Scotch Whisky Association and others v The Lord Advocate and another [2017] UKSC 76

Michaelmas Term

[2017] UKSC 76

On appeal from: [2016] CSIH 77

JUDGMENT

Between

Scotch Whisky Association and othersAppellants

The Lord Advocate and anotherRespondents

Before

Lord Neuberger

Lady Hale

Lord Mance

Lord Kerr

Lord Sumption

Lord Reed

Lord Hodge

Date 15 November 2017

Hearing Heard on 24 and 25 July 2017

Aidan O’Neill QC Morag Ross QC (Instructed by Brodies LLP) for Appellants

James Wolffe QC, The Lord Advocate, Gerry Moynihan QC, Lesley Irvine (Instructed by Scottish Government Legal Directorate Litigation Division) for Respondent (Lord Advocate)

Philip Simpson QC, John MacGregor (Instructed by Office of the Advocate General) for Respondent (Advocate General)

Introduction

[1]The Scottish Parliament has determined to address health and social consequences which can arise from the consumption of cheap alcohol. The mechanism chosen is minimum pricing. The Alcohol (Minimum Pricing) (Scotland) Act 2012 (“the 2012 Act”) will, when in effect, amend Schedule 3 of the Licensing (Scotland) Act 2005 by inserting in the licence which any retail seller of alcohol in Scotland must hold, an additional condition, to the effect that an alcohol product must not be sold at a price below a statutorily determined minimum price per unit of alcohol. The minimum price is to be set by the Scottish Ministers by secondary legislation. The current proposal is that it should be 50 pence per unit of alcohol. The Scottish Ministers have undertaken not to bring the 2012 Act into force or to make any order setting a minimum price until final determination of the present proceedings. The 2012 Act contains a requirement for the Scottish Ministers to evaluate and report to the Scottish Parliament on the operation and effect of the minimum pricing provisions after five years, and a provision terminating the operation of those provisions automatically after six years, unless the Scottish Ministers by order affirmed by the Scottish Parliament determine that the minimum pricing régime should continue.

[2]The proceedings are brought by three petitioners: The Scotch Whisky Association and two Belgian organisations which I can for economy call “the European Spirits Organisation” and “the Comité Européen des Entreprises Vins”. Their case has been presented by Mr Aidan O’Neill QC. The respondents are the Lord Advocate representing the Scottish Ministers and the Advocate General for Scotland representing the United Kingdom government. In the petitioners’ submission, the 2012 Act and the proposed system of minimum pricing are contrary to European Union law, and so outside the competence of the Scottish Parliament and the Scottish Ministers by virtue of sections 29(2)(d) and 57(2) of the Scotland Act 1998. This (with other objections not now pursued) was rejected by Lord Doherty in the Outer House:

[2013] CSOH 70

; 2013 SLT 776. On appeal to the Inner House, the Extra Division on 3 July 2014 referred six questions to the Court of Justice. In response, Advocate General Bot delivered his opinion on 3 September 2015, and the Court of Justice gave its judgment on 23 December 2015: (Case C333/14)

[2016] 1 WLR 2283

. On the matter returning to the First Division for determination, the appeal was on 21 October 2016 dismissed for reasons given in a single judgment of the court given by the Lord President, Lord Carloway:

[2016] CSIH 77

;

[2017] 1 CMLR 41

. The matter now comes to the Supreme Court with permission granted by the First Division.

[3]There are two limbs to the petitioners’ challenge under EU law to the 2012 Act and to the principle of minimum pricing. First, it is submitted that they conflict with article 34 of the Treaty on the Functioning of the European Union (“TFEU”), providing that:

“Quantitative restrictions on imports and all measures having equivalent effect shall be prohibited between member states.”

It is accepted that the proposed minimum pricing is a measure which would have equivalent effect to a quantitative restriction on imports, in that it will have an effect on, for example, actual or potential wine or beer imports from a number of other EU States. The respondents’ response is reliance on article 36 TFEU, providing:

“The provisions of articles 34 and 35 shall not preclude prohibitions or restrictions on imports, exports or goods in transit justified on grounds of … the protection of health and life of humans … Such prohibitions or restrictions shall not, however, constitute a means of arbitrary discrimination or a disguised restriction on trade between member states.”

[4]The second limb concerns wine only, and arises from Regulation (EU) No 1308/2013 (“the Single CMO Regulation”) establishing a common organisation of markets in agricultural products including wine. The objectives of the common agricultural policy (“CAP”) as set out in article 39 TFEU, include increasing agricultural productivity, stabilising markets, assuring the availability of supplies and ensuring that supplies reach consumers at reasonable prices. Common market organisations (“CMOs”) “are based on the concept of an open market to which every producer has free access under conditions of effective competition”: so the Court of Justice said in its judgment in this case at para 22. The Advocate General and Court of Justice both also accepted that a member state may adopt measures pursuing the objective of protection of human life and health, although they undermine “the system, on which the Single CMO Regulation is founded, of free formation of prices in conditions of effective competition”: paras 25-27 of the Court of Justice’s judgment. But the petitioners submit that this involves a different exercise to that arising under articles 34 and 36, in particular a different and potentially more onerous weighing of the proportionality of the measure.

The Court of Justice’s judgment

[5]Both limbs have to be examined on the basis of the guidance given by the Court of Justice. The Advocate General was clear in his advice. He took first the position under the Single CMO Regulation. He said:

“44. … I consider that the existence of a CMO covering the wine sector does not prevent the national authorities from taking action in the exercise of their competence in order to adopt measures to protect health and, in particular, to combat alcohol abuse. However, where the national measure constitutes a breach of the principle of the free formation of selling prices that constitutes a component of the single CMO Regulation, the principle of proportionality requires that the national measure must actually meet the objective of the protection of human health and must not go beyond what is necessary in order to attain that objective. 45. As the commission suggests, I consider that the examination of the proportionality of the measure must be undertaken in the context of the analysis that must be carried out by reference to article 36 TFEU. 46. Consequently, I propose that the answer to the first question should be that the single CMO Regulation must be interpreted as meaning that it does not preclude national rules, such as those at issue, which prescribe a minimum retail price for wines according to the quantity of alcohol in the product sold, provided that those rules are justified by the objectives of the protection of human health, and in particular the objective of combating alcohol abuse, and do not go beyond what is necessary in order to achieve that objective.”

[6]Turning to articles 34 and 36, he noted that the proposed minimum pricing appeared to be contrary to article 34, on which basis the next step was to consider whether this was justified under article 36. As to this, he said:

“71. A barrier to the free movement of goods may be justified on one of the public interest grounds set out in article 36 TFEU or in order to meet overriding requirements. In either case, the restrictions imposed by the member states must none the less satisfy the conditions laid down in the court’s case law as regards their proportionality. 72. In that regard, in order for national rules to comply with the principle of proportionality, it is necessary to ascertain not only whether the means which they implement are appropriate to ensure attainment of the objective pursued, but also that those means do not go beyond what is necessary to attain that objective: Berlington Hungary Tanácsadó és Szolgáltató kft v Magyar Állam (Case C-98/14) [2015] 3 CMLR 45, para 64. 73. Although the words generally used by the court seem most frequently to result in only two different stages of the control of proportionality being distinguished, the intellectual exercise followed in order to determine whether a national measure is proportionate is generally broken down into three successive stages. 74. The first stage, corresponding to the test of suitability or appropriateness, consists in ascertaining that the act adopted is suitable for attaining the aim sought. 75. The second stage, relating to the test of necessity, sometimes also known as the ‘minimum interference test’, entails a comparison between the national measure at issue and the alternative solutions that would allow the same objective as that pursued by the national measure to be attained but would impose fewer restrictions on trade. 76. The third stage, corresponding to the test of proportionality in the strict sense, assumes the balancing of the interests involved. More precisely, it consists in comparing the extent of the interference which the national measure causes to the freedom under consideration and the contribution which that measure could secure for the protection of the objective pursued.”

[7]He went on to make the important point that “judicial review of the proportionality of the measure should be marked by a certain degree of restraint” (para 82). This was for two reasons: “83. First, account should be taken of the fact that it is for the member states to decide on the degree of protection which they wish to afford to public health and on the way in which that degree of protection is to be achieved. Since the level of protection may vary from one member state to another, member states must be allowed discretion in that area … That discretion is necessarily represented by a certain relaxation of control, representing the national court’s concern not to substitute its own assessment for that of the national authorities. 84. Second, it is necessary to take into account the complexity of the assessments to be carried out and the degree of uncertainty which exists as to the effects of measures such as those at issue.” He added that a third relevant consideration in the present case was the provision for a re-evaluation and report by the Scottish Ministers after five years, coupled with the provision for automatic termination after six years unless otherwise ordered and affirmed by the Scottish Parliament (para 85).

[8]However, he added this caution:

“86. … [T]he discretion left to the member states cannot have the effect of allowing them to render the principle of free movement of goods devoid of substance. In so far as article 36 TFEU includes an exception to that principle, it is for the national authorities, even where they have a discretion, to show that the measure satisfies the principle of proportionality. … 87. Furthermore, whatever the extent of that discretion, the fact none the less remains that the reasons that may be invoked by a member state by way of justification must be accompanied by an analysis of the suitability and proportionality of the restrictive measure adopted by that state and of the precise evidence on which its argument is based ...”

[9]The Court of Justice did not either repeat or endorse the Advocate General’s above advice, but spoke in terms which give some room for argument, both as to the relationship between the principles applicable to the two limbs of the petitioners’ case, and as to the nature of any proportionality exercise which it envisaged fell to be performed under either or both of these limbs. Addressing the significance of the Single CMO Regulation, the Court (in its paras 28 and 29) adhered firmly to what Advocate General Bot had described (in his para 73: see para 6 above) as its previous general usage, distinguishing only two different stages of the proportionality test. The difficulty this raises is to know what, if any, scope there is for a more general third stage proportionality question, of the nature described by Advocate General Bot in his paras 76 and 82 to 84: see paras 6 and 7 above). The Court’s guidance in this respect is oblique, as appears from the last sentence of para 28 and from the summary in para 29 of its judgment. No doubt deliberately, the Court there suggests that the third stage, rather than involving any independent balancing of interests, can be subsumed within the second stage, that is consideration of what is necessary to achieve the desired protection of human life and health.

[10]The material parts of paras 28 and 29 of the Court’s judgment read as follows:

“28. A restrictive measure such as that provided for by the national legislation at issue must, however, satisfy the conditions set out in the court’s case law with respect to proportionality, that is, the measure must be appropriate for attaining the objective pursued, and must not go beyond what is necessary to attain that objective (see, by analogy, Berlington Hungary (Case C-98/14) [2015] 3 CMLR 45, para 64), which the Court will consider in its examination of the second to sixth questions, which specifically concern the analysis of the proportionality of that legislation. It must be observed that, in any event, the issue of proportionality must be examined by taking into consideration, in particular, the objectives of the CAP and the proper functioning of the CMO, which necessitates that those objectives be weighed against the objective pursued by that legislation, namely the protection of public health. 29. Consequently, the answer to the first question is that the Single CMO Regulation must be interpreted as not precluding a national measure, such as that at issue, which imposes an MUP for the retail selling of wines, provided that that measure is in fact an appropriate means of securing the objective of the protection of human life and health and that, taking into consideration the objectives of the CAP and the proper functioning of the CMO, it does not go beyond what is necessary to attain that objective of the protection of human life and health.”

[11]Turning to articles 34 and 36 TFEU, the Court was satisfied that the proposed minimum pricing regime appeared to be an appropriate means of attaining the objective it pursued (identified as increasing the price of cheap alcoholic drinks, so reducing the consumption of alcohol, in general, and the hazardous and harmful consumption, of alcohol, in particular): paras 36 and 39. It went on (para 40):

“As regards whether that national legislation does not go beyond what is necessary in order effectively to protect human life and health, it must be borne in mind that, in this case, that analysis must be undertaken, as stated in para 28 of this judgment, with regard to the objectives of the CAP and the proper functioning of the CMO. However, given the issue to be examined in this case, that analysis will have to be undertaken with reference to proportionality in the context of article 36 TFEU and will therefore not have to be carried out separately.”

[12]Again, this appears to subsume any third stage within the context of the second stage enquiry relating to necessity. It also indicates that the requirement, in that context, to refer to “the objectives of the CAP and the proper functioning of the CMO” adds nothing to the criteria which fall to be taken into account when deciding whether article 36 is satisfied. The petitioners’ case, that there is some important difference between the exercise to be undertaken under articles 34 and 36 and the exercise to be undertaken in relation to wine in the light of the Single CMO Regulation does not appear consistent with the Court of Justice’s guidance.

[13]The remaining paragraphs of the Court of Justice’s judgment are also noticeable for their focus on the issue now before the Supreme Court in terms of the first and second stages of the proportionality test which Advocate General Bot described. The Court thus stated:

“53. … [I]t is for the national authorities to demonstrate that that legislation is consistent with the principle of proportionality, that is to say, that it is necessary in order to achieve the declared objective, and that that objective could not be achieved by prohibitions or restrictions that are less extensive, or that are less disruptive of trade within the European Union: Criminal proceedings against Franzén (Case C-189/95) [1997] ECR I-5909, paras 75 and 76 and Rosengren v Riksåklagaren, para 50. 54. In that regard, the reasons which may be invoked by a member state by way of justification must be accompanied by appropriate evidence or by an analysis of the appropriateness and proportionality of the restrictive measure adopted by that state, and specific evidence substantiating its arguments … 55. It must however be stated that that burden of proof cannot extend to creating the requirement that, where the competent national authorities adopt national legislation imposing a measure such as the MUP, they must prove, positively, that no other conceivable measure could enable the legitimate objective pursued to be attained under the same conditions: Commission v Italian Republic [2009] All ER (EC) 796, para 66. 56. In that context, it is for the national court called on to review the legality of the national legislation concerned to determine the relevance of the evidence adduced by the competent national authorities in order to determine whether that legislation is compatible with the principle of proportionality. On the basis of that evidence, that court must, in particular, examine objectively whether it may reasonably be concluded from the evidence submitted by the member state concerned that the means chosen are appropriate for the attainment of the objectives pursued and whether it is possible to attain those objectives by measures that are less restrictive of the free movement of goods. 57. In this case, in the course of such a review, the referring court may take into consideration the possible existence of scientific uncertainty as to the actual and specific effects on the consumption of alcohol of a measure such as the MUP for the purposes of attaining the objective pursued. As Advocate General Bot stated in point 85 of his opinion, the fact that the national legislation provides that the setting of an MUP will expire six years after the entry into force of the 2013 Order, unless the Scottish Parliament decides that it is to continue, is a factor that the referring court may also take into consideration. 58. That court must also assess the nature and scale of the restriction on the free movement of goods resulting from a measure such as the MUP, by comparison with other possible measures which are less disruptive of trade within the European Union, and the effect of such a measure on the proper functioning of the CMO, that assessment being intrinsic to the examination of proportionality. 59. It follows from the foregoing that article 36 TFEU must be interpreted as meaning that, where a national court examines national legislation in the light of the justification relating to the protection of the health and life of humans, under that article, it is bound to examine objectively whether it may reasonably be concluded from the evidence submitted by the member state concerned that the means chosen are appropriate for the attainment of the objectives pursued and whether it is possible to attain those objectives by measures that are less restrictive of the free movement of goods and of the CMO.”

[14]Paragraph 59 was in substance repeated as para 3 of the Court’s ruling. Paragraph 59 echoes the two-stage approach to proportionality stated in para 56. The explanation that the court is bound to or must “examine objectively whether it may reasonably be concluded from the evidence submitted” that the means are appropriate and cannot be attained by less restrictive measures can be seen as recognising the fact that the national court is a reviewing body, not the primary decision-maker. Paragraph 57, with its reference back to para 85 of the Advocate General’s opinion, enables the reviewing court to bear in mind the uncertainties and experimental nature of the proposed minimum pricing system. Paragraph 58 might be read as suggesting a third stage proportionality issue. But the injunction to assess the nature and scale of the restriction is in terms only in order to compare them with the effects of other possible measures, and so to determine whether there are other measures less destructive of EU trade. Once it is accepted, as found here by the Lord Ordinary, that an approach based on increased taxation would be less destructive of EU trade, para 58 is on the face of it exhausted.

[15]The Court of Justice’s approach to exceptions (such as article 36) to a general principle (such as article 34) gives rise, in these circumstances, to some difficulty. The first two stages of the proportionality exercise address, respectively, the legitimacy of the aim which the legislature had in mind, and the necessity for the measures adopted if such aim is to be achieved (or, putting the latter aspect the other way round, the question whether the aim could be achieved by less extensive or restrictive measures). Neither in terms nor in logic is either stage concerned with the further question whether, on an overall balance, it is worthwhile to achieve the aim, bearing in mind the detriment that achieving it would necessarily cause to the general principle. By suppressing Advocate General Bot’s third stage, one may surmise that the Court of Justice intended at the very least to signal the appropriateness of an even greater level of restraint and respect for national authorities’ choice of measures to protect health than that which Advocate General Bot himself recognised under the third stage test which he identified (see paras 7 Page 11 and 8 above). Yet one may also infer from the Court of Justice’s references in paras 28, 29 and 40 that it intended more general objectives (in particular, those of the CAP and the CMO) to play some role, at least in relation to wine, and perhaps also other commodities. What is unclear is quite what that role might be, and how it really fits within the second stage enquiry into which the Court of Justice has inserted it.

[16]As it happens, the Supreme Court touched on the Court of Justice’s reticence about any third stage enquiry in a judgment given some six months prior to the Court of Justice’s present judgment: R (Lumsdon) v Legal Services Board

[2015] UKSC 41

;

[2016] AC 697

. In a joint judgment by Lord Reed and Lord Toulson, it said (para 33):

“33. Proportionality as a general principle of EU law involves a consideration of two questions: first, whether the measure in question is suitable or appropriate to achieve the objective pursued; and secondly, whether the measure is necessary to achieve that objective, or whether it could be attained by a less onerous method. There is some debate as to whether there is a third question, sometimes referred to as proportionality stricto sensu: namely, whether the burden imposed by the measure is disproportionate to the benefits secured. In practice, the court usually omits this question from its formulation of the proportionality principle. Where the question has been argued, however, the court has often included it in its formulation and addressed it separately, as in R v Minister for Agriculture, Fisheries and Food, Ex p Fedesa (Case C-331/88) [1990] ECR I-4023.”

The Supreme Court’s approach thus corresponded closely with Advocate General Bot’s approach. But this does not help now to explain the Court of Justice’s evidently deliberate suppression of the third stage in the present case, coupled with the insertion of one aspect of it in the limited context of the second stage test of necessity. I will have to consider how far this is significant on this appeal at a later stage in this judgment.

The issues in more detail

[17]It is common ground on this appeal that the role of a domestic court, evaluating the consistency with European law of a measure such as the 2012 Act, is not to examine or adjudicate upon the legislative process and reasoning which led to the measure, but “to examine the legislation itself in its context” (see per Lord Thomas of Cwmgiedd in In re Recovery of Medical Costs for Asbestos Diseases (Wales) Bill

[2015] UKSC 3

;

[2015] AC 1016

, para 126). The Court of Justice held (paras 63 to 65) that this examination fell to be carried out in the light of all the material available on the date when the court gives its ruling. That was the position when the matter came before the Outer House. The position on an appeal depends, as the First Division held (para 109), upon the domestic rules applicable upon appeals. In the present context of judicial review, the First Division went on to hold, and this is not now controversial, that an appellate court is “entitled to have regard to new material where it considers, in its discretion, that the interests of justice require that it be taken into account” (para 109). On this basis, a considerable amount of new material was considered by the First Division and is before the Supreme Court.

The issues

[18]The actual issues have narrowed. There is no suggestion that the proposed minimum pricing system will constitute “a means of arbitrary discrimination or a disguised restriction on trade between member states” within the last sentence of article 36 FTEU. But the respondents accept that it will affect the market in alcohol generally, including wine, and (although they maintain that the greater effect will be domestic) they also accept that imports and trade between EU member states will be impacted. The position is, therefore, that it is for the respondents to justify the EU market interference under article 36 TFEU and under the parallel principles governing wine under the CAP and Single CMO Regulation. There is also common ground, reflected in the agreed statement of facts and issues, that the 2012 Act had and has a two-fold objective. The petitioners accept the legitimacy of this objective, and they accept that minimum pricing at a rate of 50 pence per unit is an appropriate means of attaining that legitimate objective. However, the precise implications or qualifications of the agreed objective are important and, are not necessarily matters on which the parties are ad idem, and they still require examination.

The objective(s) pursued by minimum pricing

(s) pursued by minimum pricing

[19]The two-fold objective was, as put to the Court of Justice, “reducing, in a targeted way, both the consumption of alcohol by consumers whose consumption is hazardous or harmful, and also, generally, the population’s consumption of alcohol”: Court of Justice, para 34. Hazardous drinkers are in this context defined as males consuming more than 21 units and women consuming more than 14 units of alcohol a week, while harmful drinkers are defined as males drinking more than 50 units and women drinking more than 35 units a week. Both the Lord Ordinary (para 53) and the First Division (paras 171 to 172) proceeded on the basis of this agreed aim. However, the petitioners suggested to the First Division and suggest before the Supreme Court that the respondents’ justification for minimum pricing has deviated from this agreed aim, and, in particular, that they have in reality advanced a more limited aim, relating to extreme drinkers and/or the elimination of health inequality, in order to justify the 2012 Act.

[20]Even if one confines attention to the initiation of the 2012 Act, the agreed two-fold objective is more refined than might at first sight appear. The key word in the Court of Justice’s description is in this context the word “targeted”. The Scottish Government had since 2009 been aiming to address alcohol-related harm by a whole variety of measures set out in Changing Scotland’s Relationship with Alcohol (2009). The 2012 Act aimed at the particular problems created by low price alcohol. It followed a study entitled Final Business and Regulatory Impact Assessment for Minimum Price per Unit of Alcohol as contained in Alcohol (Minimum Pricing) (Scotland) Bill (“the BRIA”). The BRIA drew on a very wide range of other expert studies, including work commissioned by the Scottish Government from the University of Sheffield, Model-based Appraisal of Alcohol Minimum Pricing and Off-Licensed Trade Discount Bans in Scotland (2009, version 2: April 2010 and second update: January 2012), analysing (amongst many other aspects) the price elasticities of alcohol demand and the impact of minimum pricing as against increased taxation. The BRIA noted that Scottish per capita alcohol sales were almost a quarter higher than in England (para 2.14) and that the average consumption of alcohol in a population was directly linked to the amount of harm, in terms of illness, violence and injury and other forms of social harm (paras 2.18 to 2.29). Alcohol-related general hospital discharges and mortality rates have risen substantially over thirty years, and chronic liver disease and cirrhosis mortality rates in Scotland are way above those in England and Wales or other European countries (figures 3, 4 and 5). Paragraph 2.18 of the BRIA put the general point simply, with footnote references to prior studies:

“The average consumption of alcohol in a population is directly linked to the amount of harm as evidenced in a number of systematic reviews. The more we drink, the greater the risk of harm. As overall consumption has increased in Scotland so have the resultant harms.”

[21]However, the BRIA also recognised that the true relationship between consumption and harm was more complex, and involved other factors (particularly poverty and deprivation) of potential relevance to minimum pricing. It said, significantly, in this connection (para 2.29) that:

“Whilst alcohol-related issues impact on all socio-economic groups, it is important to recognise the greatest harm is experienced by those who live in the most deprived areas. The reasons why alcohol has a more harmful effect on people living in deprived communities are complex and not fully understood. Risky and harmful alcohol use is likely to be both a cause and effect of social deprivation. What is clear is that the level of alcohol-related harm in deprived communities is substantial, with alcohol-related general hospital discharge rates in the 20% most deprived communities (as measured by the Scottish Index of Multiple Deprivation, SIMD) around 7.5 times higher than in the most affluent fifth. Similarly, alcohol-related mortality rates are 6 times higher in the most deprived areas. Tackling alcohol-related harm has the potential to help address Scotland’s wider health inequalities.”

[22]Paragraph 2.29 of the study was taken up in a later section of the study identifying various benefits envisaged from minimum pricing. Under the heading Health Benefits for those on low incomes, para 5.24 noted that there were (at that time) insufficient data to enable the reduction in health harms across different income groups to be modelled, but that a NHS Health Scotland report (Monitoring and Evaluating Scotland’s Alcohol Strategy. Setting the Scene: Theory of change and baseline picture by Beeston, Robinson, Craig and Graham) had “confirmed strong income/deprivation patterns to alcohol-related health harm”. Para 5.24 went on to repeat the ratios quoted in para 2.29 for alcohol-related hospital discharges and mortality rates in the most deprived and most affluent communities (7.5 times and 6 times respectively). It added that:

“significantly, average weekly consumption among low income harmful drinkers was much higher than among other harmful drinkers (93 units for men and 69 for women compared to 69 and 52 units respectively for harmful drinkers in the highest income group). This helps to explain the differential harm patterns described above. In addition those on low incomes are likely to be more responsive to minimum pricing. Given this, it is therefore likely that those in lower income/more deprived groups will benefit from the greatest reduction in health harms.”

[23]The 2012 Bill, leading to the 2012 Act, was accompanied by Explanatory Notes and a Policy Memorandum, both of which identified a range of health and social and economic benefits envisaged as resulting from minimum pricing. The Policy Memorandum specifically picked up the alcohol-related hospital discharge and mortality ratios referred to in the BRIA, noting that “the Scottish Government believes alcohol plays a significant part in these inequalities” (para 10).

[24]It is therefore clear that, from the outset, concern about the health and social harms resulting from extremely heavy drinking in deprived communities was an element of targeted thinking behind the 2012 Act. The Policy Memorandum also discounted a straightforward increase in excise tax as it “would impact on high price products as well as cheap ones and so would have a proportionately greater effect on moderate drinkers than a minimum price” (para 29). The 2012 Act was, in this respect, envisaged as a balanced measure which would not target the cost of drinking generally without regard for the extra costs which this would impose on drinkers. Its aims were, as Lord Doherty found, “directed principally towards the protection of health and life, though other consequential (largely public order and economic) benefits [were] also anticipated” (para 53), and it was clear that it was not an aim that alcohol consumption be either “eradicated” or that its costs should be made “prohibitive for all drinkers (para 54). It was “intended to strike at alcohol misuse and overconsumption”, in which connection the major problem was “excessive consumption of cheap alcohol”, which the proposed measures sought to address by increasing the price of such alcohol (para 54).

[25]Even in 2013, Lord Doherty was also able to find (para 59) that:

“the harmful drinkers in the lowest income quintile consume far more alcohol per head, and are the source of much greater health related and other harm, than harmful drinkers in the higher income quintiles. There is also clear evidence that the greatest alcohol-related harm is experienced by those who live in the most deprived areas (see the evidence summarised in para 2.29 of the Final BRIA).”

And he went on to conclude, at para 60, that there was objective evidence that the proposed minimum pricing measures “are appropriate to achieve their aims”.

[26]Since the BRIA study, more work has been done to fill the lacuna to which para 5.24 referred. This consists in a University of Sheffield report Model-based appraisal of the comparative impact of Minimum Unit Pricing and taxation policies in Scotland of April 2016. This identified a number of facts not previously evident. One was that, applying the definitions mentioned above, the great majority of both hazardous and harmful drinkers were not in poverty - 20% and 6% respectively of the whole drinker population as opposed to 2% and 1% of the whole drinker population who were in poverty: table 4.3. But another side of this coin is that hazardous and harmful drinkers in poverty drink more than those not in poverty: 1,456 as against 1,396 units per annum on average in the case of hazardous drinkers and 4,499 as against 3,348 units in the case of harmful drinkers; and the link between those in poverty and cheap alcohol is clear from the fact that, although they drink noticeably more, hazardous drinkers in poverty spend less, and harmful drinkers in poverty spend only very slightly more, than those not in poverty. This corresponds with the evident likelihood, which had been accepted by Lord Doherty in the Outer House (para 57), that poorer drinkers tend to drink cheaper alcoholic drinks than better off drinkers. A further study by the University of Sheffield shortly after the passing of the 2012 Act revealed (as recorded by the Extra Division in its reference to the Court of Justice, para17) a marked difference in the average number of cheaper priced alcoholic drinks purchased by lowest and highest income quintile drinkers. The study revealed that harmful and hazardous drinkers in the lowest income quintile purchased respectively 30.8 and 7.8 units of such alcohol weekly, an average decreasing with each quintile, with harmful and hazardous drinkers in the highest quintile only purchasing respectively 13.6 and 2.7 of such units weekly. Although directed to drinks priced at less than 45 pence, rather than 50 pence, per unit of alcohol, the position in relation to drinks priced at less than 50 pence is unlikely to differ fundamentally. Still more strikingly and sadly, hazardous and harmful drinkers in poverty are involved in far more alcohol-related deaths and hospital admissions than those not in poverty. Relevant deaths and hospital admissions were for hazardous drinkers in poverty 206 and 4,563 per 100,000 drinkers as against only 83 and 1,539 respectively for hazardous drinkers not in poverty. Relevant deaths and hospital admissions for harmful drinkers in poverty were 781 and 11,555 per 100,000 drinkers as against only 371 and 6,454 respectively for harmful drinkers not in poverty.

[27]The University of Sheffield study went on to model the effect of a 50 pence per unit of alcohol minimum price on drinkers in poverty and not in poverty. It concluded that annual consumption by harmful drinkers in poverty would experience a fall of 681 units (as compared with nearly 181 units for such drinkers not in poverty), while consumption by hazardous drinkers in poverty would experience a fall of just under 88 units (as compared with a fall of only 30 units for such drinkers not in poverty). There would be 2,036 fewer deaths and 38,859 fewer hospitalisations during the first 20 years of the policy, after which when the policy had achieved its full impact, there would be an estimated 121 fewer deaths and 2,042 fewer hospital admissions each year.

[28]The 2012 Act is not yet in force, but is the subject of the present on-going proceedings in which the petitioners challenge, while the respondents seek to establish the validity of its introduction under European law. All the above material is now before the court, and is admissible on the issue of justification and proportionality. Under these conditions, it would seem artificial, and even unfair, to allow the petitioners to rely on the new material to try to undermine the justification for any aims originally advanced, but not to allow the respondents to refine the aims advanced and to demonstrate that, on the material now available, the proposed measure is justified, even if it only meets an aim which is narrower than, but still falls within the scope of those originally advanced. Accordingly, even if it is right that some of the broader assumptions about correlations between hazardous and harmful drinking and health and other social problems are not sustainable, it seems to me open to the respondents to rely on the new material as reinforcing an entirely valid correlation, developed from the outset, between the health and social problems arising from extreme drinking by those in poverty in deprived communities. The respondents are in this respect doing no more than explaining how the 2012 Act will target the particular health and social problems arising from such drinking which the new material has demonstrated.

Less restrictive measures to achieve the same aim?

[29]The focus of submissions on this appeal has been directed not to the question whether a system of minimum pricing per unit of alcohol is capable of meeting the agreed aims, including that relating to social deprivation which I have been discussing. The submissions have rather focused on the issue whether such aims could be attained by less restrictive measures. As I have indicated, but contrary to the petitioners’ case, this appears in the light of the Court of Justice’s judgment to be the same issue as whether, taking into account the objectives of the CAP and Single CMO Regulation, the proposed system is necessary to attain such aims. The petitioners object that the respondents have failed to produce appropriate and/or specific evidence or analysis to satisfy the onus on them to justify the prima facie infringement of the European legal prohibition on measures with equivalent effect to quantitative restrictions on imports and measures inhibiting free trade and effective competition. They also submit that, even on the material available, the respondents cannot show the proposed minimum pricing to be necessary to achieve the intended aims and cannot, in particular, show that there are no other ways of achieving those aims without infringing the above European legal prohibition.

[30]The core comparison here is between minimum pricing and some form of excise or tax. The comparison falls to be made on the basis that an excise or tax charge would involve less of an obstacle to free movement of goods between EU member states and competition. This is because Lord Doherty held that the respondents had not made out any case to the contrary. It is worth noting that, although it is for the domestic court to form its own conclusions as to the existence of any alternative measure(s) which would achieve the same objective(s) as minimum pricing, this is a question which was from the outset at the forefront of the Scottish Government’s mind when determining to adopt a system of minimum pricing. It is a question which was addressed in detail in para 4.3 of the BRIA and in paras 28 to 35 of the Policy Memorandum which accompanied the Bill leading to the 2012 Act. Those paragraphs are still very largely relevant to the current issues.

[31]The petitioners’ basic proposition is that an increased excise duty could achieve a similar improvement in mortality and hospital admission statistics to that envisaged by the minimum pricing system currently proposed, as set out in para 27 above. Mr O’Neill referred to a February 2016 paper by the same authors as the University of Sheffield’s later April 2016 study. That paper reported the results of a study based on an econometric epidemiological model constructed by reference to English conditions in 2014/2015. The study was to assess the differential effects of four policies on population sub-groups defined by drinking level and income or socioeconomic group. In this context, it equated the effects on health of a 13.4% increase in excise duty with those of a 50 pence per unit minimum pricing approach. Bearing in mind acknowledged differences between the scale and pattern of drinking in England and Scotland, the comparison and equation are, as the Lord Advocate submitted, not illuminating. What is worth noting is the authors’ observation that, although the predicted outcomes were overall similar, they were achieved in different ways:

“While all policies were estimated to reduce health inequalities because drinking is associated with substantially higher absolute health risks in lower socioeconomic groups than in higher socioeconomic groups, the scale of the inequality reduction varied across the policies. A £0.50 minimum unit price and a £0.22 per unit volumetric tax were estimated to reduce inequalities the most because heavy drinkers in lower socioeconomic groups buy proportionately more of the cheap alcohol most affected by these policies. Estimated impacts on health inequalities were smaller for a 4.0% alcohol ad valorem tax and a 13.4% current duty increase as price increases were more evenly distributed across the alcohol consumed by different socioeconomic groups.”

[32]The relevant study for present purposes is the University of Sheffield’s April 2016 study. It was designed with specific reference to Scottish conditions, and the conclusions it reached on the modelled effect of alcohol tax increases were as follows:

“M14 At full effect, a 50p MUP is estimated to lead to 117 fewer alcohol-related deaths per year among hazardous and harmful drinkers. To achieve the same reduction in deaths among hazardous and harmful drinkers, an estimated 28% increase in alcohol taxes is required. M15 If reductions in alcohol-related harm in specific population groups are sought, then larger tax increases would be required; for example, a 36% tax increase would be required to achieve the same reductions in deaths among harmful drinkers as a 50p MUP. This is because MUP targets large price increases on those at greatest risk from their drinking while tax increases affect all drinkers. M16 Although achieving the same reduction in deaths among hazardous and harmful drinkers as a 50p MUP, a 28% tax increase would lead to slightly larger reductions in alcohol consumption among moderate and hazardous drinkers but smaller reductions in alcohol consumption among harmful drinkers and, particularly, harmful drinkers in poverty. M17 Similarly, at full effect, the reductions in deaths under a 28% tax increase would be larger among hazardous drinkers and smaller among harmful drinkers, particularly harmful drinkers in poverty, than under a 50p MUP price. M18 These differences in how death reductions are distributed across the population mean a 50p MUP is more effective than a 28% tax increase in reducing alcohol-related health inequalities. This is because a 50p MUP better targets the alcohol consumed by harmful drinkers on low incomes who are the group at greatest risk from their drinking. M19 Increases in consumer spending on alcohol are estimated to be substantially greater under a 28% tax increase than a 50p MUP. For example, among moderate drinkers annual per capita spending would increase by £2 or 0.5% under a 50p MUP and by £17 or 4.7% under a 28% tax increase. For harmful drinkers the annual increases in spending per capita are £6 or 0.2% for a 50p MUP and £152 or 6.4% under a 28% tax increase.”

[33]On the basis of all the material before him, the Lord Ordinary considered (in paras 67 to 81 of his judgment) whether a minimum pricing system was necessary to achieve the agreed aims, or whether alternative means involving increased excise or tax would be just as effective. The whole of the Lord Ordinary’s discussion of the point is valuable, but I shall highlight three principal themes.

[34]First, he noted (para 67) that minimum pricing targets cheap alcohol products by reference to their alcohol content, whereas the effect of an increased excise or VAT charge is felt across the board on the whole category of goods to which it applies. In this connection, he rejected the argument that an effective price rise across the board would reduce consumption generally in accordance with the agreed aims (para 77), because “the legitimate aims of the measure” had not been “to reduce consumption, including consumption by hazardous and harmful drinkers, to the maximum extent possible regardless of possible economic or social consequences.” Rather, they were those he had identified in paras 53 to 54 of his judgment, set out in para 24 above. There was a relevant judgment as to which it was for the Scottish legislature and Ministers to make, what level of protection for health and life to achieve, by striking a balance between health and other interests: para 79. Second, the relevant EU directives meant that excise duty could not be used to achieve the same outcomes as minimum pricing: paras 68 and 71. Third, he said that minimum pricing was easier to understand and simpler to enforce: see para 68. It was not open to absorption, eg by “off-trade” outlets such as supermarkets selling alcohol drinks below cost in order to attract other business onto or on their premises.

[35]The petitioners challenge these propositions. As the Lord Ordinary noted, the petitioners seek to make a virtue out of the first proposition, by arguing that higher retail prices across the board can only promote the stated aim of the 2012 to reduce alcohol consumption generally. The Reference made by the Inner House to the Court of Justice was framed in terms which give some encouragement to such an argument, asking as question 5 whether it is a legitimate ground for discarding an alternative measure (in casu, an excise duty increase) that its effects “may not be precisely equivalent to the measure impugned under article 34 TFEU but may bring further, additional benefits and respond to a wider, general aim”. Not perhaps surprisingly in the light of this formulation, Advocate General Bot, in response, saw the fact “that the alternative measure entailing increased taxation is capable of procuring additional advantages by contributing to the general objective of combating alcohol abuse” as no justification for discarding that measure: para 152. However, it is right to add that he had also recognised, at paras 149 and 150, that the Lord Advocate’s case was that the “additional advantages” could only be achieved at a cost, in terms of the across the board rises in prices (for the whole market of suppliers and consumers), which it was the respondents’ case that they considered “disproportionate” and inappropriate to impose. Advocate General Bot expressed himself as “unable to see how that collateral effect … might be seen as negative in the context of combating hazardous or harmful consumption”. The Court of Justice endorsed Advocate General Bot’s approach to the fifth question (paras 47 and 58), whilst emphasising that the ultimate decision whether increased taxation would be capable of protecting human life and health as effectively as minimum pricing is for the United Kingdom courts (paras 49 and 50). Its answer to question 5 (at the end of para 50 and in para 2 of its ultimate ruling) was simply that “The fact that the latter measure may bring additional benefits and be a broader response to the objective of combating alcohol misuse cannot, in itself, justify the rejection of the measure”. The words “in itself” are here significant, because it leaves open the respondents’ case that their general objective of combating alcohol abuse was not one which they intended to pursue at all costs.

[36]The Lord Ordinary accepted the respondents’ case on this point (in para 77 of his judgment, cited in para 34 above). The First Division also accepted it, saying (para 200) that:

“Furthermore, assuming that any practical tax increase within the EU setting would involve across the board increases, albeit perhaps on different types of product, such increases would have a disproportionate, undesirable and unnecessary effect on moderate drinkers, who do not generally represent a significant problem in societal terms, at least of the type requiring to be addressed.”

The First Division also said (at para 181) that:

“The fact that minimum pricing may not, to the same extent, affect those who are more affluent, is of peripheral significance. These richer persons tend not to suffer to the same extent as harmful and hazardous drinkers in the lower quintile of affluence, whose health and life is at greatest risk.”

[37]Mr O’Neill submits that a desire not to impose a tax burden on moderate or other drinkers not at serious health risk cannot itself constitute or justify a measure taken for the protection of health or human life within article 36. That can readily be accepted. But it misses the point, which is that it was never, and is not now, the aim or target of the Scottish Parliament and Ministers to reduce consumption, even by hazardous and harmful drinkers, and still less by moderate drinkers, to the maximum extent possible regardless of possible economic or social consequences: see para 34 above. The more recently available information from the University of Sheffield study of April 2016 merely underlines the appropriateness of a more targeted approach in this connection. It follows that it is legitimate to balance any possible health advantages across the board against the unwanted burden which increased taxation across the board would impose on drinkers falling within the hazardous and harmful categories who are not (for reasons of affluence or whatever) at extreme risk and on moderate drinkers who are at no risk at all. Further, the April 2016 study makes clear that even the level of tax increases which would achieve similar overall reductions in mortality and hospitalisations would not have the same effect in targeting those in poverty, who, as the statistics tellingly show, are the group by far the most heavily affected by extreme drinking and consequent health and social problems. I consider therefore that there is no basis on which the Supreme Court should depart from the Lord Ordinary’s conclusions on this point.

[38]The second point raises for consideration how far the framework of the EU Directives allows a member state, if it wishes, to assimilate by reference to alcoholic content the excise rates applicable to different categories of alcoholic beverage. In Commission of the European Communities v French Republic (

Case C-434/97

)

[2000] ECR I-1129

, para 244, the Court of Justice summarised the difference between VAT and excise as being that the former is levied on price, “whereas excise duty is primarily calculated on the volume of the product”. The position under Council Directive 92/83/EEC on the harmonisation of the structures of excise duties on alcohol and alcoholic beverages (also known as “the structures Directive”) is, on examination, more nuanced. This Directive identifies five categories of alcoholic beverage to which member states must apply an excise duty in accordance with the Directive. These are i) beer (article 2); ii) wine, still and sparkling with an alcoholic strength between either 1.2% and 15% or 15% and 18% (still wine) or 1.2% and 15% (sparkling wine) (article 8); iii) other fermented beverages, still and sparkling (article 12); iv) intermediate products (other products not within articles 2, 8 or 12) with an alcoholic strength between 1.2% and 22% (article 17); and v) ethyl alcohol, defined to cover(a) products falling “within CN codes 2207 and 2208” or

(b) products within “CN codes 2204, 2205 and 2206 which have an actual alcoholic strength by volume exceeding 22%” or

[39]Subject to Directive 92/84/EEC (which sets minimum rates for beer and intermediate products but effectively no minimum rate for wine, since the rate stated is ECU 0 per hectolitre), Directive 92/83/EEC allows different categories to carry different rates. In the case of beer and ethyl alcohol products, the rate stated is, broadly, chargeable according to alcoholic content (articles 3(1) and 21). In the case of fermented beverages and intermediate products, it is to be fixed by reference to the number of hectolitres of finished product (articles 13(1) and 18(1)). Within each category, there are requirements to fix the same rate in respect of the whole category, or in respect of each of certain defined sub-categories. Thus, in relation to wine, article 9(2) requires member states, first, to “levy the same rate of excise duty on all products chargeable with the duty on still wine”, and, second, to “levy the same rate of excise duty on products chargeable with the duty on sparkling wine” (article 9(2)), with the member state being free to decide whether or not the rates for still and sparkling wines should be equated with each other. There are however exceptions to the requirement to have a single rate for each category or sub-category, in the case of lower alcoholic strength beverages; that is: beer with an alcoholic strength not exceeding 2.8% by volume (article 5(1); wine or fermented beverages not exceeding 8.5% (articles 9(3) and 13(3)); intermediate products not exceeding 15%, subject to certain conditions (article 18(3)); and ethyl alcohol products within code 2208 with an alcohol strength not exceeding 10% (article 22(5)). Hence, the low rates applied in the United Kingdom (under article 9(3)) to various defined categories of cider with alcohol content not exceeding 8.5%. There are also exceptions allowing reduced rates under certain conditions for beer brewed by independent small breweries (article 4(1)) and for ethyl alcohol products produced by small distilleries (article 22(1)).

[40]However, it is clear that this framework precludes any complete assimilation by reference to alcoholic strength. A single rate must be levied on all still or sparkling wines with an alcohol content between 8.5% and 15%. Further, a single rate must be levied on each category or sub-category of alcoholic beverage, whatever its retail price. To ensure that the cheapest drinks were sold at a price, inclusive of excise duty and VAT, equivalent of 50 pence per unit of alcohol, the excise rate would have to be set high. But this would mean a correspondingly high excise rate for more expensive drinks which were already being priced at more than 50 pence per unit of alcohol.

[41]Before the Lord Ordinary and the Inner House, the fact that the Scottish Parliament and Ministers had no power to raise taxation on alcoholic drinks was (although referred to at one point as “the elephant in the room”: Inner House, para 192) disregarded on the basis that it arose from the internal division of powers within the United Kingdom. But two assumptions were evidently made, first, that legislation could (by cooperation between the relevant United Kingdom and Scottish Parliaments and/or Governments) be enacted to impose additional excise duty in Scotland alone, but, second, that any such legislation would have to fit within the framework of Directive 92/83/EEC. The Lord Advocate, for the first time, sought in his written case to challenge the first assumption, by arguing that any increase in excise duty could not be restricted to Scotland, under either United Kingdom or EU law. During his oral submissions, he, however, conceded, in the light of article 1(2) of Directive 2008/118/EC and Court of Justice caselaw, that this particular challenge must fail.

[42]Article 1(2) of Directive 2008/118/EC provides:

“Member states may levy other indirect taxes on excise goods for specific purposes, provided that those taxes comply with the Community tax rules applicable for excise duty or value added tax as far as determination of the tax base, calculation of the tax, chargeability and monitoring of the tax are concerned, but not including the provisions on exemptions.”

Article 1(2) was in materially the same terms and has materially the same effect as article 3(2) of the predecessor Directive 92/12/EEC: Tallinna Ettevõtlusamet v Statoil Fuel & Retail Eeesti AS (

Case C-553/13

) EU:C:2015:149. In that case, as in the previous case Transportes Jordi Besora SL v Generalitat de Catalunya (Case C82/12) EU:C:2014:108, the Court of Justice proceeded on the basis that article 3(2) or 1(2) was available for use by a city or region. The Court also considered more generally the preconditions for use of the article. It stated both that, where alternative interpretations of the meaning of the article are possible, “preference must be given to that interpretation which ensures that the provision retains its effectiveness”: Commission of the European Communities v French Republic (

Case C-434/97

), para 21; and that “a derogating provision such as article 1(2) must be interpreted strictly”: Tallinna, para 39. The basic, and cumulative, preconditions are that, first, the tax must be levied for one or more specific purposes and, second, it must comply with the EU tax rules applicable to excise duty and VAT as far as determination of the tax base, calculation of the tax, chargeability and monitoring of the tax are concerned, not including the provisions on exemptions: Tallinna, para 35.The purpose must be a purpose which is not merely budgetary: Commission of the European Communities v French Republic (

Case C-434/97

), para 19, Transportes, para 23 and Tallinna, para 37. It is not therefore sufficient that the tax will be used, or is hypothecated, to promote an activity which the taxing authority is anyway obliged to undertake and to fund: Tallinna, paras 38-40.

[43]What article 1(2) does permit is a tax with the specific purpose “to guide the behaviour of taxpayers in a direction which facilitates the achievement of the stated specific purpose, for example by way of taxing the goods in question heavily in order to discourage their consumption”: Tallinna, para 42. That is precisely the basis on which the petitioners submit that an additional excise tax or VAT could be imposed by the Scottish Parliament and Ministers under article 1(2). The tax would still however have to satisfy the second precondition. What that means, and whether and how far any such tax would have to reflect or respect the categorisation or banding provided by Directive 92/83/EEC, is, as Mr O’Neill accepts, much less clear. Commission of the European Communities v French Republic, on which he relies in this context, stands for the proposition (para 27) that article 3(2) (or now article 1(2)) “does not require member states to comply with all rules applicable for excise duty or VAT purposes as far as determination of the tax base, calculation of the tax, and chargeability and monitoring of the tax are concerned. It is sufficient that the indirect taxes pursuing specific objectives should, on these points, accord with the general scheme of one or other of these taxation techniques as structured by the Community legislation.”

[44]The Court observed that, bearing in mind the different bases on which excise tax and VAT are imposed (see inter alia para 31 above), it would commonly be impossible to comply with the tax rules relating to both simultaneously (para 24) and said that the general aim was to prevent additional indirect taxes “from improperly obstructing trade”: para 26). The tax in issue in the case itself was imposed on beverages with an alcoholic strength exceeding 25% alcohol by volume. The Commission challenged this tax on the basis that the threshold of 25% did not correspond with the threshold of 22% provided in Directive 92/83/EEC (see para 38(v) above). That complaint was summarily rejected by the Court, on the basis that it related to the “substantive scope” of that Directive, and that article 3(2) of Directive 92/12/EEC (or now article 1(2) of Directive 2008/118/EC) “does not, on this point, demand compliance with the tax rules applicable for excise duty or VAT purposes”: para 30. Mr O’Neill relies on this decision in submitting that an excise tax or VAT could, under article 1(2), be levied by reference to bands of alcoholic strength quite different from and much more refined than those specified in Directive 92/83/EEC. Each band of alcoholic strength could, for example, attract a different rate - the greater the strength, the higher the rate.

[45]Since the Lord Advocate did not address any detailed submissions to this point, as discussed in Commission of the European Communities v French Republic, or submit that the second precondition would preclude additional excise duties or VAT rates by reference to narrowly defined bands alcoholic strength, I am prepared for present purposes to accept the correctness of Mr O’Neill’s analysis of the likely effect of the case law. Had the point been critical, it might have been necessary to make a further reference to the Court of Justice, for clarification of the second precondition. But, as will appear, I do not consider it is critical. It follows that, for present purposes, the second point on which Lord Ordinary relied (paras 34 and 38 above) is no longer available to the respondents.

[46]The third point made by the Lord Ordinary (para 68) is that minimum pricing is easier to understand and simpler to enforce. It would not be open to absorption, eg by off-trade outlets such as supermarkets selling alcohol drinks below cost in order to attract other business onto their premises. That remains a valid point, if one considers an excise duty or VAT charge by itself and without more. However, Mr O’Neill counters it by submitting that a combination of measures could achieve the same result as a minimum price. Retailers could be prohibited from making sales below “cost”, with excise duty or VAT being levied at a rate which would be bound, on that basis, to ensure the desired minimum retail sales price. A prohibition on sales at a loss, or giving rise to an artificially low profit margin, applying to all traders within a particular member state, is consistent with European law: Criminal Proceedings against Bernard Keck and Daniel Mithouard (Joined Cases C-267/91 and C-268/91)

[1993] ECR I-6097

and Groupement National des Négociants en Pommes de Terre de Belgique (Belgapom) v ITM Belgium SA and Vocarex SA (

Case C-63/94

)

[1995] ECR I-2467

. The practical difficulties of operating and enforcing any such system are however evident. Alternatively, an excise duty or VAT charge could be levied at a rate which would, by itself, ensure that, even the cheapest, or at least the great majority of the cheapest, drinks were sold at whatever minimum price per unit of alcohol was intended, and retailers could be prohibited from themselves carrying or subsidising all or any part of an excise duty or VAT charge. Both these suggestions are however open to the fundamental objection that they would in practice be bound to lead to a generalised increase in prices across the board, which brings one back to the Lord Ordinary’s first and basic point.

The lack of market impact analysis and proportionality stricto sensu

[47]As I have indicated in para 14, it is not easy to know or to understand the conceptual framework within which to address these topics. It is in particular unclear how the EU market impact of the proposed minimum pricing fits into the exercise which a domestic court must undertake. Assuming (as the Court of Justice’s judgment indicates) that it is to be considered as an aspect of the issue of necessity arising at the second stage identified by both Advocate General Bot and the Court of Justice, it is unclear how it bears on “necessity”. It is clear that the Court of Justice refrained deliberately from endorsing the Advocate General’s analysis of a threestage approach. While that is so, and whether or wherever it fits into the legal analysis, it is nonetheless appropriate to address the basic point, that an appreciation of the likely EU market impact seems on the face of it a sensible precondition to action interfering with EU cross-border trade and competition. Put rhetorically, can it be that, provided an objective is reasonable and can only be achieved in one way, it is irrelevant how much damage results to the ordinary operation of the EU market?

[48]The first response that can be made to this rhetorical question is that the proposed comparison is, in the present case, between two essentially incomparable values. One is the value of health, in terms of mortality and hospitalisation, coupled moreover with the evident desirability of reducing socioeconomic inequalities in their incidence. The other is the market and economic impact on producers, wholesalers and retailers of alcoholic drinks across the European Union. A second observation is that this comparison is yet further complicated by the fact that it is not for any court to second-guess the value which a domestic legislator may decide to put on health. It is “for the member states, within the limits imposed by the Treaty, to decide what degree of protection they wish to assure”: as the Court of Justice reiterated in the present case, para 35, with reference to prior case law. The circularity deriving from the qualification “within the limits imposed by the Treaty” does not help resolve the question what limits there may be on the value that may be placed on life. Would or should a court intervene because it formed the view that the number of deaths or hospitalisations which the member state sought to avoid did not “merit” or was not “proportionate to” the degree of EU market interference which would be involved? I very much doubt it. Any individual life or well-being is invaluable, and I strongly suspect that this is why the Court of Justice did not endorse the Advocate General’s third stage enquiry, and treated the issue very lightly indeed. But it follows that I see very limited scope for the sort of criticism that the petitioners make about the absence of EU market evidence.

[49]As a matter of fact, it appears that the petitioners’ case on this aspect was not prominent before the Lord Ordinary. It was however clearly raised before the First Division (paras 165 and 201-205). The First Division approached this case on two bases. First, it concluded on the material before it (para 203) that “In EU market terms the effect [of minimum pricing at 50 pence per unit] might be described as relatively minor. The ontrade is unlikely to be adversely affected at all. No doubt some wine from Bulgaria, Romania and Portugal may lose a competitive edge. Their share of the market too is very small, but there will be an effect on the competitive nature of some wines and beers from other EU states. Cheap French brandy may be affected, even if, so far as spirits are concerned, the greater impact will be on domestically produced vodka, whisky and cider.” As a broad conclusion on the information available, this does not appear to be challenged. Second, the First Division went on to reject the petitioners’ case that the information on which it was based was inadequate, taking the view that the detailed exercise of market prediction based on the production of models, for which the petitioners were contending, was neither necessary nor practicable.

[50]This leads to a third observation. Whatever the position as to the first two observations, there are also strong reasons for thinking that any attempt to assess the EU market impact in the present area would itself have involved incalculables, which cannot presently be further or more precisely assessed in any way which would be relevant. That conclusion is foreshadowed both in material available at the outset and further material produced to date. At the outset, prior to the 2012 Act, the Scottish Government did not attempt itself or commission any analysis focusing specifically on the EU market. But the BRIA noted under the heading Effect on Market (para 5.114) that: “There is no consensus from industry on what will happen to pricing of products and hence the effect on the market in relation to the introduction of a minimum price per unit of alcohol. Some consider all prices will be affected ie those above a minimum price will also be adjusted, others believe it will only be those below the minimum price that will be affected, and others consider it will be somewhere in between.”

[51]The BRIA continued by recording various possibilities, including switching between categories of alcoholic drink, switching to premium brands once the price differential became small, decimation of own label brands and concentration by retailers on particular products, though which these might be was unknown (paras 5.114-118). Similarly after noting that minimum alcohol pricing would apply to all products, irrespective of which country produces them (para 5.119), the BRIA said that:

“It has proved extremely difficult to access the level of data required to analyse which individual products are likely to be most affected, and the country of origin of such products.”

Again, at para 6.7 the BRIA recorded that the Scottish Government “is not able to predict how individual companies and retailers will react to the introduction of a minimum price per unit.” A survey had shown no consensus. As regards the effect on producers, again, there was no consistent view among (it appears, Scottish) industry representatives.

[52]The BRIA did summarise material indicating that spirits were predominantly, though not exclusively, of domestic origin (paras 5.120-123) and that beer, cider and other alcoholic drinks were both domestically produced and imported. The vast majority of wine was, in contrast, imported from a large number of countries retailing across the range of prices (para 5.124), with the top ten countries of origin of wines selling on the UK market being (in descending order of market share) Australia (21.5% of the market), the USA (14.3%), Italy (14.2%), France (13.9%), South Africa (9.1%), Chile (8.6%), Spain (7.5%), New Zealand (5.3%), Germany (2.3%) and Argentina (1.2%). The BRIA observed that a 50 pence per unit minimum price regime would require an uplift in the average bottle price of wines from each of these countries, except France and New Zealand (the average uplift being 49p for Australia, 60p for the USA, 58p for Italy, 85p for South Africa, 69p for Chile, 60p for Spain, 45p for Germany and 24p for Argentina).

[53]Annex A to the BRIA was a Competition Assessment, which identified markets and sectors potentially affected by minimum pricing, including indirectly affected sectors upstream, in the form of drinks manufacturers and distributors/wholesalers (para 4). Under the heading International Competition, it noted (para 30) that:

“The legislation would apply equally to international producers, wholesalers and retailers trying to enter the Scottish market. Any firms wanting to import high strength, low price products would have to raise their retail prices to comply with the minimum price per unit legislation. This could impact on a foreign company’s ability to compete in the domestic market if the company was currently selling at very low margins in order to be competitive with domestic products.”

The Competition Assessment noted that the initial change effected by minimum pricing would be a reduction in the quantities sold of products whose original price lay below the minimum, though the extent would depend on the elasticity of demand (para 36). Retailers would however benefit by the higher prices of the quantities actually sold and might, as in British Columbian experience, benefit by a general raising of the price of higher value products to maintain a differential with those now affected by minimum pricing (paras 36 and 38). The likely distribution of the increased revenues across the supply chain was not known (para 42).

[54]In August 2013 and in an updated version in December 2014, NHS Scotland produced for the Scottish Government, and the Court of Justice had before it, a table analysing the price distribution of wine from various countries of origin sold in Scotland’s off-trade (where the great bulk of cheap wine is sold). This demonstrates that the majority of the impact of minimum pricing will fall on wine imported from outside the EU, though Italy (with 14.6% of off-trade wine sales), Spain (with 11.5%) and France (with 10.6%) would be affected, selling respectively 31%, 56% and 25% of their wine in Scotland at below 50 pence per unit of alcohol. Germany, Portugal, Bulgaria and Romania had respectively 1.3%, 0.7%, 0.3% and 0.1% of the market, with respectively 2%, 39%, 97% and 84% of their wines being sold at below 50 pence per unit. Another table, which was before the First Division on the reference back from the Court of Justice, showed that none of the 15 wines with the largest off-trade sales values was produced in an EU country. In response to the Court of Justice’s request, the Scottish Government also produced a table stating in general terms which other alcoholic drinks imported into Scotland would be, or be likely to be, affected. Those thought likely to be affected were all brandy and cognac, about 15% of the branded lager sales market in Scotland, part of the stout market, 87% of which was produced in Ireland, but most of which sold at below 50 pence per unit, part of the cider market, 36% of which comes from EU countries and part of the fortified wine market (sherry and port), though most of this sells at more than 50 pence per unit. Some effect on other products was thought possible, but unlikely.

[55]The petitioners have referred to general statements by the Commission about the wine market and the balance of supply and demand and increased competitiveness reached after many years of structural surpluses. They have also referred to statistical information on wine production within the EU and intra- and extra-EU trade. It is not, however, suggested that this material gives answers to the questions which the petitioners submit that the respondents must answer if they are to satisfy the evidential onus on them. The petitioners suggest that it was incumbent on the respondents to analyse “the structure of the wine industries in, say, Romania, Spain, Portugal and Italy, and/or assess how much of the total wine exports of each member state are sold in Scotland, and therefore get some idea of how much MUP [minimum unit pricing] in Scotland might impact upon the wine producers in those countries” (written case, para 4.65). Bearing in mind the impossibility of obtaining information about or analysing even the effect on the Scottish retail market and on the relationship between retailers and their suppliers, this appears an unrealistic counsel of perfection.

[56]This is to my mind confirmed by reports received in October 2012 and May 2016 by the petitioners from Professor George Yarrow and Dr Christopher Decker entitled Economic Analysis of the impact of minimum pricing on alcoholic beverages in Scotland. These set out in broad economic terms various possible outcomes of a minimum pricing regime, and they advance some firm views about the desirability of a taxation, rather than a minimum pricing, approach. But the reports also suggest that the petitioners’ criticisms about lack of specificity are misguided. To my mind, they confirm that lack of specificity is essentially inherent in the present situation. Paragraph 2 of the first report states:

“The detailed analysis is necessarily non-exhaustive, not only because of the time constraints for delivery of this opinion but also because, for reasons to be explained, regulatory policies with the types of characteristics possessed by the MUP scheme are liable to lead to chains of unintended consequences. Whilst it is possible to identify and analyse the tendencies involved in these chains of consequences, they are impossible to pin down with anything approximating total precision, because in part they are governed by future adaptations and innovations to changed incentive structures, knowledge of which is today necessarily limited.”

I note that, even when examining differences between studies by HMRC in 2010 and 2014 of price elasticities in the alcohol market, the authors in their second report identified a problem of uncertainty, arising from lack of sufficient evidence to make it possible to know on what assumptions the available data should be analysed (underdetermination) (para 13).

[57]The authors also stated in their first report (para 6) that, because taxation is, in their view, an obvious and more effective alternative to minimum pricing:

“[T]here is no need in this case to consider balancing trade-offs between health policy goals and other aspects of economic policy, such as the promotion of unimpeded trade flows and the promotion of competition.”

Paragraph 6 of this report means that the authors did not attempt an exercise in comparison of opposing considerations. Those considerations are not only incommensurate on their face; their comparison would, in the light of para 2 of the report, involve weighing inherently unknowable uncertainties regarding the nature and impact of minimum pricing on EU trade against the value which it is for national legislatures and governments to place on health policy goals: see para 48 above.

[58]The Yarrow and Decker reports explain as a matter of general economic theory why and how minimum pricing will be likely to distort the market, by, in effect, suppressing competition or cartelising a part of the market, formerly occupied by lower priced alcoholic drinks, and precluding new entrants into it. This can also be expected to reduce imports. “The economic results to this effect are almost selfevident”, as the first report states (para 51). But the first report also contains material checking the general theory by reference to a First Brand Ltd survey using retail prices in Scotland, Italy and Spain, with a lesser contribution from Portugal and some limited imports from Bulgaria and Cyprus. That distortion of this nature is likely to occur is not however in issue. What is notable throughout the reports is the repeated caveat that the precise nature and effects of minimum pricing on the market cannot at this stage be assessed. It remains uncertain whether it will lead to destocking or, because of the greater retail profit margin, to retailer concentration on the brands whose price has to be increased to the minimum price. As to this paras 63 and 65 of the first report contain the following passages:

“63. … [T]he purpose of this analysis [by reference to the First Brands Ltd survey] was principally to capture a more general point that, whilst the MUP will, by definition, lead to a change in prices for those products which are currently priced below the relevant threshold, it is also possible that products currently priced above the MUP may also be affected by such a policy in the longer term … 65. Indeed, although predicting retailer strategies is a somewhat speculative exercise, we think economic logic points Page 32 to the de-stocking of higher-priced products as a likely outcome.”

[59]Section 4 of the report entitled The Economic Impacts of MUP in more detail starts with two introductory paragraphs, which include the following:

“96. The general conclusion to which economic analysis leads in this case is that it is possible to be very confident that distortionary/discriminatory effects will eventuate, but that it is not possible to evaluate those effects in a comprehensive and precise way.”

[60]The second report examines new evidence available from a Cardinal Research survey of off-trade prices, and concludes that this does not materially affect the general conclusions reached in the first report, regarding the distorting or discriminating effects of minimum pricing on the market and EU trade. While accepting that “the benefits of adherence to Single Market principles (alternatively the costs of setting them aside) are manifestly unquantifiable in any precise way” (para 65) and that it would be for the courts, not economists or other experts, to determine what relative weight should be attached to such principles (paras 53 and 67), the authors repeat their view (see para 6 of their first report, above) that the present case is not one where there is any “trade-off” to resolve (para 68).

[61]Among the factors to which the authors refer is the fact that taxation would increase the Scottish Government’s general revenues, enabling it to devote more funding to promote health, while minimum pricing will increase retailers’ and, it may be others’, profit margins. It is however essentially for the Scottish Government to decide what burden by way of taxation it wishes to impose or, looking at the matter another way, what taxation it requires to raise. It was well aware of the difference in this respect between increased excise or VAT and minimum pricing. Both the BRIA (para 4.3) and the Policy Memorandum (para 29) mentioned it. The BRIA noted that the Scottish Government already had power in other legislation to impose a social responsibility levy on retailers of alcohol on social and health grounds, the proceeds of which would then be available to tackle health issues.

[62]In any assessment which is appropriate of the general proportionality of the proposed system of minimum pricing, due weight must be given to the requirement under the 2012 Act that the system be reviewed after five years, and the “sunset” provision that it will expire after six years unless renewed by a ministerial decision receiving the positive approval of the Scottish Parliament. The proposed system was therefore explicitly provisional, requiring the authorities to take stock of its effectiveness after a period of years and placing the onus of justifying its continuation in the light of experience firmly on the Scottish Parliament at the end of that period. Both the Advocate General (para 85) and the Court (para 57: para 13 above) regarded these provisions as relevant on the issue of proportionality. The Advocate General, at para 85, described the proposed system as “somewhat experimental”. The Court referred, at para 57, to “the possible existence of scientific uncertainty as to the actual and specific effects on the consumption of alcohol of a measure such as the MUP for the purposes of attaining the objectives pursued”. When using the word “scientific”, it cannot have been referring to chemistry or physics. It was clearly referring to the uncertainties experienced even by experts in predicting the precise reactions of markets and consumers to minimum pricing. As the examination above of the available material shows, this applies as much to the effect on EU trade as to any other aspect. The logic of paras 85 and 57 applies as much to the issue presently under discussion as to any other aspect of the proposed system.

Conclusion

[63]The Lord Ordinary and First Division decided that it could reasonably be concluded, on an objective examination of the differing material put before them and now before the Supreme Court, that the proposed system of minimum pricing was proportionate in the sense required by European Union law and now explained by the Court of Justice. It is for the Supreme Court to determine whether this was a judgment that they were entitled to reach. Despite the forceful and very well presented submissions of Mr O’Neill, I consider that they were. A critical issue is, as the Lord Ordinary indicated, whether taxation would achieve the same objectives as minimum pricing. Although not all of the points on which he relied for his conclusion on this issue can still stand, the main point stands, that taxation would impose an unintended and unacceptable burden on sectors of the drinking population, whose drinking habits and health do not represent a significant problem in societal terms in the same way as the drinking habits and health of in particular the deprived, whose use and abuse of cheap alcohol the Scottish Parliament and Government wish to target. In contrast, minimum alcohol pricing will much better target the really problematic drinking to which the Government’s objectives were always directed and the nature of which has become even more clearly identified by the material more recently available, particularly the University of Sheffield’s April 2016 study. As to the general advantages and values of minimum pricing for health in relation to the benefits of free EU trade and competition, the Scottish Parliament and Government have as a matter of general policy decided to put very great weight on combatting alcohol-related mortality and hospitalisation and other forms of alcohol-related harm. That was a judgment which it was for them to make, and their right to make it militates strongly against intrusive review by a domestic court. That minimum pricing will involve a market distortion, including of EU trade and competition, is accepted. However, I find it impossible, even if it is appropriate to undertake the exercise at all in this context, to conclude that this can or should be regarded as outweighing the health benefits which are intended by minimum pricing. In the overall context of the Scottish or, on the face of it, any other market, it appears that it will be minor, though it will hit some producers and exporters to the Scottish market more than others. Beyond that, the position is essentially unpredictable. Submissions that the Scottish Government should have gone further to predict the unpredictable are not realistic. The system will be experimental, but that is a factor catered for by its provisions for review and “sunset” clause. It is a significant factor in favour of upholding the proposed minimum pricing régime.

[64]For these reasons, I consider that the appeal should be dismissed.

Scotch Whisky Association and others v The Lord Advocate and another [2017] UKSC 76

Introduction

The Court of Justice’s judgment

The issues in more detail

The issues

The objective(s) pursued by minimum pricing

(s) pursued by minimum pricing

Less restrictive measures to achieve the same aim?

The First Division also said (at para 181) that:

The lack of market impact analysis and proportionality stricto sensu

Conclusion

Cited in 2 later judgments