objective exist; with respect, the MR’s view is to be preferred to that of Arden LJ. Indeed, it is a cornerstone of the proportionality principle that the least restrictive measure be taken, and it must be for the decision-maker to address its mind to the issue.

The critical question, therefore, is how far the decision-maker needs go in considering alternatives.  It is here that the MR and Arden LJ converge, both taking a hands-off approach. The MR frames the issue in terms of the reasonableness of the SoS’s view as to the alternative; for Arden LJ, it was more that the appellant had not discharged the burden of proving the adequacy of the alternative. Of course, the “alternative” for them was a voluntary code – statutory regulation had been removed from the equation by Parliament. Laws LJ, on the other hand, considered that the SoS had to give consideration to regulation through ARM even though he had no power under the primary legislation to impose it and even though he may ultimately reject it. This was rejected by the majority: the focus should be on the substantive merits rather than on process (for once, perhaps, the more searching scrutiny of EU judicial review came to the rescue of the decision-maker).

On one point, however, all members of the Court agreed: there was certainly no obligation on the SoS to engage in respective financial cost/benefit analyses of the various alternatives: as Arden LJ put it, “It is surely difficult to put a price on human health, just as it is on human happiness. The benefits of health policy include changes in behaviour and lifestyle, which enable people to lead longer and more fulfilling lives” (§178).

In the end, it is difficult to determine the exact ratio of the judgment in Sinclair Collis.  But what can be said with confidence is that whilst the proportionality doctrine has now, thanks to both EU and ECHR law, been part of domestic law for decades, its precise meaning and application in individual cases is often very difficult to determine.

(As a concluding footnote, it might be noted that in May 2011 the Scottish Court of Session (Outer House) came to a similar conclusion to that of the Court of Appeal in respect of the ban in Scotland of vending machines: see [2011] CSOH 80.)


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