KELLOGG MARKETING AND SALES COMPANY (UK) LIMITED & Anor (R on the application of) v SECRETARY OF STATE FOR HEALTH AND SOCIAL CARE

The Food (Promotions and Placement) (England) Regulations 2021 (SI 2021/1368 – “the 2021 Regulations”) are part of the Government’s strategy to tackle childhood obesity. They introduce restrictions on the promotion, in supermarkets or other large outlets and online, of food which is classified as high in fat, sugar or salt (“HFSS”) and therefore “less healthy”.

Under Schedule 1 to the 2021 Regulations, breakfast cereals are included in the categories of food which may be “specified food” and therefore subject to the relevant restrictions. Whether a given product within one of these categories is in fact classified as “less healthy” depends on the score which it is given under the Food Standards Agency’s Nutrient Profiling Model (“NPM”), which was devised in 2004/2005 and is required by the 2021 Regulations to be applied in accordance with its associated technical guidance (the “Technical Guidance”). Both form part of a document entitled “Nutrient Profiling Technical Guidance” (“the NPTG”) which was published by the Department of Health in January 2011. The NPM requires that the nutrient content of a given product is analysed per 100g of the food or drink itself, rather than taking account of what the food or drink may be consumed with.

The Claimants (“Kellogg”) are part of a well-known group of companies which manufactures breakfast cereals. Under the 2021 Regulations, the majority of Kellogg’s breakfast cereal products, and therefore 54.7% of the breakfast cereal currently sold by Kellogg, will be classified as less healthy assuming that it does not take steps to reduce the levels of sugar, fat and/or salt in those products. Kellogg makes use of agreements with retailers to place its products in parts of stores (e.g. near the checkout, in a queuing area, at the end of an aisle) which maximise sales (“location promotions”), and to promote its products on the retailers’ websites. 30% of its HFSS products are sold through location promotions. It also sells 3.5% of its breakfast cereal on price promotions (e.g. “buy one get one free” or “25% off if you buy 6”). Kellogg says that the 2021 Regulations will therefore impact on sales of its products. Its marketable goodwill will also be damaged through a combination of reduced brand visibility, and damage to its reputation more generally as a result of the majority of its products being classified as HFSS.

Kellogg’s fundamental complaint about the 2021 Regulations is that, under the NPM, the fact that a portion of, for example, Kellogg’s “Frosties” will typically be consumed with milk is not taken into account in assessing whether this product is HFSS. If the consumption of milk with breakfast cereal were taken into account, fewer Kellogg products would be classified as HFSS because the nutrient values of the added milk would contribute to the scoring. Kellogg argue that an approach which measures the relative levels of fat, sugar or salt in the product itself, rather than the health impact of the product as typically consumed, is disproportionate and irrational.

Kellogg point out that in the case of products which require to be mixed or cooked with liquid before they are consumed (e.g. custard powder, cocoa powder, dried soup and dried pasta) the Technical Guidance specifies that the calculation of the nutrient profile is to be based on 100g “of the product as reconstituted according to the manufacturer’s instructions”. They argue that breakfast cereals are a similar product, and a similar approach should therefore be taken.

The 2021 Regulations are therefore challenged on four grounds. The first two of these are vires objections to two aspects of the Regulations, and the third and fourth directly raise Kellogg’s commercial concerns.

Ground 1 contends that regulation 10 of the 2021 Regulations, which provides for food authorities to be able to issue improvement notices where there are reasonable grounds for believing that a person is failing to comply with the Regulations, is ultra vires. It is contended that the Defendant does not have the power, under the relevant enabling legislation – the Food Safety Act 1990 (“the FSA”) and the Regulatory Enforcement and Sanctions Act 2008 (“the 2008 Act”) - to make regulations which provide for improvement notices in relation to the subject matter of the 2021 Regulations.

Ground 2 contends that it is impermissible for regulation 3(4) of the 2021 Regulations to incorporate the NPTG by reference as the basis for scoring a given product, so as to determine whether the product is “less healthy”. Kellogg argues that section 16(1) of the FSA, which permits the Defendant “by regulations [to] make…provision for…regulating” the relevant matters requires that all applicable rules or criteria for determining whether a given specified food should be classified as HFSS must be contained in the statutory instrument itself. The Defendant’s approach in this case is therefore ultra vires, and it is unlawful in that it frustrates sections 16 and 48(2) of the FSA and sections 1(1) and 5 of the Statutory Instruments Act 1946 (“the 1946 Act”).

Ground 3A alleges irrationality of reasoning on the part of the Defendant. The decision to include breakfast cereals on the list of potential “specified food” is in Schedule 1 to the 2021 Regulations is not challenged, but it is said that this increased the need for a rational consideration of how “less healthy” breakfast cereals were to be identified. Kellogg’s pleaded case in relation to this issue is that the Defendant failed to have regard to a relevant consideration, namely that there had been no assessment of the appropriateness of assessing the nutrient profile of breakfast cereals without taking into account the fact that they are typically consumed with milk. The Defendant was not asked to, and did not, consider whether breakfast cereals should be assessed with or without milk and was not aware of “the over-inclusive nature of the Regulations as they applied to breakfast cereals (by contrast…. to other dried foods)”. Nor did he take account of any prior consideration of the application of the NPM to breakfast cereals when the model was under development: see the Amended Statement of Facts and Grounds (“ASFG”) at [85].

It is also contended by Kellogg that the Defendant failed to discharge his Tameside duty of reasonable inquiry by reviewing the work which was done between 2004 and 2009 in relation to the NPM and the Technical Guidance and, in particular, the extent to which there was expert consideration of whether it was appropriate to assess breakfast cereals on an “as consumed” rather than an “as sold” basis. Had the Defendant done so, he would have been discovered that this was an issue which had not been resolved and he would have been bound to take steps to consider it further: ASFG at [85A].

Ground 3B alleges: (a) that the assessment of the nutrient profile of breakfast cereals without including added milk disproportionately infringes Kellogg’s right to peaceful enjoyment of its possessions contrary to Article 1 of Protocol 1 of the European Convention on Human Rights (“A1P1 ECHR”) and/or its right to freedom of expression, contrary to Article 10 ECHR. It is said that the Defendant cannot show that the assessment of breakfast cereal without milk is rationally connected to its aim of reducing the consumption of HFSS products and childhood obesity, and nor can he show that the approach under the NPTG is the least intrusive means of achieving its aims given that a similar approach could be taken to that which is taken to products which require to be “reconstituted” before they are eaten. It is also alleged: (b) that there is irrationality of outcome: “for the same reasons”, the approach to the assessment of breakfast cereals falls outside the range of approaches reasonably open to the Defendant: ASFG at [100].

The 2021 Regulations were laid before Parliament in draft on 21 July 2021. They were approved by the House of Commons on 16 November 2021 and by the House of Lords on 24 November 2021, and they were made on 2 December 2021. They are due to come into effect, in respect of location promotions, on 1 October 2022. Mr Hickman confirmed that the date of the decision under challenge for the purposes of the Claim is 21 July 2021.

Part II of the FSA comprises sections 7 to 26 and it contains the “Main Provisions” of the Act. Section 10 is part of a section on “food safety”. It provides, so far as material:

“10.— Improvement notices.

(1)

If an authorised officer of an enforcement authority has reasonable grounds for believing that the proprietor of a food business is failing to comply with any regulations to which this section applies, he may, by a notice served on that proprietor (in this Act referred to as an “improvement notice” )—

(a)

state the officer's grounds for believing that the proprietor is failing to comply with the regulations;

(b)

specify the matters which constitute the proprietor's failure so to comply;

(c)

specify the measures which, in the officer's opinion, the proprietor must take in order to secure compliance; and

(d)

require the proprietor to take those measures, or measures which are at least equivalent to them, within such period (not being less than 14 days) as may be specified in the notice.

(2)

Any person who fails to comply with an improvement notice shall be guilty of an offence.

(3)

This section and section 11 below apply to any regulations under this Part which make provision—

(a)

for requiring, prohibiting or regulating the use of any process or treatment in the preparation of food; or

(b)

for securing the observance of hygienic conditions and practices in connection with the carrying out of commercial operations with respect to food or food sources.” (emphasis added)

Under sections 37 and 38 of the FSA there is a right of appeal to the Magistrates Court, and thereafter to the Crown Court, against a decision to issue an improvement notice. Such appeals are not limited to appeals on points of law.

Section 11 empowers a court to impose a prohibition order on the proprietor of a food business where they are convicted of an offence under any regulations to which sections 10 and 11 apply, provided the business or its practices give rise to a risk of injury to health as defined.

Sections 12 and 13 of the FSA go on to deal with emergency prohibition notices and orders, and emergency control orders, respectively, before sections 14 and 15 deal with selling food which is not of the nature, substance or quality demanded and falsely describing or presenting food. The latter two sections appear under the heading “Consumer protection”.

The FSA contains a number of provisions which enable further provision to be made by regulations or orders: see e.g. section 6(4) which concerns enforcement, section 10 (above), sections 16-19 and 26. Section 16, which provides so far as material as follows, is in issue in this case:

“16.— Food safety and consumer protection.

It will be noted that sections 16(1)(c) and (d) replicate sections 10(3)(a) and (b). Parliament therefore intended that the power to issue improvement notices and prohibition orders under section 10 would apply to breaches of regulations made pursuant to sections 16(1)(c) and (d). The relevant enabling provisions of section 16 for the purposes of the 2021 Regulations are sections 16(1)(e) and (f) and it is said that this sheds light on an issue between the parties as to whether, for the purposes of section 48(1) of the FSA, set out below, regulation 10 of the 2021 Regulations “deals with matters similar to those being dealt with by” section 10 of the FSA.

Section 26(3) of the FSA provides, so far as material:

“26.— Regulations and orders: supplementary provisions.

…(3) Regulations under this Part may—

(a)

provide that an offence under the regulations shall be triable in such way as may be there specified; and

(b)

include provisions under which a person guilty of such an offence shall be liable to such penalties (not exceeding those which may be imposed in respect of offences under this Act) as may be specified in the regulations.” (emphasis added)

Section 48(1) forms part of Part IV to the FSA: “Miscellaneous and Supplemental”. It provides, so far as material, as follows:

“48.— Regulations and orders.

(1)

Any power of the Secretary of State to make regulations or an order under this Act includes power—

(a)

to apply, with modifications and adaptations, any other enactment (including one contained in this Act) which deals with matters similar to those being dealt with by the regulations or order;

(b)

to make different provision in relation to different cases or classes of case (including different provision for different areas or different classes of business); and

(c)

to provide for such exceptions, limitations and conditions, and to make such supplementary, incidental, consequential or transitional provisions, as the Secretary of State considers necessary or expedient.

(2)

Any power of the Ministers or the Minister to make regulations or orders under this Act shall be exercisable by statutory instrument.

(3)

Any statutory instrument containing—

(a)

regulations under this Act; or

(b)

an order under this Act other than an order under section 60(3) below,

shall be subject to annulment in pursuance of a resolution of either House of Parliament.” (emphasis added)

It is because part of the 2021 Regulations was enacted pursuant to section 62 of the Regulatory Enforcement and Sanctions Act 2008 that the affirmative resolution procedure was required to enact them: see section 62(3) of the 2008 Act. Section 62(3)(a) provides that the statutory “instrument may not be made unless a draft has been laid before, and approved by resolution of, each House of Parliament…”. Section 5 of the Statutory Instrument Act 1946, to which I return below, deals with the annulment or negative resolution procedure.

As noted above, the 2021 Regulations apply restrictions to the promotion of “specified food”. Regulation 3 provides so far as material as follows:

“Specified food

3.—(1) For the purposes of these Regulations, “specified food” is food contained in a prepacked food item which—

(a)

is Schedule 1 food,

(b)

is less healthy (as defined in paragraph (4)), and

(c)

is not food to which paragraph (5) (charity food sales) applies.

…

(4)

For the purposes of this regulation—

(a)

food that is not a drink is less healthy if it scores 4 or more points in accordance with the Nutrient Profiling Technical Guidance;

(b)

a drink is less healthy if it scores 1 or more points in accordance with the Nutrient Profiling Technical Guidance….”

Under regulation 2, “the Nutrient Profiling Technical Guidance” referred to in regulation 3(4) is defined as:

“..the guidelines published by the Secretary of State on 14 January 2011 about the application of the 2004-2005 Nutrient Profiling Model”.

Schedule 1 to the 2021 Regulations then sets out 13 categories of food which may be “less healthy” depending on their nutrient profiling score. These include certain types of prepared soft drink containing added sugar, certain types of savoury snack, confectionery, ice cream and similar frozen products, cakes, biscuits, buns, desserts, sweetened yoghurt, pizzas, roast and fried potato products as well as ready-made foods and meals which are marketed as merely needing to be cooked or reheated, and certain sorts of breaded or battered foods. Category 3 on Schedule 1 is:

“Breakfast cereals including ready-to-eat cereals, granola, muesli, porridge oats and other oat-based cereals”.

The 2021 Regulations apply to “qualifying businesses”, which are defined in regulation 4. In broad terms, they are businesses which offer prepacked food items for sale, whether in person or online, and which employ 50 or more employees, other than care homes and educational institutions and, for most purposes, restaurants.

As for the restrictions, in broad terms:

As to enforcement, regulation 9 then designates each food authority as the enforcement body within its area. regulation 10 provides as follows:

“Improvement notice

10.

If a food authority has reasonable grounds for believing that a person is failing to comply with one or more of regulation 5, 6, 7 or 8, it may, by a notice served on that person (in these Regulations referred to as an “improvement notice”)—

(a)

state the authority’s grounds for believing that the person is failing to comply with, as the case may be, regulation 5, 6, 7 or 8;

(b)

specify the matters which constitute the person’s failure so to comply;

(c)

specify the measures which, in the authority’s opinion, the person must take in order to secure compliance; and

(d)

require the person to take those measures, or measures that are at least equivalent to them, within such period as may be specified in the notice.”

Regulation 11 then creates an offence of failure to comply with an improvement notice served under regulation 10, and regulation 12(1) provides that a person who is guilty of an offence under regulation 11 is liable on summary conviction to a fine. Regulation 12(2) and Schedule 2 to the 2021 Regulations make provision for fixed monetary penalties for such an offence pursuant to the powers under section 62 of the 2008 Act.

The NPTG is addressed to food manufacturers, retailers and advertisers. As the introduction to the NPTG states, the NPM was developed by the Food Standards Agency (“the Agency”) in 2004-2005 to provide Ofcom with:

“a tool to differentiate foods on the basis of their nutritional composition, in the context of television advertising foods to children”. (emphasis added)

Responsibility for nutrition was transferred from the Agency to the Department of Health in 2010 and the document was therefore “rebranded” a Department of Health document in 2011. The introduction goes on to explain that:

“The model uses a simple scoring system where points are allocated on the basis of the nutrient content of 100g of a food or drink. Points are awarded for ‘A’ nutrients (energy, saturated fat, total sugar and sodium), and for ‘C’ nutrients (fruit, vegetables and nut content, fibre and protein). The score for ‘C’ nutrients is then subtracted from the score for ‘A’ nutrients to give the final nutrient profile score. (emphasis added)

Foods scoring 4 or more points, and drinks scoring 1 or more points, are classified as ‘less healthy’ and are subject to Ofcom’s controls on the advertising of foods to children on TV.

The model applies equally to all food and drink; there are no exemptions or category-specific criteria.” (emphasis in the original)

The NPM itself is then set out in less than 2 pages in Section 1. It comprises:

Importantly, the sample for the purposes of both A points and C points is clearly stated to be “100g of the food or drink”, rather than 100g of the food or drink as typically consumed or with anything else. Just as the introduction says, the NPM is a tool to differentiate foods on the basis of their nutritional composition so as to decide which foods have a less healthy nutrient profile. It therefore measures the composition of a given food rather than a meal or snack of which it may form part, or a dish of which it may be an ingredient or to which it is added. In the case of breakfast cereals, the food is the breakfast cereal. The NPM therefore measures the nutritional values of the breakfast cereal itself rather than the breakfast, just as Schedule 1 to the 2021 Regulations applies to savoury snacks “whether intended to be consumed alone or as part of a complete meal”.

The second section of the NPTG is entitled “2. How to Calculate Scores for the Fruit, Vegetable and Nut Content of Food and Drink”. This runs to 2.5 pages. The Section begins with a link to another document where, it is said, “more detailed information can be found”. That document, dated December 2005, purports to provide “definitions”, “exceptions”, “recommendations” and “guidance” as to how to apply the NPM. The information in Section 2 overlaps with the information in the December 2005 document, but the latter contains further detail. Section 2 of the NPTG also describes itself as “guidance” but, for example, it deals with the definition of fruit and vegetables, specifying that: “The definition for the 5 A DAY programme should be used…Potatoes and other starchy vegetables such as yams do not count.”. The December 2005 document then provides detail as to which foods are to be regarded as fruit or vegetables for this purpose. Section 2 specifies that only intact fruit and vegetables which have been minimally processed (e.g. peeled or sliced) can be included when calculating a score. There is then a definition of “Nuts”, with specific rules as to how to score coconut and how seeds should be dealt with, and there are sections dealing with scoring dried and pureed fruit and vegetables and fruit juice, and whether the score must be calculated before or after the product has been cooked. The text of these provisions is prescriptive.

Section 3 of the NPTG is entitled “Frequently Asked Questions”. It sets out 11 questions which include one which specifically addresses the case of breakfast cereals:

It appears from the evidence that this question may have been included because there had been arguments from breakfast cereal manufacturers, in the course of the decision making in relation to the NPM and the Technical Guidance in 2004-2009, that a “with milk” approach should be adopted or, alternatively, the unit for measurement should be smaller, for example 50g. I will return to the discussions on these issues which took place in greater detail below.

The FAQs also include answers which reiterate that the unit for assessment is 100g of the food or drink even where the product is consumed in quantities of less than 100g (Question (b)) and regardless of whether the product is measured by volume rather than weight (Question (d)). Again, these questions and answers reflect issues which arose in the course of the development and review of the NPM. There is then this Question and Answer:

“e.

Should I calculate nutrient profile scores for products as sold or consumed?

Nutrient profile scores should usually be calculated for a product as sold. In cases where a product needs to be reconstituted before it is eaten, for example custard powder, the nutrient profile score should be based on 100g of the product as reconstituted according to the manufacturers instructions.” (emphasis added)

Question (f) then applies this principle to drinks to which liquid needs to be added before they are consumed (e.g. squash, milkshake powder or syrup, hot chocolate powder, cocoa powder and malted milk powder). Question (g) deals with dried and ready-made soups and makes clear that dried soups should be approached in the same way as other foods which need to be reconstituted but ready-made soups should be assessed in the ready-made form. Question (h) deals with dried pasta, noodles or dried rice products and, in effect, repeats answer (e). These questions therefore, in effect, answer questions as to what is the food or drink for the purposes of the 100g measurement.

Mr Hickman makes the point, which I accept, that although the term “reconstituted” is used in relation to these foods and drinks, it is not accurate to say in relation to all of them that they are reconstituted, in the sense that they are restored to their original hydrated state, before being consumed. I discuss these products further below.

There is then a question (i) which deals with calculating the nutrient score for dairy milk given that it may be whole milk, semi skimmed or skimmed and given that there may be geographical and seasonal variations in the nutritional components of milk, and there is a variety of processing treatments. The assessor is required to apply McCance and Widdowson’s “The Composition of Foods 2002”, Sixth Summary Edition or, where this does not give a value, the whole milk average is to be used, adjusted for fat content.

Drinking yoghurts are then addressed (Question (j)), and then the question of what to do if the product is reformulated after submission of the nutrient profile is addressed in Question (k).

Section 4 of the NPTG is a series of 6 worked examples.

Mr Hickman’s argument starts with section 10 of the FSA which, he points out, does not confer a power to enact regulations which expand the scope for issuing improvement notices in this context, and which expressly states that the power to impose improvement notices applies to alleged breaches of regulations to which the section applies (section 10(1)). Those regulations are specified in section 10(3) and they are regulations under Part II which make provision for “(a) requiring, prohibiting or regulating the use of any process or treatment in the preparation of food; or (b) securing the observance of hygienic conditions and practices in connection with the carrying out of commercial operations with respect to food or food sources”. He submits that regulation 10 of the 2021 Regulations is concerned with very different matters.

Mr Hickman notes that the Defendant relies, in his pleaded case, on section 48(1)(a) of the FSA. He says that this enacts a Henry VIII power to amend primary legislation but, he submits, relying on the decision of the Court of Appeal in McKiernon v Secretary of State for Social Security (1990) Admin LR 133, that this power has not been validly exercised in the present case. He submits that, for there to be a valid exercise of a Henry VIII power, the fact that a given regulation is applying or modifying primary legislation must be expressly stated.

In answer to a question from the court, Mr Hickman submitted that in this type of case there are two questions:

Mr Hickman draws a contrast between regulation 10 and regulation 16 of the 2021 Regulations which, he says, demonstrates the legislative technique which must be deployed where Henry VIII powers are exercised. Regulation 16 provides, so far as material:

“Application of various sections of the Food Safety Act 1990

16.—(1) The provisions of the Food Safety Act 1990 set out in paragraph (2) apply for the purposes of these Regulations—

(a)

as if any reference in those provisions to the Act, or to any Part of the Act, were a reference to these Regulations;

(b)

with any modifications specified in paragraph (2)…”

Paragraph (2) then sets out a list of 15 sections of the FSA which apply for the purposes of the 2021 Regulations. In a small number of cases there are modifications to these sections which are intended to ensure that the language of the section reflects the fact that it is applies in the context of the 2021 Regulations. Where this is the case, in the usual way the section is said to apply “as if” and then the modification is stated. For example:

“(g)

section 33(2), as if the reference to “any such requirement as is mentioned in subsection (1)(b) above” were a reference to any such requirement as is mentioned in that subsection as applied by sub-paragraph (f)”.

Section 10 of the FSA is not expressly applied or modified by regulation 16, although regulation 16(2)(m) does provide that section 37 of the FSA (appeals to the Magistrates Court from an improvement notice issued under section 10) will apply “as if the reference to an improvement notice were to an improvement notice served under regulation 10”.

Mr Hickman submits that even if, contrary to his argument, section 48(1) permits what he describes as the application, modification or adaptation of primary legislation “by implication”, regulation 10 does not do this. Rather, it sets out an entirely separate provision to section 10 of the FSA.

Second, he submits that the power to issue improvement notices is expressly limited by section 10 to alleged breaches of regulations making the provisions described under section 10(3)(a) and (b). Relying again on McKiernon, he submits that these express statutory limitations cannot be “overridden” by a general power to make regulations which apply section 10 with modifications or adaptations. The approach to construction has to be strict (in this regard he also relies on R v Secretary of State for Social Security ex parte Britnell [1991] 1 WLR 198) and the terms “modification” and “adaptation” are indicative of slight or consequential changes rather than transformation or disapplication of the limits stated in section 10(3), which is what he says regulation 10 does.

Mr Hickman also relies on what he says are the extensive powers conferred on local authorities by regulation 10, emphasising that an improvement notice may be issued where there is merely a reasonable belief that there has been a breach of the relevant regulations, and that there are criminal sanctions for failure to comply with the notice. There does not have to be an actual breach of the 2021 Regulations for an offence to be committed. It is also for the food authority to decide whether to prosecute. He submits that Parliament cannot have intended that such powers would be conferred by statutory instrument. He adds that any doubt has to be determined against the Defendant given the strict approach to construction and he relies in this regard on McKiernon, Britnell and the PLP case.

Third, Mr Hickman emphasises that the provision which may be applied pursuant to section 48(1) has to deal “with matters similar to those being dealt with by the regulations or order”. Section 10 is concerned with food safety whereas the 2021 Regulations are concerned with consumer protection. The focus has to be on a comparison between the precise matters dealt with. The regulation of food preparation, treatment and/or hygiene is not similar to the regulation of the presentation, promotion and marketing of foods, even if this is done for reasons of public health. The aim of the measures is not to the point. If, as the Defendant contends, the subject matters referred to in section 16 are all to be regarded as similar to each other then the limitation in section 48(1) is denuded of meaning.

For his part, Sir James Eadie QC emphasised the breadth of the enforcement powers in relation to the FSA which are provided for by Parliament. In addition to the extensive powers to enact substantive provisions by subordinate legislation, the FSA provides, in section 6(4), for regulations or orders to specify which of the relevant “authorities are to enforce and execute them, either generally or in relation to cases of a particular description or particular area”. Under Schedule 7 to the 2008 Act, it is provided that the FSA is one of the enactments in respect of which there may be a civil penalty regime alongside any criminal offences which are created by statutory instrument: see section 62(2), in particular. This power has been exercised in the form of Schedule 2 to the 2021 Regulations, as I have noted. Sir James emphasises the power, under section 26(3) of the FSA, to enact regulations which provide for the trial of offences under the regulations and for criminal penalties where a person is found guilty. This, he submits, is a reflection of the fact that regulations made pursuant to the FSA may create criminal offences for breach.

Sir James’s point was that it would be surprising if, in this context, Parliament had nevertheless set its face against the use of improvement notices as part of the enforcement regime unless such notices were specifically authorised by section 10 of the FSA. He characterised section 10 as Parliament positively deciding that improvement notices could be used to enforce regulations which made provision for the matters specified in section 10(3) - they were “automatically” allowed, without further legislative steps being required - but leaving for further consideration the question of their use to enforce other regulations which made provision for other aspects of the FSA. This was not a question of section 10 being “overridden” in the sense that regulation 10 went against a decision of Parliament as expressed in primary legislation.

Secondly, adopting a potential line of argument raised by the court and departing somewhat from his pleaded case and his skeleton argument, Sir James argued that the words of section 48(1) of the FSA make clear that they are merely spelling out what the various powers to make regulations or orders which are set out earlier in the FSA “include”. As he put it, if section 48 did not exist, it would not be necessary to invent it because the powers under, for example, section 16 would be sufficient. Section 48 was a supplemental provision which emphasised the breadth of earlier provisions.

Thirdly, he argued that section 48(1) permitted the enactment of regulation 10 in any event. In this connection he contested Mr Hickman’s proposed interpretation of McKiernon as establishing a rule that the effective exercise of Henry VIII powers requires an express statement on the face of the relevant regulation that it is modifying the primary legislation in question. His submission, by reference to the PLP case in particular, was that the question in every case of this type is one of the construction of the relevant power in the primary legislation to determine whether the purported exercise of that power fell within the class of action that Parliament must have contemplated when delegating the power to the Executive. He accepted that a strict approach is to be adopted but, he submitted with reference to R (Black) v Secretary of State for Justice [2017] UKSC 81; [2018] AC 215 and R (O) v Secretary of State for the Home Department [2022] UKSC 3; [2022] 2 WLR 343, there had been some softening of the so-called strict approach and that legislative purpose plays an important part in the task of interpretation.

Fourthly, Sir James contested each of Mr Hickman’s arguments on the text of section 48(1) and submitted that the language of the provision clearly authorised the enactment of regulation 10 of the 2021 Regulations.

In McKiernon there was a fundamental inconsistency between section 165A(2) Social Security Act 1975 and regulation 25 Social Security (Industrial Injuries) (Prescribed Diseases) Regulations 1985. Section 165A applied to all social security benefits and provided that, save in specified cases, no entitlement arose unless there was a claim made in the prescribed manner and within the prescribed time. Section 165A(2) required that:

“(2)

Regulations shall provide for extending, subject to any prescribed conditions, the time within which a claim may be made in cases where it is not made within the prescribed time but good cause is shown for the delay.”

Regulation 14(2) Social Security (Claims and Payments) Regulations 1979 had made such provision by allowing for an extension of time where “good cause” for a failure to claim within the statutory deadline was proved by the claimant. However, regulation 25 of the 1985 Regulations subsequently purported to disapply regulation 14 of the 1979 Regulations in cases of occupational deafness and replaced it with, in effect, a hard deadline. The net position was that, if regulation 25 was to have the effect contended for by the Secretary of State, it had to be held that it overrode, at least in occupational deafness cases, the express requirement under section 165A(2) that there be provision for an extension of time in relation to all claims for social security benefits. The Secretary of State’s case was that the relevant power to do this was to be found in section 77(2) of the 1975 Act which provided, so far as material:

“In relation to prescribed diseases ... Regulations may provide –

(a)

for modifying provisions of this Act relating to injury benefit and disablement benefit, and the administration of such benefit." (emphasis added)

As to Mr Hickman’s primary argument, that there must be an express indication that the primary legislation is being modified, it is true that the Secretary of State’s argument in McKiernon was rejected by Russell LJ on the basis of an alternative submission by Mr Drabble for the claimant that:

“..where, as in this case, a statute enables the Secretary of State by Regulations to modify any provision contained in primary legislation, the appropriate modification should be expressly stated in the statutory instrument and· is not to be inferred or implied from the content of the regulation. Subsection (2) of s.165A is a mandatory provision requiring the Secretary of State to make Regulations incorporating provisions for extending the time within which a claim for benefit may be made outside the prescribed time where good cause for delay can be shown: That subsection cannot be overriden by any statutory instrument unless the instrument itself, and in specific terms, modifies subs. (2).” [137F-H]

Russell LJ went on to say that a “narrow and strict” construction was required when Parliament delegates a power to the Executive to amend primary legislation. He emphasised that the power in section 77(2) of the 1975 Act was to “modify” the primary legislation and that regulation 25 made no reference to section 165A, still less did it contain any words which purported to modify that provision. It was silent in relation to the very provision in the primary legislation which it was said to have overridden:

“In my judgment, the reality of the situation is that reg. 25, far from modifying s. 165A(2), ignores it and, accordingly, I take the view that s.165A(2) prevails, together with reg. 14 of the 1979 Regulations, which introduce the concept of good cause for delay.” [138C]

Nourse LJ agreed that “a power to modify the provisions of an Act of Parliament by statutory instrument must be restrictively construed”. On this basis he accepted Mr Drabble’s first submission (which Russell LJ did not consider it was necessary to decide) that section 77(2) did not confer a power to modify generally applicable provisions of the 1975 Act such as section 165A. The power under section 77(2) was limited to a power to modify provisions which specifically related to disablement benefits and the administration of this particular type of benefit. [139D-E]. Nourse LJ went on to accept Mr Drabble’s alternative submission, albeit obiter, but he characterised this argument as being that “reg 25 did not ‘modify’ s. 165(2). At best….there is an inconsistency between the two provisions…that cannot be a modification of one provision by another”. Nourse LJ went on to say that “In accepting this submission also, I gratefully adopt the fuller reasons which have been stated by Lord Justice Russell”. [139F-G]

Lord Donaldson MR said that he agreed that the appeal should be allowed for the reasons given by Russell LJ. He said:

“Primary legislation represents the expression of the will of Parliament after full debate with considerable opportunities for amendment. Subordinate legislation, at any rate when subject to the negative resolution procedure, represents the will of the Executive exercised within limits fixed by primary legislation. Whether subject to the negative or affirmative resolution procedure, it is subject to much briefer, if any, examination by Parliament and cannot be amended.

The duty of the courts being to give effect to the will of Parliament, it is, in my judgment, legitimate to take account of the fact that a delegation to the Executive of power to modify primary legislation must be an exceptional course and that, if there is any doubt about the scope of the power conferred upon the Executive or upon whether it has been exercised, it should be resolved by a restrictive approach.” [140B-D] (emphasis added)

Like Nourse LJ, Lord Donaldson agreed with Mr Drabble’s first submission, that section 77(2) should be construed as only conferring a power to modify those provisions of the 1975 Act which specifically related to disablement benefit or the administration of this benefit, rather than to modify generally applicable provisions. But his acceptance of this submission was “in principle” and he went on to say that whether or not the submission was correct it also depended on the Secretary of State being able to show an exercise of his powers under section 77(2):

“It must be remembered that any such modification would apply not only to this statutory instrument, but to all future statutory instruments until it was further modified.

Bearing in mind the exceptional nature of this power, I think that reg. 25 …must be held to be ultra vires unless it can be shown that either previously or simultaneously there was some amendment of the mandatory requirements of s.165(2) that Regulations shall contain a good cause provision or of s. 165A(3)(c) deleting the reference to disablement benefit. That he cannot do. All that he can say is that it is implicit in the instrument since otherwise it would be ultra vires.” [141A-C] (emphasis added)

In my view a careful reading of McKiernon itself, and subsequent authority, shows that Mr Hickman’s arguments seek to derive more from the decision of the Court of Appeal than is warranted by what the Court said and decided. The principle on which the Court based its conclusion was that the issue is always one of construction — there the construction of section 77(2) of the 1975 Act; here the rather different words of sections 16 and/or 48(1) of the FSA – in order to determine the scope of the enabling provision. Because of the difference in the degree of Parliamentary scrutiny, where there is a doubt the court should err on the side of a restrictive approach to the enabling provision because to do so errs on the side of greater Parliamentary scrutiny and the supremacy of Parliament. There may also be issues of construction in relation to the amending provisions in the subordinate legislation in order to determine whether they are within the scope of the enabling provision, but the two issues are bound up with each other, rather than separate. In McKiernon, the majority of the Court of Appeal was of the view that the enabling power in question was to modify specific rather than generally applicable provisions of the 1975 Act but that, even if it was to modify all provisions of that Act, what had been done went beyond modification and amounted to overriding section 165A(2).

It is true that Russell LJ interpreted the word “modifying” in section 77(2) of the 1975 Act as envisaging words which referred to section 165A(2) and said how it was being modified. I also agree that typically the provision which purports to exercise Henry VIII powers will include such words because the subordinate legislation is amending primary legislation and therefore refers to the provisions which are to be amended. But I do not accept that Russell LJ or either of the other two members of the Court of Appeal were purporting to establish an invariable rule that there must be express modification in any case where subordinate legislation is enacted pursuant to a power to modify primary legislation. The Court’s observations were about the particular provisions in the case which they were considering in their proper context. The view of the Court was also very much influenced by the improbability of Parliament conferring on the Executive a power to override mandatory requirements in the primary legislation, and their impression appears to have been that the section 165A(2) had been overlooked rather than modified: i.e. that there had been a mistake.

I note that Leading Counsel were agreed that an objective approach to the words of the relevant provisions was and is required. What matters is whether there were the vires to enact the subordinate legislation rather than whether the Minister in question had the relevant provisions in mind at the relevant time. Russell LJ’s reference to regulation 25 “ignoring” section 165A(2) therefore should not be read as requiring the Court to be satisfied that the drafter had section 165A(2) in mind when drafting regulation 25 if they were to be held to have modified this provision. The logic of the objective approach is, however, that if the power exists the court need not necessarily search for evidence, in the form of express words, that the drafter was intending to modify the primary legislation.

Ex parte Britnell is an example of where the subordinate legislation, regulation 20(2) of the Social Security (Payments on account, Overpayments and Recovery) Regulations 1987, purported to extend the scope of a provision in the primary legislation - section 53 of the Social Security Act 1986 - rather than override it. This was held by the House of Lords to be authorised by section 89(1) of the 1986 Act, which provided for regulations to make transitional provision “(including provision modifying any enactment contained in this or any other Act)” (emphasis added). Whereas section 53 contemplated the recovery of overpayments of benefit occurring after its coming into force section 89(1), properly construed, also permitted recovery of past overpayments of benefit as a transitional measure.

Britnell therefore supports the view that the courts may more readily find that a provision of subordinate legislation which enlarges the scope of a provision in primary legislation is a modification. A provision which cuts down the scope of primary legislation and therefore goes directly against the will of Parliament as specifically expressed at the time of its enactment, will require clearer evidence of the power to do so in the language of the enabling provision. The PLP case may be another illustration of this point although various considerations led to the result in that case.

It is also relevant to note that Britnell does not detract from the analysis of McKiernon which I have set out above. Lord Keith of Kinkel, with whom the rest of their Lordships agreed, considered McKiernon and said at 204E-F:

“The principal ground for the decision was that it did not purport to modify section 165A(2) but simply ignored it. It would in any event have been strange if a power to modify had been construed as authorising the annulment of a mandatory provision. The judgments contain passages to the effect that a power to modify the provisions of a statute should be narrowly and strictly construed, and that view is indeed a correct one.”

In R v Secretary of State for the Environment, Transport and the Regions ex part Spath Holme Ltd [2001] 2 AC 349 the House of Lords considered the scope of a power under section 31(2) of the Landlord and Tenant Act 1985 to enact “provisions excluding, adapting or modifying any provision made by or under any enactment (whenever passed) relating to rent or the recovery of overpaid rent”. Their Lordships held that this permitted an Order which placed maximum limits on fair rent increases which could be registered in relation to regulated tenancies under section 70 of the Rent Act 1977. Lord Bingham cited the passage from the judgment of Lord Donaldson in McKiernon which I have highlighted at [62] above, where he said that if there was any doubt about the scope of the power or whether it has been exercised, it should be resolved by a restrictive approach. Lord Bingham said that this principle had been endorsed in Britnell and added, at 382H:

“Recognition of Parliament's primary law-making role in my view requires such an approach. But it is an approach which is only appropriate where there is a genuine doubt about the effect of the statutory provision in question. Here, the language used seems on its face to leave little room for doubt about the scope of the power in section 31(2).”

In the PLP case, the issue was as to the scope of the power of the Lord Chancellor under section 9(2) of the Legal Aid, Sentencing and Punishment of Offenders Act 2012, by order, to “vary or omit” the legal services described in Part 1 of Schedule 1 to the 2012 Act. The Supreme Court held that, on its true construction, this power did not authorise an amendment to Schedule 1 to provide that people who failed a residence test would, subject to certain exceptions, be removed from the scope of Part 1.

At [23] Lord Neuberger PSC with whom the other members of the Court agreed, set out the role and task of the court in all cases where the vires of subordinate legislation are challenged:

“Subordinate legislation will be held by a court to be invalid if it has an effect, or is made for a purpose, which is ultra vires, that is, outside the scope of the statutory power pursuant to which it was purportedly made. In declaring subordinate legislation to be invalid in such a case, the court is upholding the supremacy of Parliament over the Executive. That is because the court is preventing a member of the Executive from making an order which is outside the scope of the power which Parliament has given him or her by means of the statute concerned. Accordingly, when, as in this case, it is contended that actual or intended subordinate legislation is ultra vires, it is necessary for a court to determine the scope of the statutorily conferred power to make that legislation.”

As to the particular type of subordinate legislation under consideration in the PLP case, at [25] he noted that:

“As explained in Craies on Legislation, 10th ed (2012), ... para 1.3.9: “The term “Henry VIII power” is commonly used to describe a delegated power under which subordinate legislation is enabled to amend primary legislation”.

He went to say, at [26]:

“The interpretation of the statutory provision conferring a power to make secondary legislation is, of course, to be effected in accordance with normal principles of statutory construction. However, in the case of an “amendment that is permitted under a Henry VIII power”, to quote again from Craies, para 1.3.11:

‘as with all delegated powers the only rule for construction is to test each proposed exercise by reference to whether or not it is within the class of action that Parliament must have contemplated when delegating. Although Henry VIII powers are often cast in very wide terms, the more general the words used by Parliament to delegate a power, the more likely it is that an exercise within the literal meaning of the words will nevertheless be outside the legislature’s contemplation’.” (emphasis added)

Lord Neuberger then noted that in Britnell and Spath Holme the House of Lords cited with approval the passage from the judgment of Lord Donaldson in McKiernon which I have highlighted at [62] above and which Lord Neuberger said was “to much the same effect” as the passage from Craies which he had approved. He then summarised Lord Bingham’s observations in Spath Holme, to which I have referred above, as including that:

“where there is “little room for doubt about the scope of the power” in the statute concerned, it is not for the courts to cut down that scope by some artificial reading of the power.”

Finally, in R (Coughlan) v Minister for the Cabinet Office [2022] UKSC 11; [2022] 1 WLR 2389 the approach to the interpretation of Henry VIII clauses was summarised by Lord Stephens JSC, in the light of the authorities discussed above. At [17] he said:

“..the rule for construction is to test each proposed exercise by reference to whether or not it is within the class of action that Parliament must have contemplated when delegating. However, in relation to the power to amend primary legislation “if there is any doubt about the scope of the power conferred upon the Executive or upon whether it has been exercised, it should be resolved by a restrictive approach.”

Sir James also referred me to Black and to the O case, as I have noted. His submission was that the restrictive approach referred to in the authorities on Henry VIII clauses is equivalent to the requirement to demonstrate a given construction by express words or necessary implication, and that this test had been modified in Black where the issue was the rule that statutes do not bind the Crown save by express words or necessary implication. He relied on [36] of the judgment of Baroness Hale in that case, where she said:

“(3)

The goal of all statutory interpretation is to discover the intention of the legislation.

(4)

That intention is to be gathered from the words used by Parliament, considered in the light of their context and their purpose. In this context, it is clear that Lord Hobhouse of Woodborough’s dictum in R (Morgan Grenfell & Co Ltd) v Special Comr of Income Tax [2003] 1 AC 563, 616, para 45, that “A necessary implication is one which necessarily follows from the express provisions of the statute construed in their context” must be modified to include the purpose, as well as the context, of the legislation.”

Sir James also referred me to various passages in O, including [39]-[43] of the judgment of Lord Hodge JSC, in particular. In O the issue was whether the fee chargeable for registering children as British citizens under regulations made pursuant to section 68 of the Immigration Act 2014 impermissibly restricted their right to citizenship under section 1(4) of the British Nationality Act 1981. The regulations were said to be contrary to the presumption, recognised in R v Secretary of State for the Home Department ex parte JCWI [1997] 1 WLR 275, that subordinate legislation enacted pursuant to a later Act of Parliament will not cut down rights established pursuant to an earlier statute. Sir James submitted that this was conceptually very similar to the issue in the present case.

In O the Supreme Court held that the issue was one interpretation of the enabling power in the later statute. The relevant presumption should be taken into account in the interpretive exercise, but it was not to be elevated into a rule which predetermined the vires of the later statute. Lord Hodge noted that the Court was not dealing with a vested right at common law or under statute but, rather, with a statutory procedure for registration by which a person can acquire citizenship [43], and he concluded at [42]:

“If the court, having taken into consideration the established assumptions or presumptions concludes that statute 2, expressly or by necessary implication, has empowered the executive to make subordinate legislation which has the effect of removing rights conferred by statute 1, the principle enunciated by the Court of Appeal in JCWI… imposes no additional hurdle for the Secretary of State.”

I agree with Sir James that a rather more flexible approach is required than the rule based approach advocated by Mr Hickman. Whether the issue relates to Henry VIII clauses or other types of enabling provision, the question is always as to the scope of the power conferred on the Executive by the primary legislation under consideration. That involves interpretation of the words of the enabling provision in its proper context and having regard to the purpose of the legislation. In the particular case of Henry VIII clauses, because what is contended for is a power to amend primary legislation the court is required to have particular regard to the fact that Parliament is the primary law maker and has come to a view on the particular matter, which view is expressed in the words of the statute. For this reason, if the court is in doubt as to the scope of the enabling provision it should err on the side of caution and accord supremacy to Parliament.

I am not convinced that this case is concerned with the exercise of a Henry VIII power in the sense discussed in the authorities referred to above. In those cases the question was whether there was a power to amend the primary legislation in question so that it remained the source of the relevant right, power, duty or obligation but its scope and operation were altered. Regulation 10 does not purport to amend section 10 of the FSA. It enacts a free standing mechanism for the enforcement of the 2021 Regulations, albeit that mechanism imitates the mechanism in section 10 of the FSA.

This does not mean that the task in relation to any challenge to the vires of regulation 10 is anything other than to consider whether, on their true construction, the relevant provisions of the FSA confer a power on the Defendant to enact regulation 10. Nor does it mean that the principle of Parliamentary scrutiny and the supremacy of Parliament as law-maker are irrelevant. They remain important considerations given the lower degree of scrutiny which is applied to subordinate legislation. I also accept the point underlying Mr Hickman’s argument that regulation 10 “overrides” section 10, namely that the extent to which the subordinate legislation is inconsistent with the will of Parliament, as expressed in the relevant substantive provisions of the primary legislation, is relevant to the construction of the enabling provisions: the greater the departure from or inconsistency with the substantive provision, the greater the need for clear words which authorise this in the enabling provision.

But it does mean, in my judgment, that the key question in the present case may not be whether it was permissible to enact regulation 10 pursuant to a power, under section 48(1) of the FSA “to apply” section 10 “with modifications and adaptations” when, on one view, this is not what the Defendant has done. Mr Hickman is right that regulation 16 is a clear exercise of the section 48(1) power in that the sources of the relevant rights, duties, powers and obligations remain the sections of the FSA listed in regulation 16(2). They are “applied” for the purposes of the 2021 Regulations in this sense. Regulation 16, in effect, amends the relevant provisions of the primary legislation so that they have a wider application than previously. Regulation 10 does not apply section 10 in this sense, and the authorities on Henry VIII powers should be considered with this in mind.

I therefore prefer the view that the power to enact regulation 10 is to be found in sections 16(1)(e) and (f) and/or 26 of the FSA. It seems to me that having regard to the aims of the FSA, the breadth of the powers highlighted by Sir James in his submissions summarised above, and the extensive role which Parliament left to regulations and orders to enact the practical detail of the FSA, Parliament authorised the Defendant to devise appropriate methods of enforcement as part and parcel of making “provision for…regulating” the relevant activities, subject to the negative resolution procedure under section 48(3). Mr Hickman accepted that the Regulations did so, albeit he said that the vires for any methods of enforcement probably came from section 26 rather than section 16. Given that, as Mr Hickman also accepted, the permitted methods of enforcement included the creation by regulations of criminal offences and fines (section 26(3)), I accept that the class of action which Parliament must have contemplated included measures, such as improvement notices, which gave the recipient an opportunity to take corrective steps so as to avoid criminal proceedings but which created a criminal offence of breach of the notice.

Mr Hickman argued that such a conclusion would be contrary to the terms of sections 10 and 11 of the FSA, which provide for improvement notices and prohibition orders to be issued only in relation to the food safety measures described in section 10(3)(a) and (b). Indeed, on this argument, since section 10 is the only express source of the power to issue improvement notices, and is not itself an enabling provision, Parliament decided that there could never be provision for improvement notices in regulations made under the FSA, not even in regulations dealing with the subject matter referred to in section 10(3)(a) and (b). I see the force of the argument that the fact that Parliament made specific provision for improvement notices in section 10 may indicate that it did not intend to confer a power to make wider provision for improvement notices pursuant to section 16 (or, indeed, 48(1)). But one can just as easily ask the question why, given the measures which Parliament “ruled in”, it would have ruled out the use of improvement notices in relation to any subject matter other than that to which section 10 applies. What is it about the regulations made pursuant to regulation 16(1)(a), (b), (e) and (f) which suggests that Parliament decided that it could only approve the use of improvement notices by primary, rather than secondary, legislation? Ultimately, no answer to this question was forthcoming from Mr Hickman.

I agree with Sir James that it is not a case of section 10 being “overridden” in the sense in which section 165A(2) of the Social Security Act 1975 was alleged to have been overridden by the subordinate legislation in the case of McKiernon i.e. the subordinate legislation being in direct contradiction of an express requirement of the primary legislation. The scope and operation of sections 10 and 11 are not altered in any way by regulations 10-12. Rather, this is in my judgement a case of the question of further applications of improvement notices being left open and the subsequent regulations then enlarging the power to issue improvement notices in a way which is consistent with the aims of the FSA.

In this connection I note that, subject to his point on similarity under section 48(1), Mr Hickman accepted that if the Defendant had included section 10 on the list of sections of the FSA in regulation 16(2) which were to be applied with modifications and adaptations for the purposes of the 2021 Regulations, this would have been permissible. There is a tension between this concession and his argument that section 10 shows that Parliament set its face against the use of improvement notices other than to enforce the regulations referred to in section 10(3). The concession seems to me to involve an acceptance that Parliament did contemplate the enactment, in regulations rather than primary legislation, of further provision for improvement notices provided the relevant regulations dealt “with matters similar to those being dealt with by” section 10.

The previous point also tends to suggest that Mr Hickman’s argument is one of form. He accepted this when it was put to him but he argued that form is important in this context – because of the importance of Parliamentary scrutiny, transparency and legal certainty it needs to be apparent on the face of the statutory instrument that a provision of primary legislation is being applied and/or modified – but he did not cite any authority other than McKiernon in support of this argument, and I do not see any principled reason why this would be so in the present case. Parliamentary and public scrutiny were not impeded by regulation 10 being enacted in the terms in which it was, rather than listed in regulation 16(2) with appropriate adaptations, and nor was legal certainty undermined. The objective approach to deciding the vires of subordinate legislation is also inimical to arguments along the lines that enacting a given power is permitted provided the enactment uses one form of words rather than another. As Leading Counsel agreed, what matters is whether the enactment of the power is or is not authorised by the primary legislation.

On this analysis, the Defendant does not need to bring regulation 10 within the terms of section 48(1), which spells out powers that exist elsewhere in the FSA, rather than placing limits on those powers. However, I accept Sir James’ submission that the case can be brought within section 48(1) in any event. Much of the reasoning as to the nature and effect of regulation 10 and section 10 set out above continues to apply, but specifically by reference to the terms of section 48(1):

I therefore reject Ground 1.

In summary, Mr Hickman’s argument under this heading was as follows:

Sir James Eadie argued, on the basis of Alvi, that reference to an extraneous document in a statutory instrument is permissible “provided that the reference was to an existing document and there was no question of any sub-delegation” ([40] Alvi) and/or “the document was fixed and not changeable at the Secretary of State’s discretion” ([23]). The NPTG satisfied these tests and there therefore could be no objection to its incorporation by reference. He then contested each of the steps in Mr Hickman’s argument summarised above.

Ultimately, there was little controversy about the relevant principles. Mr Hickman agreed the following summary when I put it to him:

I would add that I agree that the incorporated document must be in existence at the time when the statutory instrument is laid before Parliament, and the effect of it becoming law is that it cannot then be changed without following whatever legislative process is required to amend or replace the statutory instrument itself. These requirements are a necessary consequence of the need for legal certainty and the application of the constitutional principles referred to above, although I accept Mr Hickman’s point that they are not sufficient in themselves: the primary legislation must also permit incorporation of the document by reference. I leave for consideration in another case whether a statutory instrument could provide for amendment by a Minister of the incorporated document. That question does not arise here as Sir James accepted that any proposed changes to the NPTG in the future would need to be laid before Parliament.

Looking more closely at the ex parte Camden case, there paragraphs 1(2) and 2(1) of Schedule 1 to the Supplementary Benefits Act 1976 provided for a person’s resources and their requirements for the purposes of supplementary benefit to be “calculated in the prescribed manner” by reference to such “categories …as may be prescribed”. The word “prescribed” was defined by section 34(1) of the Act as meaning “specified in or determined in accordance with regulations”. Section 33(1) specified that the power to make regulations was to be exercised “by statutory instrument” and there was a requirement, under section 33(3), that a draft of the regulations be “laid before Parliament and approved by resolution of each House.”. The 14 page statutory instrument in question made reference to a publication which had not been laid before or approved by Parliament, namely a directory known as the “Supplementary Benefit Maximum Amounts, Initial Periods and Board and Lodging Areas” which, as the title suggests, specified the financial limits of benefits and periods of time for a large number of boarding and lodging areas.

Having considered the Statutory Instruments Act 1946, the Court of Appeal held that the approach which had been taken was permissible. The primary argument of Mr Drabble for the claimant was that the directory formed part of the regulations. Section 33(3) of the 1976 Act therefore required that it be laid before Parliament. Since it had not been, the statutory instrument was invalid. Slade LJ’s answer to this submission is encapsulated in the following passage at 827B-C:

“In view of the provisions of section 33(1) of the Act of 1976, read in conjunction with section 1(1) of the Act of 1946 , I am prepared, for the purposes of this appeal, to accept that the wording of section 33(3) of the Act of 1976 obliges the Secretary of State to place before Parliament a draft of the entire document by which he proposes to exercise his power to make regulations. Nevertheless, in my judgment, that was what he did. The document by which he exercised his power to make regulations was the 14-page document to which I have referred earlier in this judgment. Before he exercised the power, he duly placed before Parliament a draft of that 14-page document and the draft was duly approved by a resolution of each House.”

For these particular statutory purposes, the directory was not to be regarded as part of the statutory instrument. What mattered was that the document “by which the relevant power [was] exercised” was laid before Parliament. The 14 page statutory instrument was that document and the directory was not.

I note that Slade LJ did not consider that there was any inconsistency between this analysis and the constitutional principles to which I have referred. He looked at the matter very much as one of practical reality, noting at 824D-F that:

“the directory was clearly identified in the draft regulations laid before each House. It is not suggested that the members of either House could not have obtained access to a copy of it if they so wished or were misled in any way by the form of the draft or did not fully appreciate the substance of what they were approving.”

At 827F-G he said:

“…no circumvention of Parliamentary control was either attempted or achieved……If Parliament had objected to the manner in which the Secretary of State had chosen to identify these maximum awards (by reference to a document not contained in the statutory instrument) or, after scrutiny, had objected to the proposed maximum awards themselves, the remedy lay in its own hands; it could have withheld approval from the draft. It must be taken not to have objected to the draft on these or any other grounds.”

He also associated himself with the following observations of the first instance judge (Macpherson J) [827H-828E]:

In Alvi the issue was whether it was permissible for paragraph 82 of Appendix A to the Immigration Rules to provide that no points would be awarded under a new points-based system for determining permission for non-EEA nationals to work in the United Kingdom if the job was not on the Border Agency’s list of skilled occupations, and did not pay the appropriate rate for the job listed. The list was not laid before Parliament.

Section 3(2) of the Immigration Act 1971 provided, so far as material, that the Secretary of State was obliged to “lay before Parliament statements of the rules, or any changes in the rules, laid down by him as to the practice to be followed in the administration of this Act” and that these would then be subject to the negative resolution procedure. The issue was therefore whether the materials referred to in paragraph 82 of Appendix A did or did not contain “rules” for the purposes of section 3(2), and the Supreme Court held that they did insofar as they contained criteria which were determinative of the rights of an applicant for leave to enter or remain.

It was in this context that the Supreme Court Justices sought to define what a “rule”, as opposed to guidance or further detail, was: see paragraphs [57], [94], [97], [115], [122] and [128]. But, as the following passage from the judgment of Lord Dyson illustrates, this was because this was the determinative issue under the applicable provisions in that case. The Court’s definition was also specifically for the purposes of section 3(2) of the 1971 Act, and was based on what the terms of that section indicated as to the extent of the scrutiny which Parliament required.

“94….a rule is any requirement which a migrant must satisfy as a condition of being given leave to enter or leave to remain, as well as any provision “as to the period for which leave is to be given and the conditions to be attached in different circumstances” (there can be no doubt about the latter since it is expressly provided for in section 3(2)). I would exclude from the definition any procedural requirements which do not have to be satisfied as a condition of the grant of leave to enter or remain. But it seems to me that any requirement which, if not satisfied by the migrant, will lead to an application for leave to enter or remain being refused is a rule within the meaning of section 3(2). That is what Parliament was interested in when it enacted section 3(2). It wanted to have a say in the rules which set out the basis on which these applications were to be determined.” (emphasis added)

I therefore reject Mr Hickman’s argument that any analogy can be drawn between Alvi and the present case on the basis that the NPTG can be regarded as containing rules. No doubt it can be, as Sir James accepted, but that is not the determinative issue in the present case, which turns on the wording of section 16 of the FSA rather than the terms of section 3(2) of the Immigration Act 1971.

The second point which emerges from Alvi is that, at [40], Lord Hope referred with approval to the ex parte Camden case and, in particular, Slade LJ’s approval of the observations of Macpherson J. It is on this basis that Sir James relies on Alvi for the proposition that what matters is that the reference in the statutory instrument is to an existing document and that there is no question of sub-delegation. In Alvi, Mr Drabble (now QC), on behalf of the Joint Council for the Welfare of Immigrants, had also made submissions which accepted the reasoning in ex parte Camden but had emphasised that the document referred to could not be open to change at the discretion of the Secretary of State without further reference to Parliament [23].

Finally, in relation to Alvi, it is true, as Mr Hickman points out, that in referring with approval to ex parte Camden, Lord Hope included Slade LJ’s approval of Macpherson J’s observation that there was no objection to the practice of referring to extrinsic documents in the eyes of the Joint Committee on Statutory Instruments, and that there had been an increasing tendency to resort to this technique. Mr Hickman’s argument was that views have since changed and that this is now regarded as bad practice. In this connection, he referred to passages from reports of committees of the House of Lords in Session 2021-22 including:

I agree with Sir James that this debate does not take the matter further in the present case.

Returning, then, to the central issue, as noted above, section 16(1) of the FSA permits the Defendant “by regulations [to] make…(e) provision for imposing requirements or prohibitions as to, or otherwise regulating, the…presenting or advertising of food” and “(f) such other provision…for prohibiting or regulating the carrying out of commercial operations with respect to food”. I do not accept that the effect of the formulation “by regulations make…provision for”, which is used repeatedly in the FSA, is that the whole of the provision has to be set out or contained in one document, nor that all of the determinative criteria or rules have to be. Plainly, Parliament contemplated that provision for regulating a given activity in relation to food could be made by regulations which referred to or incorporated external benchmarks, formulae for assessment or other standards. The detail of how scientific analysis would be carried out need not be spelt out in the statutory instrument, albeit that detail was important and/or determinative. Indeed, it would be surprising if the position were otherwise given that the subject matter of the legislation is issues of food safety and public health, which are likely to engage scientific processes.

This leaves the question whether, nevertheless, all of the documents which regulated the relevant activities were required to be laid before Parliament by virtue of section 62(3) of the 2008 Act or section 5 of the 1946 Act. Here, again, it seems to me that the answer is that which was given by Slade LJ in ex parte Camden, namely that for the purposes of the Statutory Instruments Act 1946 the document by which the power to make regulations was exercised – SI 2021/1368 — was duly laid before Parliament and approved by a resolution of each House of Parliament.

Mr Hickman raised the spectre of one-line statutory instruments, incorporating multiple documents, if the Defendant’s argument is accepted. He argued that this should lead the court to conclude that section 16 of the FSA required the whole of the rules to be set out in the statutory instrument itself as, otherwise, there would be no logical basis for objecting to a one line statutory instrument. But in my view the question is one of degree, where it is for the court to determine whether the extent or nature of the incorporation by reference goes beyond what was authorised by the statute, or means that the statutory instrument is not the document by which the power to make regulations is being exercised, or amounts to an unlawful attempt to circumvent or frustrate the required level of Parliamentary scrutiny. Here, the statutory instrument which was laid before Parliament set out the applicable law in detail and cross referred, in effect, to an established industry benchmark as the basis for the assessment of a particular aspect of the regulatory framework. For the reasons I have given, I regard it as realistic to describe the statutory instrument itself as exercising the relevant power and I do not accept that the Defendant’s approach in this case was contrary to constitutional principle or frustrated the aims of section 16 of the FSA and/or the 1946 Act.

I therefore reject Ground 2.

Kellogg’s pleaded case is summarised at [9] and [10] above. Mr Hickman’s case in his skeleton argument was consistent with the pleading. He complained that “the over-inclusive nature of the restrictions as they apply to breakfast cereals, and the significant fact that they are assessed without regard to their actual health impacts, was not drawn to the Defendant’s attention…[he] was not made aware of the longstanding issue and longstanding concerns of the industry when he took the decision..”. There had been no detailed or meaningful assessment of the appropriateness of assessing breakfast cereals without milk and/or this issue had not been resolved at the time of the creation of the NPM and the publication of the Technical Guidance: “the issue was never subject to detailed assessment and was never resolved”. In any event, the Defendant was not aware of this issue at all. The skeleton argument then engaged in an examination of the evidence, including the contemporaneous documents, with a view to establishing these assertions as facts.

Sir James Eadie’s essential answer to this complaint, as developed in his skeleton argument, was that the issue was, in fact, considered and resolved in the course of the development of the NPM and the Technical Guidance in the period 2004-2009. The matter had been settled, and the NPTG had been applied by Ofcom since then. Moreover, there had been two consultations in relation to what became the 2021 Regulations, in the course of which the issue had not been raised by Kellogg or any other manufacturer of breakfast cereals. This, therefore, clearly was not a matter which was so relevant that it had to be taken into account by the Defendant and/or that he had to make further inquiries so as to appraise himself of the position.

Mr Hickman’s oral submissions covered the ground foreshadowed in his pleaded case, although he did not contest the issues in relation to the development of the NPM and the Technical Guidance in the 2004-2009 period in any detail, perhaps because he recognised that these were not his strongest points. However, he also gave a great deal of emphasis to the questions of the extent to which the Defendant had been aware of the Technical Guidance and its contents at the time when the statutory instrument was laid before Parliament, and whether he had been aware of Kellogg raising issues about the approach under the NPM on and after 11 June 2021. There was also a degree of uncertainty as to whether the Defendant had personally considered the arguments which Kellogg put forward in its Pre Action Protocol (“PAP”) letter of 3 September 2021.

Sir James dealt with these issues “on his feet”, as it were, but on the day after the hearing before me the Defendant submitted a second witness statement for Mr Dodds together with a “Note on the outcome of the NPM Review”, both dated 29 April 2022. There was then a “Claimants’ Note in Reply” dated 5 May 2022, and a letter dated 10 May 2022 from the Government Legal Department which corrected two points in the Claimants’ Note in Reply. In the absence of any objection, I have taken these materials into account in coming to my conclusion on Grounds 3A and 3B. However, I have done so for the purposes of deciding the case as pleaded rather than any wider case which Mr Hickman may or may not have been pursuing.

This question is dealt with in the witness statement of Dr Alison Tedstone MBE dated 3 March 2022. At the time of giving her statement, she was the Chief Nutritionist at the Office for Health Improvement and Disparities (previously Public Health England) in the Department of Health and Social Care. She has worked as a government nutritionist since 2001 and has a distinguished record as an academic and of public service in the field. She worked for the Food Standards Agency at the time of the development and review of the NPM and she took part in the decision-making process in her then capacity of Principal Nutritionist, and then as Head of Nutritional Science. She gives her recollection of the process and refers to a number of contemporaneous documents which evidence that process.

Kellogg put forward their arguments as to what the contemporaneous evidence shows, principally through the witness statements of Dr Alexa Hoyland — Wellbeing Market Activation Director of Kellogg Management Services Europe Limited — dated 8 November 2021, and Mr Neil McGowan - Senior Director of Global Regulatory Affairs and Wellbeing at Kellogg Europe Trading Limited — dated 30 March 2022. Obviously, they also rely on the written and oral arguments of Mr Hickman.

In the light of the dispute between the parties as to what the evidence shows, I undertake my own assessment below, albeit recognising that the passage of time means that the evidence is unlikely to be complete and bearing in mind that the context is a claim for judicial review. However, it does appear that a reasonably clear account of what happened can be derived from the contemporaneous documents which remain, combined with the evidence of Dr Tedstone. My findings are also relevant to the issues of proportionality and rationality of outcome under Ground 3B.

By way of overview, the NPM was originally developed in 2004-2005 and there was then a review of its operation in 2007-2009. As far as the initial development phase is concerned, the Food Standards Agency commissioned a team of researchers from the British Heart Foundation Health Promotion Research Group based at Oxford University (“the Oxford University team”). Their work was overseen by an Expert Group whose members were drawn from the Department of Health and other organisations whose focus was health and nutrition, as well as the British Retail Consortium, the Food and Drink Federation (“FDF”) and the Consumers Association. The proposed model was also subjected to external scrutiny through a scientific workshop attended by academics with expertise in nutrition, and by the Scientific Advisory Committee on Nutrition (“the SACN”). There were also two public consultations. The first was from November 2004 to February 2005 and the second, which was in relation to the final model, took place between July and September 2005.

The important starting point is to note that the criteria by which the success of the model was to be judged, as noted by the Expert Group at a meeting in July 2004, were simplicity (bearing in mind that it would be used by manufacturers, retailers, regulators and others), transparency, accuracy (taking into account sensitivity and specificity), workability in tandem with the Department of Health’s guidelines on “5 a day” and the capacity to extend its use to adults. The aim was to develop a pass/fail test which could be carried out on the basis of the nutritional information and ingredients provided on the packaging for a given product.

The Oxford University team began its work in May 2004. It is clear from its highly detailed report, dated October 2004, that a large number of models were considered as the basis for categorising a given food as healthy or less healthy. The approach in other countries was also examined as part of a literature review. The recommendation, in the light of the extensive research which was carried out, was to develop and refine a scoring based model which measured a group of nutrients known as Group C nutrients per 100g of the food. As will be appreciated, this was fundamentally the approach adopted under the NPM in that it measured the inherent nutritional quality of the food rather than the food as consumed and/or in combination with other foods with which it was typically consumed.

It is clear from a 7 March 2005 summary of the more than 80 responses to the first consultation in relation to the preferred model that there was a high degree of participation, including by the Association for Cereal Food Manufacturers (“ACFM”), the FDF, Breakfast Cereals UK (“BCUK”) and Kellogg itself. The arguments that the approach should be based on the actual health impacts of the food, so that the basis for measurement should be portion size and/or “as consumed”, were clearly advanced. Specifically in relation to breakfast cereal it was argued that “The method based on per 100g is not a true reflection of the meal occasion which is based on a serving of milk”, and that the approach based on 100g of the product itself penalised low moisture foods. There was also a complaint that the model operated unfairly given that foods such as rice and pasta were modelled “as eaten” whilst breakfast cereals were modelled “dry.” It was argued that they should be modelled “as eaten, with milk” and it was pointed out that if they were modelled with semi skimmed milk they were often classed as healthier.

It is also clear that the arguments that breakfast cereals should be assessed with milk were considered carefully as part of the development and decision making process. For example, a report of a scientific workshop on 25 February 2005, which was convened as part of the consultation process, and was chaired by a member of the SACN and attended by 30 experts in nutrition and public health, shows that it was concluded that, amongst other tests which should be carried out, “It might be useful to run the model for foods as eaten e.g. breakfast cereal with milk”. I note that the argument that the approach should be based on portion size rather than a 100g base was also considered but this was regarded as problematic given that it would be a huge task to agree a typical portion size, and given that portion size would tend to vary with the age and appetite of the child. It was therefore rejected. The conclusion of the academic workshop was that the portion size approach added an unnecessary complication and should only be used if demonstrable problems arose from the 100g approach.

The evidence also shows that the model was then run for breakfast cereals with milk and this exercise was overseen by the Expert Group. A food composition table, dated 7 April 2005, shows that 15 wholegrain and high fibre breakfast cereals, 8 hot breakfast cereals and 11 other breakfast cereals were tested in relation to various serving sizes of the cereal with various quantities of skimmed milk, and in relation to a large number of different nutrients.

At a subsequent meeting of the Expert Group, including representatives from the SACN, on 11 April 2005, a series of issues which had been raised in the course of the consultation were discussed, including various arguments that the model discriminated against cereal based products and particularly against breakfast cereals. One of the arguments was that the assessment of breakfast cereals should include semi skimmed milk, but the meeting considered that the results which the model was producing in relation to breakfast cereals were appropriate and assigned low priority to this issue.

Thereafter, the Oxford University team and the Expert Group worked together to refine the model and, in July 2005, the Agency launched a second public consultation in relation to the now revised model. More than 70 responses were received from a range of stakeholders. It appears from a table of responses, dated September 2005, that consultees were asked whether refinements which had been made had improved the classification of individual foods. The arguments for an “as consumed” approach, taking into account consumption of milk with the breakfast cereal, and the complaints about discrimination against breakfast cereals put forward in the course of the first consultation, which I have summarised above, were then reiterated by Kellogg and other breakfast cereal manufacturers. These arguments were answered by the Agency as follows:

“The nutritional content of foods is affected by how it is prepared and eaten and foods can be eaten in a variety of different ways – there is no way of knowing how the consumer will eat the food as purchased. For example, there will be people who eat breakfast cereals with different amounts and types of milk, such as whole milk, skimmed milk, soya milk or other liquids such as orange juice. Assessing foods on an “as purchased” basis is the only way of allowing foods to be compared on a “like for like” basis” (emphasis added)

The Food Standards Agency Board then decided to recommend the NPM to Ofcom at a Board Meeting on 13 October 2005, and to carry out a review of the impact of the model after a year. In coming to this decision, the Board also accepted the SACN’s advice. It is clear from the minutes of this meeting, and related documents, that detailed consideration had been given to the issues in relation to the NPM in the year since the Oxford University team made its recommendation, and also by the members of the Board as part of the process of deciding whether to recommend it. Indeed, the “100g approach” was specifically discussed at the 13 October meeting, including arguments which Kellogg had put forward in a letter which was circulated to the members of the Board, and the recommendation to adopt this approach was probed. The minutes record that Professor Rosemary Hignett, the then Head of the Nutrition Division of the Agency:

“…explained that this issue had been considered on a number of occasions. It had been concluded that the portion size approach had no advantages but some disadvantages. It moved products that were high in fat, sugar or salt but eaten in small portions, such as snacks, towards the healthier end of the spectrum. This was considered to be a disadvantage by the expert group and the academic seminar. There was also added complexity as there were a number of foods that were used in different ways. Also, individuals acted in different ways and behaviour between different age groups of children differed. The use of a 100g method as a basis for advice and legislation on claims was well accepted and used in the UK and internationally. Professor Jackson [Chair of the SACN] added that portion size as a unit could not be measured or quantified with any reliability and also took no account of frequency of consumption.” (emphasis added)

Having considered the arguments for an “as consumed” approach to the assessment of foods generally, including an approach in the case of breakfast cereals which took account of the fact that they are generally consumed with milk, the decision on this fundamental issue was to assess a standard unit of the food itself rather than attempt to predict and measure what it might be consumed with. This was consistent with the recommendation of the Oxford University team in October 2004 but the model which they proposed had been refined in other ways in the course of the deliberations in the course of the year which followed.

In February 2007, Ofcom announced its intention to use the NPM to restrict the broadcast advertising of HFSS foods to children. The restrictions were then progressively introduced between April 2007 and 31 December 2008.

The review of how the NPM was working in practice was carried out by an independent review panel chaired by Professor Mike Kelly, who was the Public Health Excellence Centre Director of the National Institute for Health and Care Excellence (“NICE”). The Review Panel included members from the British Nutrition Foundation and various other experts in the field. It began its work in September 2007, and the aim was to make final recommendations to the Agency Board in early 2009. The Review included an academic workshop on nutrient profiling, which was held on 28 February 2008, consideration by the SACN and a 12 week public consultation which began in July 2008.

As part of the Review, the panel met with a large number of stakeholders, including Jenny Walton representing Kellogg/the ACFM, on 15 October 2007. The minutes of this meeting record that there were complaints, amongst other things, about how the model operated in relation to breakfast cereals in that the 100g dry weight of product approach was said to be resulting in 90% of breakfast cereals being classified as HFSS. The thresholds in the NPM for sugar and salt were said to be too stringent and not to encourage reformulation:

“Also in relation to this point, it was noted that as the NP model scores on 100g of dry weight of product, not as eaten, a greater proportion of breakfast cereals fail the model and cannot be advertised. It was noted that this issue had been considered at length when developing the model but that standard portion sizes remained an issue for the industry. It was noted that the issue also applied to some other foods eaten in quantities less than 100g.” (emphasis added)

Stakeholders were apparently asked a series of questions for the purposes of the Review. I note from a summary of the Panel’s responses to the points made by them that, amongst other things, stakeholders were asked for suggestions as to how the NPM could be improved. Having noted that it had proposed five alternative options for breakfast cereals to the Agency in April 2007, the ACFM said that:

“most ACFM members feel that the best approach for breakfast cereals would be to use a reference amount of 50g because breakfast cereals are a low moisture product. There are similar allowances in the model to account for products which are customarily consumed in larger quantities such as liquids and we feel that a similar derogation could be made in dehydrated or dry/low water content products, generally consumed in smaller quantities and hydrated or consumed with either milk, soya or water”.

In other words, the argument was that the unit for assessment should be reduced to 50g, as a smaller unit for assessment would mean a higher pass rate for dry products. This was not an argument that the unit for assessment should include milk or any other liquid.

The Panel undertook to consider the issue of the low pass rate of breakfast cereals further. In the light of the stakeholders’ concerns, the academic workshop in February 2008 was asked to consider four main issues including whether the assessment should be based on 100g of the product or on portion size, and whether a less stringent approach to the application of the model should be taken in relation to certain categories of food, so as to recognise their contribution to a balanced diet.

The discussions and presentations at the academic workshop included a presentation from Dr Tedstone which considered, in detail, the pros and cons of assessing food “per 100g, per 100kJ, per serving/portion”. However, the academic workshop rejected a move away from the 100g base. The note of this workshop records:

“Secondly, while recognising that a 100g based model did not reflect the actual amounts of some foods being eaten, the group recommended that the base of the model should not be changed. This was because the group was content with the way the model categorised foods. It also reflected concern about the practical difficulties in defining portion sizes, especially for children, together with a concern that a move to per portion could generate unforeseen anomalies in categorisation.”

This conclusion reflected the views of the breakout group which discussed this issue that the key disadvantages were:

As far as a category based approach is concerned, the group discussed whether it was “necessary and justified to make the current model more (or less) stringent for any food type through the addition of separate criteria to the NP model for specific food categories”. It was noted that:

“Concerns were raised by stakeholders representing the dairy, snack food, confectionery, breakfast cereal and chewing gum industries, that the model should be category based because the current model was felt to be too stringent for some categories and it was hindering reformulation. Stakeholders also felt that if a food contributes a significant proportion of micronutrients (e.g. fortified cereals) to children’s diets then advertising restrictions could result in children become micronutrient deficient.”

In relation to breakfast cereals specifically, it was noted that:

“There was a consensus among the group that there was insufficient evidence to support the argument over micronutrient insufficiencies/malnutrition following limiting breakfast cereal advertising to children. It was noted that currently approximately 25% of all breakfast cereals currently meet the model, and that more should be encouraged to pass through reformulation. The group agreed therefore that there was insufficient justification to have a separate category for breakfast cereals.” (emphasis added)

The note records the conclusion of the group as follows:

“Thirdly, the group discussing category based models recommended that the current NP model should not be changed to a category based model or be modified to include additional categories for dairy products, breakfast cereals or dried fruit. The group considered a number of specific options but did not consider in any of the cases that there was a justifiable nutritional argument to include additional categories.”

The Review Panel subsequently concluded that the NPM was “an appropriate and [scientifically] robust tool in categorising less healthy foods” and it agreed with the views of the academic workshop set out above. It then made draft recommendations which went out for consultation in July 2008. One of the key issues consulted upon was the view of the Review Panel that the base for the model should be per 100g rather than per portion.

A key complaint about the model was that it did not “reflect how a food is consumed because it fails to consider any measure of intake, even portion”. It therefore merely measured the inherent properties of a food rather than its actual nutritional impact in practice. Arguments for a portion size approach, or alternatively a different approach to certain categories of food, were put forward by various respondents including the ACFM.

As far as the move to a portion based approach is concerned, the Agency noted that the issue had been considered by the Review Panel:

“[7]…. It concluded that there are distinct disadvantages in moving to a ‘per portion’ base, such as the variation in manufacturers’ declared portion sizes for similar foods and the variability in amounts consumed by children of different ages and as such it supported the 100g approach applied in the FSA model.”

The AFCM had noted that 70% of breakfast cereals were being classified as HFSS despite that fact that breakfast cereals make a beneficial contribution to the diets of children:

“One main reason for this is that it does not reflect how food is consumed – e.g. by portion, as in the case for cereals with milk. ACFM therefore feels that more consideration should be given to nutritional consumption which reflects how the product is actually consumed, but we recognise the review panel have considered this option once.” (emphasis added)

The fact that breakfast cereals are generally consumed with milk therefore was not forgotten at the Review stage but the AFCM acknowledged that it had been considered. The AFCM went on to argue for what it said was a viable alternative of a 50g base unit to which the NPM would be applied. It argued that breakfast cereals have designated and agreed portion sizes from 30-45g according to density “a proposal to judge the cereal on a nominal 50g basis creates an acknowledgement of this” and a modification of this sort in the application of the NPM would be analogous to the approach to drinks which had been taken.

The FDF also noted that the model did not reflect how food is consumed because it failed to consider portion size. It acknowledged that the Review Panel had considered an approach based on portions rather than 100g but expressed disappointment that this had not been adopted.

The Agency response to these arguments reiterated answer [7] on a per portion approach, referred to at [142] above, and supported the Review Panel’s view that:

“there was no justifiable argument to suggest that a category-based approach would be better than a per 100g approach at differentiating foods based on their nutrient content in the context of the broadcast advertising restrictions.”

The Board of the Food Standards Agency was then presented with a detailed paper which set out the history of the decision making in relation to the NPM in 2004-2005, the main issues which had been raised by stakeholders in the Review process (which did not include whether breakfast cereals should be assessed with milk), the conclusions of the Review Panel and the agreement of the SACN so far as the relevant issues for present purposes are concerned. At a meeting on 25 March 2009 the findings of the Review Panel were accepted.

The first version of the Technical Guidance was published in April 2009. It was in materially the same terms as the January 2011 version and, as will have been appreciated, it reflects the decisions which had been taken by the Agency.

I note that the evidence in the 25 March 2009 report to the Food Standards Agency Board suggests that the restrictions introduced by Ofcom had a substantial impact on the level of food and drink advertising directed at children. However, there was no rationality or human rights challenge from any food manufacturer to the use of the NPM or the Technical Guidance as the basis for Ofcom’s decision making nor, indeed, any legal challenge. This, and the lack of any evidence of the issues raised in Grounds 3A and 3B being raised during the subsequent decade during which the NPTG was in operation, lend further support to Sir James’ submission that these issues had been resolved even if not in the way that Kellogg would have preferred.

I also reject Mr Hickman’s claim that in the period 2004 to 2009 the issues which Kellogg now raises had not been considered, or had not been considered adequately or in sufficient detail. They plainly had been the subject of detailed research and consideration by a substantial number of experts and expert bodies over a lengthy period of time as part of the fundamental issue as to whether the approach to identifying less healthy foods should be “as consumed” or “as sold”. The issue from the outset was whether the approach should be based on the inherent nutritional content of the food or drink, or whether attempts should be made to assess the nutritional impact of the food or drink in practice, by reference to how a typical portion was consumed. Initially the main arguments as to how to achieve the latter approach in the case of breakfast cereals involved arguing that the consumption of milk should be taken into account, and comparisons were made with the approach to so called reconstituted foods for the purposes of this argument. These are the arguments which Kellogg now puts forward under Ground 3B and they were considered and rejected by the experts.

In the Review which took place from 2007-2009, the main arguments were that an “as consumed” approach should be achieved by assessing a quantity of the food which was closer to the quantity per portion consumed in practice, and comparisons were made with the approach to drinks under the NPM. The category based modifications approach was put forward as an alternative basis on which to recognise the beneficial impact of the consumption of breakfast cereals on diet. These arguments were about the application of a nutrient profiling model to the food in question rather than assessing it in combination with other foods and ingredients. They called, in effect, for a modification of the NPM rather than a departure from this fundamental principle.

I agree with Mr Hickman that what he called “the milk issue” does not appear to have been the focus of the Review in 2007-2009, although it was considered. However, this appears to have been because the relevant stakeholders, including Kellogg, recognised that the issue had been considered in detail when the model was developed, and had been resolved. They had the opportunity to reopen the issue, if they wished, but they chose to emphasise different arguments for bringing the approach closer to an “as consumed” approach, including that the unit for assessment should be adjusted and/or there should be modification of the criteria applied to breakfast cereals as a category.

By the end of the five year period during which the NPM was under consideration, a range of different variations of the “as consumed” approach had been considered and rejected by the experts save, arguably, in the case of foodstuffs which had to be prepared before they could be eaten and where the manufacturer’s instructions for preparation provided a basis for a standard approach to assessment. The limitations of the “as sold” approach were recognised and understood – the NPM itself did no more than differentiate foods according to their nutritional content on a like for like basis – but this was determined to be an acceptable outcome in terms of achieving the relevant policy objectives. The “as consumed” approach was rejected because of the very significant practical problems with achieving a reliable, transparent and fair test, and its potential for undermining the relevant policy objectives. The objections to the “with milk” approach were essentially the same as the objections to the other “as consumed” approaches which were advocated by the various stakeholders.

On 25 June 2018, the Department of Health and Social Care published chapter 2 of the Childhood Obesity Plan which included its intention to restrict the promotion of HFSS food and drink by location and price. It was also announced that the Government intended to consult on the measures which were to be proposed.

On 12 January 2019, the Government then launched a public consultation in relation to what was to become the 2021 Regulations (“the 2019 Consultation”). The closing date for responses was 6 April 2019. The consultation document, entitled “Consultation on restricting promotions of products high in fat, sugar and salt by location and by price” (emphasis added), explained the relevant policy aims and covered all of the potential issues in relation to the proposals. Of particular relevance in the present context, it explained that:

“Our aim is to restrict promotions of those products that contribute the most sugar and calories to children’s diets. Therefore, we propose that the restrictions should only apply to products which are classed as HFSS and are included in the PHE’s sugar and calorie reduction programmes….

We propose that the 2004/5 Nutrient profiling model (NPM) should be used to define HFSS food and drink products because it is based on scientific evidence and it is already used by industry to determine which products can and cannot be advertised to children. Please see Annex 4 for further details on the 2004/5 NPM.” (emphasis added)

The consultation document then set out a series of “Consultation Questions”. In a set of three questions on “Definitions” Question 27 asked consultees:

“Do you think that the 2004/5 Nutrient profiling model (NPM) provides an appropriate way of defining HFSS products within the food and drink categories proposed for inclusion in this policy (see Annex 4)? Yes/No. If you answered no, what other ways could we use? Please explain your suggestions.”

Annex 3 then listed the product categories which were included in the Government’s sugar reduction programme. As was well known in the sector in any event, these included breakfast cereals.

Annex 4 then explained the NPM, beginning with an overview which included the following:

“The 2004/5 Nutrient profiling model (NPM) was developed by the Food Standards Agency (FSA) to provide Ofcom, the broadcast regulator, with a tool to differentiate foods on the basis of their nutritional composition. Ofcom uses the outputs from the model to regulate the television advertising of foods to children.”

The Annex then set out the NPM itself, including the system for calculating A points, C points, and the overall score, albeit the Annex did not include the Technical Guidance. The fundamental principle that “The points for each nutrient are determined based on the amount of each per 100g of the food or drink” was therefore stated in terms.

Significantly:

On 28 December 2020, the Government then published its response to the 2019 Consultation. This document set out the background, the policy objectives pursued, and why, and it gave a detailed account of the responses to Question 27 in the 2019 consultation document. It said that, having considered the responses, it had been decided that the NPM would be used to determine which products were HFSS, and that the revised NPM, which had yet to be published, would not be used. This document contained, at Annex 5, what was essentially Annex 4 to the January 2019 document with links to the NPTG in the text and in a footnote.

Also on 28 December 2020, the Government launched a consultation on how the proposed regulations should be enforced (“the Enforcement Consultation”). The Enforcement Consultation document included, at Annex A, a draft of the proposed Regulations which defined “specified food” by reference to “the 2004/05 Nutrient Profiling Model”. This, in turn, was defined as “the tool developed by the Food Standards Agency to identify food which is high in fat, salt or sugar”, with a footnote which contained a link to the NPTG and said that hard copies of “guidance on the application of” the NPM could be obtained from the Obesity Team at the Department of Health and Social Care. The closing date for the Enforcement Consultation was 22 February 2021.

There were then written responses to the Enforcement Consultation from BCUK, the FDF and Kellogg itself. The Kellogg response was dated 22 February 2021. It emphasised the nutritional benefits of “cereal with milk” and went on to address various questions which had been raised in the Enforcement Consultation document, albeit with a focus on enforcement measures. These included questions about how the draft Regulations dealt with HFSS products and a general question, number 31: “Are there any comments on the draft of the regulations?”. But no issue was raised as to the use or application of the NPM.

In her evidence, Dr Hoyland draws attention to the fact that the text of the consultation documents and the original draft Regulations referred to the NPM but not the Technical Guidance. She says that therefore Kellogg could not have been expected to raise the fundamental flaw in the NPM which it now contends there is “as there was no way to know whether the Defendant may have been rethinking the approach in the Technical Guidance”. I note that she does not actually say that it was not appreciated that the proposal was that the Technical Guidance would also continue to apply, still less that it was not appreciated that the assessment under the NPM would be based on 100g of the breakfast cereal. I regret to say that I found this argument disingenuous and was not surprised that it was not at the forefront of Mr Hickman’s submissions.

In my view no one in the sector could have failed to appreciate from the outset that the proposal was that the NPM would be used to identify HFSS products in the same way as it had been for more than a decade in the context of television advertising to children. The essential point – that the model would assess the nutritional profile of 100g of the food or drink itself – which had been argued about at length in 2004-2009, was stated in terms in the NPM and, in any event, it was clear from the documents as a whole that the proposal was to take the same approach as had been applied in the past. None of the consultees who provided responses sought clarification on this point.

In my judgment, the true position is that the fact that, in their detailed responses, none of the breakfast cereal manufacturers raised the issue during the consultation period of more than a year tends to support the view that the “as sold versus as consumed” issue had long since been resolved, was well understood and was accepted in the sector. They also understood that the approach under the NPM was in accordance with the Government’s stated objectives, as set out in the consultation documents including the passages which I have quoted above. I return to this point in considering Ground 3B below.

After consideration of the responses to the 2019 Consultation, there was then a detailed Ministerial submission to the Defendant dated 4 October 2019. The proposal was that the restrictions under the draft Regulations should apply to categories of product which are the most significant contributors to the sugar or calorie intakes of children and which are heavily promoted to customers. Breakfast cereals were clearly identified as one of the categories which it was proposed to include.

As Mr Hickman points out, the submission did not specifically draw attention to the fact that the ACFM consultation response had included arguments that breakfast cereals should not be subject to the proposed restrictions. This would have drawn more attention to the case of breakfast cereals. However it was made clear, in the submission, that there had been arguments for the same approach as was taken in Scotland, which was based on a list of discretionary products which the submission set out, and which did not include breakfast cereals. The submission went on to recommend that the Scottish approach was not taken. The categories of potential specified foods would include the discretionary products but would be wider and would include breakfast cereals. As Mr Dodds confirms in his first witness statement, the Defendant was expressly advised that breakfast cereals should be included. The reason for doing so was that they are one of the most significant contributors to the sugar and calorie intake of children in the United Kingdom and the relevant data were provided in Annex B to the submission.

The 4 October 2019 submission specifically recommended that “the 2004/05 Nutrient Profile Model (NPM) is used to define HFSS products”. There was a section of the submission headed “Model to define HFSS products” which said the following:

“The consultation proposed that the 2004/05 NPM is used to define HFSS products. The NPM is based on scientific evidence and it is already used by industry to determine which products can or cannot be advertised to children on TV during children’s viewing times. The NPM provides an overall assessment of the nutritional content of products as it accounts for nutrients of concern (fat, sugar, salt and calories) as well as beneficial nutrients (fibre, fruit and vegetable content).” (emphasis added)

It was perfectly clear from this that the approach was to assess the nutritional content of the product, rather than the nutritional content of a dish or meal of which it formed part. The Defendant was therefore made aware that the assessment of the product under the NPM was “as sold” rather than “as consumed”.

There was then an account of the consultation responses on the proposal to use the NPM which included that 40% of respondents were in favour, 39% were against and 21% did not provide a view. There was also a summary of some of the arguments put forward by business, including that the NPM was regarded as too strict and would require unrealistic reductions in the sugar, fat or salt content of their products. But it was noted that, as was the case, no alternative model had been put forward. Annex A to the submission also contained a detailed analysis of the consultation responses. The submission specifically asked the Defendant whether he agreed with the use of NPM.

I was shown a further submission to the Defendant dated 10 March 2020, which annexed the October 2019 submission for ease of reference. The aim of this submission was to confirm that the Defendant was content with the decisions which had been taken, particularly in relation to three aspects of the proposals, in the light of further data which had been gathered.

I was then shown part of a follow up submission, apparently dated 10 June 2020, from which it appeared that the Defendant had asked various questions about the proposals, including pertinent questions about the proposed categories of food to which the draft Regulations would apply. Mr Hickman drew attention to the response to a request for “more background on the HFSS NPM definitions”, which effectively repeated what had been said in the October 2019 submission in the passage cited at [168] above. It then added:

“The 2004/5 NPM was developed by the Food Standards Agency to provide Ofcom, the broadcast regulator, with a tool to differentiate foods on the basis of their nutritional composition. Ofcom uses the outputs from the model to regulate the television advertising of foods to children. The NPM scores foods based on their nutritional content. The nutrients considered are split into two categories – A and C. The score for ‘C’ nutrients is subtracted from the score for ‘A’ nutrients to give the final score. A higher score indicates a less healthy food.

‘A’ nutrients consist of energy, saturated fat, total sugar and sodium. ‘C’ nutrients consist of fruit, vegetables and nut content, fibre and protein. Therefore, a food scoring highly on ‘A’ nutrients is not automatically classified as less healthy, only if it additionally scores little on ‘C’ nutrients.

Foods scoring 4 or more points, or drinks scoring 1 or more points, are classified as ‘less healthy’. Therefore if a food category in scope has score of 4 or more (or drinks in scope 1 or more) then the restrictions will apply.”

Mr Hickman’s point was that the submissions to the Defendant made no reference to the Technical Guidance. Here, the Defendant had asked for information about how the NPM worked but the response did not explain in any detail, or materially add to what was known from the consultation documents. It did not mention the “as consumed” approach, or the approach in relation to so called reconstituted foods, still less any issue in relation to breakfast cereals.

I do not accept that these criticisms carry any real weight. The fundamentals of the approach remained perfectly clear in the light of the further explanation provided. If the Defendant asked himself whether the NPM measured the nutritional content of a breakfast cereal or the breakfast of which it formed part, the answer was quite apparent. It was not incumbent on civil servants to draw attention to the myriad issues which had been considered and resolved in 2004-2009, of which “the milk issue” was but one, nor on the Defendant to consider them, particularly in circumstances where “the milk issue” had not been raised in the course of the 2019 Consultation. If the Defendant considered that the answer with which he had been provided was inadequate, he could, and no doubt would, have asked for further information.

There were then two lengthy and detailed Impact Assessments in relation to the proposed measures dated 11 November 2020. These examined various options against the relevant policy considerations and costed them in terms of the impact on business and the potential benefits to the National Health Service and the wider economy. It was estimated that affected manufacturers could suffer losses of profit of in the order of £231.8m annually. However, the Government considered that this was well within the bounds of proportionality given the estimated health benefits both in human and financial terms. The value of the health related benefits of the 2021 Regulations was estimated to be around £77 billion over 25 years compared with a cost to business of around £5.5 billion over the same period.

In his second witness statement Mr Dodds says, and I accept, that on 17 November 2020 there was a submission to the Defendant which attached a draft of the consultation response of 28 December 2020 referred to at [161] above. He exhibits the relevant parts of the draft - i.e. draft Annex 5 – which set out the NPM in materially the same terms as it appeared in the published draft. A weblink to the NPTG itself was also included at the end of the draft submitted to the Defendant.

On 27 April 2021, a draft of the Government response to the Enforcement Consultation was also submitted to the Defendant. This made five references to the Technical Guidance which also explained its function, three of which references were weblinked to the NPTG itself. In four of the references it is explained, amongst other things, that “This document aims to answer frequently asked questions about the application of the model to different types of products through a simple guide, Q&A section and worked examples”.

It was on 11 June 2021 that a Mr Chris Silcock of Kellogg sent a letter to the Permanent Secretary to the Department of Health and Social Care which raised, for the first time, a claim that Kellogg had “uncovered” an inconsistency in the treatment of the breakfast cereal category in the Technical Guidance. This was said to be that certain cereals, such as porridge oats, were assessed on an “as prepared” basis whereas other breakfast cereals were assessed on an “as sold” basis despite the fact that the vast majority of breakfast cereals are consumed with milk. This was said to place a disproportionate burden on ready-to-eat cereals and not to reflect how cereals are actually eaten. Arguments as to the health benefits of breakfast cereals were also put forward, and the letter requested a meeting to discuss the matter and for civil servants to engage with the Kellogg’s team “on this technicality”.

A “technical dossier” was also enclosed with the letter of 11 June 2021, which said that 87% of breakfast cereals were consumed with milk and developed various arguments which Kellogg put forward in favour of an “as consumed” approach to the assessment of the nutritional value of breakfast cereals. The proposal was that the FAQ section of the Technical Guidance should be “updated” so that the answer in relation to breakfast cereals would provide that “for the entire breakfast category” the score:

“should be calculated based on the recommended serving size of the product including an appropriate quantity or milk, to account for the manner in which breakfast cereal is frequently consumed” . (emphasis added)

I note that the inconsistency alleged was in the treatment of porridge oats as compared with ready-to-eat breakfast cereals, rather than breakfast cereals compared with foods which needed to be reconstituted or prepared before they could be eaten. In fact, the required approach for porridge oats was the same as for other breakfast cereals, so there was no such inconsistency. The Kellogg arguments did not address the point that the NPM itself provided that the nutritional profile would be calculated “on the basis of the nutrient content of 100g of [the] food or drink” or propose any amendment to the NPM in this regard. In effect, it argued for the Technical Guidance to exempt breakfast cereals from this principle. The “update” proposed was also vague and problematic for the reasons identified when essentially the same approach was advocated in 2004-2009. Dr Hoyland’s argument that a fixed ratio of milk to cereal could be adopted was not put forward.

There was then a follow up letter from Mr Silcock, on 24 June 2021, which drew attention to data which suggested that there had been a decline in the consumption of breakfast cereals during the pandemic in favour of an increase in cooked breakfast. This was said to be bad for the nation’s health and the argument appeared to be that food which might form part of a cooked breakfast should be included in the specified food categories.

There was then a meeting between Kellogg and officials from the Department and Public Health England on 29 June 2021 about which very little evidence was put before the court.

The Permanent Secretary then followed up on the 29 June 2021 meeting with a letter to Mr Silcock dated 5 July 2021. This letter defended the “as sold” approach under the NPM and explained the decisions which had been taken when the NPM had been developed and then reviewed in 2007. It also corrected Kellogg’s misapprehension as to the approach to porridge oats and undertook to ensure that the approach to them was clear under the Technical Guidance. The letter addressed other arguments which had been put forward by Kellogg and indicated a willingness to continue a dialogue.

On 9 July 2021, Mr Silcock wrote a further letter which developed Kellogg’s arguments for an “as consumed” approach. Contrary to Mr Hickman’s case, this letter specifically referred to FAQ (c) in the NPTG, which said that breakfast cereal should be assessed on the basis of 100g of the product as sold, and said “whilst the logic of such a stipulation may have been unclear, such a requirement has been relatively uncontroversial given the prior function(s) served by the NPM” (emphasis added). Mr Silcock did not explain why the wider use of the NPM now rendered it controversial but it is to be inferred that this was because of its wider impact, including on Kellogg’s commercial interests. As for whether the logic of the approach was or was not clear, the letter also acknowledged that the “as consumed” approach would create complexity albeit this was said not to justify “avoiding the issue”. The substance of the arguments which the letter advanced for assessing breakfast cereals with milk, including by reference to a comparison with the treatment of foods which need to be prepared before they can be consumed, had been considered and rejected in 2004-2005. But, despite the meeting of 29 June 2021 and the letter of 5 July 2021, Mr Silcock appeared to be ignorant of this when he said that his “clear impression” was that officials had not given “due consideration” to the issues which he raised.

On 15 July 2021, a further submission to the Defendant was made which sought confirmation that he was content to lay the Regulations now that they had been scrutinised by the Joint Committee on Statutory Instruments, and to clear the Explanatory Memorandum. A draft of the Explanatory Memorandum was included with the submission and it included the following passage, at [10.4], as part of a summary of the decisions made by Government as a result of the consultation which had taken place:

“The technical guidance specified in the regulations (currently known as the “Nutrient Profiling Technical Guidance”) published by the Department of Health on 1 January 2011 which supports the 2004/5 Nutrient Profile Model….will be used to define HFSS products within the specified list of product categories…..”

The draft Regulations were laid before Parliament on 21 July 2021 as I have said. Also on 21 July 2021, the Government published its response to the Enforcement Consultation.

On 3 September 2021, Kellogg submitted its PAP letter. This led to two submissions to the Defendant which are subject to legal advice privilege which is not waived, and therefore were not disclosed, although Mr Dodds has summarised the material parts of them in his second witness statement. The first, on 30 September 2021, notified him of the nature of the potential challenge. The second, dated 13 October 2021, asked for the proposed response to Kellogg’s PAP letter to be cleared. The PAP letter itself was annexed to both of these submissions.

In the 30 September 2021 submission, the reasons for the approach under the NPM were explained. It was also noted that this approach had been decided upon after a review by the Expert Group in 2007, and the NPM had been in use ever since in the context of advertising on children’s television.

The 13 October 2021 submission gave the examples of two Kellogg cereals – “Frosties and Krave” – which were classified as less healthy on an “as sold” basis but were not if they were assessed with 125ml of semi skimmed milk. The point that the approach under the NPM had the approval of the Expert Group after consultation and a review was reiterated. The argument that the approach to breakfast cereals was inconsistent with the approach to foods which required to be reconstituted was addressed, and it was said that these foods are different in that they are added to other foods and ingredients in accordance with manufacturer’s instructions before they are consumed. It was also said that the approach under the NPM was consistent with the approach in other countries.

The Defendant’s response to the PAP letter was approved by him and it was sent on 18 October 2021. This said that Kellogg’s arguments in the exchanges from 11 June 2021 had been considered by the Defendant albeit it is clear that this was in the context of considering the pre-action correspondence.

The legal tests to be applied are very familiar. I was taken by Mr Hickman on a tour of the decision of the Court of Appeal in R (National Association of Health Stores & another) v Department of Health [2005] EWCA Civ 154 but the test applied in that case, as in Re Findlay [1985] AC 318, 334B, was that which was stated in CREEDNZ v Governor General [1981] NZLR 172 where, as Sedley LJ put it at [63] of the Health Stores case:

“Cooke P drew the distinction, which our courts had previously failed to draw, between things which are so relevant that they must be taken into account and things which are not irrelevant and so may legitimately be taken into account. It is axiomatically only a failure to take into account something in the former class that will vitiate a public law decision”.

At [64] Sedley LJ referred to the issue as being whether “either the statutory purpose or the nature of the issue before the minister made [the consideration] so relevant that a lawful decision could not be taken in ignorance of [it]”. At [75] Keene LJ referred to such matters as ones which “the decision-maker is bound to take into account”.

As to the Tameside duty, the position is summarised by Underhill LJ in the now very well-known passage from the judgment in R (Balajigari) v Secretary of State for the Home Department [2019] EWCA Civ 673; [2019] 1 WLR 4647 at [70]:

“First, the obligation on the decision-maker is only to take such steps to inform himself as are reasonable. Secondly, subject to a Wednesbury challenge … it is for the public body and not the court to decide upon the manner and intensity of inquiry to be undertaken…. Thirdly, the court should not intervene merely because it considers that further inquiries would have been sensible or desirable. It should intervene only if no reasonable authority could have been satisfied on the basis of the inquiries made that it possessed the information necessary for its decision. Fourthly, the court should establish what material was before the authority and should only strike down a decision not to make further inquiries if no reasonable authority possessed of that material could suppose that the inquiries they had made were sufficient. Fifthly, the principle that the decision-maker must call his own attention to considerations relevant to his decision, a duty which in practice may require him to consult outside bodies with a particular knowledge or involvement in the case, does not spring from a duty of procedural fairness to the applicant but rather from the Secretary of State’s duty so to inform himself as to arrive at a rational conclusion. Sixthly, the wider the discretion conferred on the Secretary of State, the more important it must be that he has all the relevant material to enable him properly to exercise it.”

Ultimately, and consistently with Mr Hickman’s written submissions of 5 May 2022, the evidence about what happened after 21 July 2021 does not take the matter further in terms of whether Ground 3A is established, although it might have been relevant to relief if the issue arose. This is because the decision under challenge is the decision to lay the 2021 Regulations before Parliament and because, in any event, what happened after this did not materially affect the factual basis for Mr Hickman’s pleaded case. In particular, the evidence did not show that the specific matters pleaded under Ground 3A had been considered by the Defendant before the submission of 30 September 2021, when he was considering it in the context of threatened litigation. There is no suggestion that he remade his decision at this point. I therefore accept that Ground 3A falls to be determined on the basis that, as stated in the Defendant’s 1April 2022 Response to the Claimants’ Request for Information, he was not asked in any relevant Ministerial submission to consider the issues in relation to breakfast cereals which were considered or decided in 2004-2005 or 2007-2009, nor the quality of the process by which they were decided.

By the same token the question under Ground 3A is, in effect, whether the Defendant was bound to consider the pleaded matters and/or to make further inquiries into them or into the decision-making process which was undertaken between 2004 and 2009. I am not persuaded that he was.

Firstly, the information provided to the Defendant in the relevant submissions to him made perfectly clear that the approach under the NPM was to assess the nutritional value of the food itself i.e. he was well aware of the rule which is under challenge in these proceedings and evidently approved it. It was open to him to make such further inquiries about the NPM as he considered necessary; indeed, he did make inquiries in June 2020 as I have noted. He also had access to the NPTG. He evidently considered that he was sufficiently well informed about the approach which would be taken to identifying foods which were HFSS.

Second, as Sir James Eadie submitted and as I have found above, the relevant matters had been carefully considered in the course of the development and review of the NPM. The issues which Kellogg raised then, and raises now, had been resolved. The approach stated in the NPM had been applied for more than a decade without controversy. There had then been ample opportunity to raise these issues in response to the 2019 Consultation and, indeed, for a year after it closed in April 2019, but the approach under the NPM had not been challenged. Until Mr Silcock’s letter of 11 June 2021, there was no issue as to the approach which Kellogg now challenges.

Third, the fact that there had been an issue more than a decade earlier was not a fact that the Defendant was bound to take into account in making his decision. Nor was there anything which would make it irrational for him to fail to make inquiries into issues which had not arisen in the context of the 2019 Consultation or the Enforcement Consultation, and how they had been resolved in the past. His decision was no more than that a well-established industry benchmark would be used for the purpose for which it had been used in the past and in a related context.

Fourth, in my view the fact that a form of Kellogg’s present argument was raised on 11 June 2021 and developed between then and 21 July 2021 does not alter this conclusion. By now, the public consultations, which gave Kellogg and other breakfast cereal manufacturers ample opportunity to raise the so called “milk issue”, had long since closed. The decision that the NPM would be applied had long since been taken in the light of the responses to the 2019 Consultation and had been publicly announced on 28 December 2020. There had then been further decisions as to enforcement mechanisms taken in the light of the Enforcement Consultation. The fact that Kellogg came forward at the eleventh hour, seeking to reopen decisions as to the “as sold vs as consumed” approaches which had been taken more than a decade earlier, and re-taken in 2020/2021, did not render those decisions irrational or mean that it would be irrational for the Defendant to proceed with those decisions rather than pause the process and make further inquiries.

Nor did the merits of Kellogg’s arguments mean that they had to be taken into account or inquired into again. Leaving on one side the point that Kellogg’s initial argument based on inconsistency was misconceived, for reasons which I will explain in relation to Ground 3B, the argument which they have since developed does not have sufficient merit to render it irrational to proceed with the decisions which had been taken. Nor did it render the substantive decision itself irrational. There was no fact or consideration which the Defendant was bound to take into account, or inquire into, but did not.

I therefore reject Ground 3A.

Kellogg’s pleaded case is summarised at paragraph [11], above. There was no dispute that A1P1 and Article 10 ECHR are engaged:

The question was therefore whether the Defendant could “justify” the relevant infringements. Here, again, there was no issue that the aim of the impugned measure is legitimate, that is, “public safety….the protection of health” (Article 10(2)) and “the public interest” (A1P1).

The dispute is as to proportionality. In this connection there was only limited controversy as to the applicable legal principles. The burden of proof is on the Defendant. The test is as stated by Lord Reed in Bank Mellat v Her Majesty’s Treasury (No 2) [2013] UKSC 39; [2014] AC 700 at [74]:

“74.

….it is necessary to determine (1) whether the objective of the measure is sufficiently important to justify the limitation of a protected right, (2) whether the measure is rationally connected to the objective, (3) whether a less intrusive measure could have been used without unacceptably compromising the achievement of the objective, and (4) whether, balancing the severity of the measure's effects on the rights of the persons to whom it applies against the importance of the objective, to the extent that the measure will contribute to its achievement, the former outweighs the latter. …. I have formulated the fourth criterion in greater detail than Lord Sumption JSC, but there is no difference of substance. In essence, the question at step four is whether the impact of the rights infringement is disproportionate to the likely benefit of the impugned measure.”

In the case of measures relating to public health, the courts will accord a measure of discretion to the public authority. As the late Laws LJ put it in R (Sinclair Collis Ltd) v Secretary of State for Health [2011] EWCA Civ 437; [2012] QB 394 at [49]:

“Public health, and perhaps especially the health of minors, is surely the particular responsibility of elected government. It is in my judgment a strategic goal or aspiration of such importance as to confer a broad margin of appreciation on the decision-maker. So much is well demonstrated on the cases.”

However, Mr Hickman submitted that this principle is applicable where the matter has been considered. Where it has not been, there is nothing to which a margin of appreciation can be afforded. In this connection he relied on Belfast City Council v Miss Behavin’ Ltd [2007] UKHL 19; [2007] 1 WLR 1420 at [27], [37], [46]-[47] and [90]-[91]. He argued that there had been no consideration of the proportionality of including breakfast cereals on the list of potential specified foods on a basis (the NPM) which did not reflect how they are consumed and so materially affects whether they are classified as HFSS.

The principle on which he relied is, in effect, that, all other things being equal, the greater the consideration given to the affected interests by the decision maker, the greater the deference which the court is likely to accord to the view of the decision maker. As Baroness Hale put it at [37]:

“So the court has to decide whether the authority has
violated the Convention rights. In doing so, it is bound to acknowledge that
the local authority is much better placed than the court to decide whether the right of sex shop owners to sell pornographic literature and images should be restricted for the prevention of disorder or crime, for the protection of health or morals, or for the protection of the rights of others. But the views of the local authority are bound to carry less weight where [it] has made no attempt to address that question...where there is no indication that this has been done, the court has no alternative but to strike the balance for itself, giving due weight to the judgments made by those who are in much closer touch with the people and the places involved than the court could ever be.”

The principle is therefore more flexible and less likely to be binary than Mr Hickman appeared to suggest. He added that where “ex post facto” reasons are relied on, the Court should scrutinise the application of the proportionality standard with particular care, and he referred to Re Brewster [2017] UKSC 8; [2017] 1 WLR 519 at [52] and [64]. Again, the point to which he referred was that, as the late Lord Kerr put it, “where the question of the impact of a particular measure on social and economic matters has not been addressed by the government department responsible for a particular policy choice, the imperative for reticence on the part of a court….is diminished” and the level of scrutiny of the validity of the “justification” will intensify to take account of the fact that it is put forward ex post facto [64]. But, consistently with the passage from Miss Behavin’ cited above, Lord Kerr also stated that “Even retrospective judgments…if made within the sphere of expertise of the decision maker, are worthy of respect, provided they are made bona fide” [52].

Mr Hickman went on to argue, in any event, that although there was no challenge to the inclusion of breakfast cereals in Schedule 1 to the 2021 Regulations, the objection is to the assessment of breakfast cereals without any reference to how they are actually consumed. Relying on the evidence of Dr Hoyland and Mr McGowan, he drew a contrast with other specified foods such as snacks and biscuits which, he said, did not have recognised importance as part of a daily meal, were not an important source of vitamins and minerals and did not significantly contribute to the consumption of milk. He also emphasised that breakfast cereals are not designed to be eaten dry, are not marketed to be eaten in their dry form, and are not in practice eaten dry. Assessing them on a dry weight basis, he submitted, was therefore artificial and meant that the assessment of breakfast cereal under the NPM was not an assessment of their impact on health. There was therefore no rational connection between the assessment and the aim of the relevant measure.

As to the third limb of the Bank Mellat formulation, Mr Hickman argued that there was a less intrusive means of achieving the aim in question. He argued that there is no satisfactory explanation for the approach to so called reconstituted foods despite the evidence that they may be prepared and/or consumed in a variety of ways, and not necessarily in accordance with the manufacturer’s instructions. They are also often eaten with, or added to, other foods. He advocated what he called a “less intrusive and more rationally tailored approach” based on assessing a portion of the breakfast cereal with the amount of milk specified in what he described as “the manufacturer’s recommended servings”, or applying a fixed ratio of cereal to semi skimmed milk for all breakfast cereals.

For the same reasons, Mr Hickman submitted that a “dry weight” approach was irrational because it was not within the range of approaches reasonably open to the Defendant in the circumstances.

In my view, this is a case in which a significant degree of deference should be accorded to the decision maker(s).

This, then, is not a case where no thought was given to the relevant matters and there is nothing to which the court ought to show deference. It is a case where the expert judgment of the public health, nutrition and other experts who worked on the development and review of the NPM, as well as the judgment of the Defendant, should be accorded respect by the court.

Even if that were not so, however, in my view the relevant aspects of the Defendant’s approach under the 2021 Regulations are proportionate and rational. The measure under challenge is the placing of the relevant restrictions on the promotion of food products which are classified as HFSS. The aim of this measure, as stated from the outset of the January 2019 Consultation (see [155], above) and at [2.2] of the Explanatory Memorandum to the 2021 Regulations is:

“…to reduce overconsumption of HFSS products that can contribute to children being overweight or living with obesity….to shift the balance of promotions towards healthier options and maximise the availability of healthier products available on promotion…”

It is important to note all of the elements of this aim. It entails identifying “products” which are high in fat, sugar or salt relative to other products, so that they can be made subject to measures which restrict their promotion. The promotion of products which are lower in fat, sugar or salt will thereby be encouraged because there are no relevant restrictions placed on them, or because popular products which are HFSS are reformulated so as to reduce their levels of fat, sugar and/or salt so that they can continue to be promoted in the relevant ways.

It seems to me that when these points are focussed on, a number of Mr Hickman’s arguments fall away or do not engage with the point. Crucially, there plainly is a rational connection between these aims and the impugned measures for reasons which, for the most part, hardly need be spelt out. The approach of the 2021 Regulations specifically reflects their stated aims.

The context is that, as is well known, there is an epidemic of childhood obesity. The findings of the Government’s National Child Measurement Programme for England are that, amongst 4-5 year olds, levels of obesity have increased from 9.9% in 2019/2020 to 14.4% in 2020/2021. Amongst 10-11 year olds, they have increased from 21% to 25.5%. Although Mr McGowan argued in his witness statement that these increases were a result of the Covid-19 pandemic, it was not clear why he thought this point made any difference to the analysis. Children who are obese are more likely to suffer health issues including type 2 diabetes, musculoskeletal pain and asthma. They are more likely to be bullied at school and to suffer psychological issues, and they are more likely to be obese as adults and, as a result, run an increased risk of heart disease, diabetes, stroke and certain cancers. The latest estimate of the cost to the NHS in England of obesity related health issues is £5.1 billion (in 2014/2015) and the estimated total cost to the wider economy is in the order of £27 billion.

Second, the root cause of the prevalence of obesity and excess weight in the population is excess calorie consumption. Excess sugar consumption is associated with increased calorie consumption. This was the conclusion of a report by the SACN on “Carbohydrates and Health” in 2015. The SACN’s finding was that current intakes of sugar are too high and are contributing to the levels of obesity in the population. The nature of the causal link is complex but the SACN found that the higher a person’s consumption of sugars, the more likely they are to exceed their estimated average requirement for energy. It recommended that average intake of free sugars (i.e. sugars added to the food or drink plus sugars naturally present in honey, syrups and fruit juices) should not exceed 5% of total dietary energy and that lowering intakes of free sugars made it more likely that a person’s estimated average requirement for energy would not be exceeded, and could go some way to addressing the problem of obesity. The SACN also found that in the United Kingdom, children are consuming more than double the recommended maximum amount of free sugars.

A third, important, point is that food promotion and advertising has a marked effect on the choices which the public make in relation to food. Public Health England’s report “Sugar Reduction: the Evidence for Action” (2015) found that foods on promotion account for around 40% of all expenditure on food and drinks consumed at home, and that higher sugar products are promoted more than other foods. The view that advertising and promotions affect people’s choices, and therefore the content and balance of their diet, was recently endorsed by the World Health Organisation in its report: “Monitoring and restricting digital marketing of unhealthy products to children and adolescents”. It is therefore a key aspect of the Government’s obesity reduction strategy to incentivise businesses to re-balance their portfolios of products in favour of healthier products by placing restrictions on the promotion of HFSS products.

Fourth, Kellogg does not challenge the inclusion of breakfast cereals on the list of foods and drinks in Schedule 1 to the 2021 Regulations. This concession is rightly made. It renders a number of the arguments advanced by Dr Hoyland and Mr McGowan, as to whether breakfast cereals as a category are good or bad for the health and/or as to the extent to which they contribute to obesity, somewhat peripheral to the issues in this case. However, breakfast cereals are a significant contributor of calories and free sugars in children’s diets. According to the National Diet and Nutrition Survey, for children aged 4-10 years, they contribute 7.2% of the daily intake of free sugars and 5.7% of the calories. The figures are 7% and 4.8% respectively in the case of 11-18 year olds. These percentages for free sugars are roughly equivalent to the percentages contributed by confectionary, jam and sweet spreads, and not far short of the percentage contributions of buns, cakes and pastries, biscuits and fruit juices. Breakfast cereals are therefore an appropriate category of food to regulate and it is for this reason that they were included in Public Health England’s sugar reduction programme: “Sugar Reduction: Achieving the 20%” (March 2017). This, in turn, was a factor which led to them being included in Schedule 1 to the 2021 Regulations, as I have noted in quoting from the January 2019 consultation document at [155] above.

Fifth, as Dr Tedstone and Mr Dodds point out, in any event, the 2021 Regulations do not restrict the promotion of all breakfast cereals. They merely restrict the promotion of breakfast cereals which are HFSS. Again, this tends to render a number of Dr Hoyland and Mr McGowan’s arguments about the merits of breakfast cereals somewhat peripheral to the issue. There is no dispute that breakfast cereals can be part of a healthy balanced diet, contain fibre and can be fortified with micronutrients which bring health benefits to the consumer. Indeed, the Government actively encourages the consumption of healthy breakfast cereals.

But the argument that there are nutritional benefits to the consumption of a given breakfast cereal does not affect the point that if it contains excess fat, sugar or salt, that feature of the product is adverse to a child’s health. Still less is it an argument against seeking to encourage, for health related reasons, the promotion and consumption of breakfast cereals which contain less fat, sugar or salt. Nor does mixing a breakfast cereal which is high in, for example, sugar, with milk alter the fact that it is high in sugar; the “as consumed” approach merely calls attention to, and seeks to rely on, the nutritional benefits of the food or drink with which the breakfast cereal is consumed. The notion that the approach advocated by Kellogg measures the “actual health impact” of a breakfast cereal is therefore problematic. In fact, it does not confine the measurement to the health impact of the breakfast cereal itself; it measures the impact of the cereal combined with other products and seeks to take advantage of the fact that the other products are lower in fat, sugar and/or salt or contain other compensating nutrients.

Kellogg’s argument is not that its products are themselves lower in fat, sugar or salt; it is that they should be assessed in combination with other foods and ingredients, namely semi skimmed milk. At least 21% of consumers of “Frosties” are children aged 0-15. On Dr Tedstone’s evidence, 100g of “Frosties” contains more sugar (37g) than the average chocolate biscuit (28.5g). The suggestion that “Frosties” should not be regarded as a less healthy product because of the nutritional value of the milk with which they may be consumed is surprising. Conversely, it is unsurprising that Kellogg products such as “Crunchy Nut Clusters Milk Chocolate Curls”, “Krave Choc Nut”, “Krave Milk Chocolate”, “Krave White Chocolate” and “Krave Chocolate, Caramel and Peanut” are classified as “less healthy” products under the 2021 Regulations. The proposition that they somehow become healthy products if they are consumed with milk is wholly unconvincing, as the addition of milk does not alter the nutritional profile of the products themselves. These considerations further illustrate the point that Kellogg’s preferred approach is actually inconsistent with the aims of the 2021 Regulations.

Sixth, it is unsurprising that the role of the NPM in this context is merely to distinguish those breakfast cereals which are HFSS from those which are not. As is quite apparent, it is a tool to differentiate products according to their nutritional composition rather than to identify how they are eaten or to assess them, or their impact, on this basis. Again, this is consistent with the aims of the 2021 Regulations. The argument that the approach under the 2021 Regulations should be based on attempting to assess how products are consumed in practice is therefore besides the point or, at least, is an argument for a fundamentally different approach and/or for seeking to achieve a different aim to that which the NPM and the 2021 Regulations seek to achieve.

Seventh, even if this is wrong, and the aim of the 2021 Regulations should be seen more broadly in terms of combatting childhood obesity, Kellogg’s proposed approach unacceptably compromises that aim for the reasons which were recognised in the course of the development and review of the NPM in 2004-2009. In this connection, it is important to bear in mind the criteria by which the model for differentiating less healthy from more healthy foods was to be judged, as set out in the July 2004 Expert Group meeting referred to at [120] above, namely simplicity, transparency, accuracy and workability. The assessment of 100g of the food under the NPM ensures a transparent, like-for-like comparison which is workable, scientifically accurate and robust. I agree with Dr Tedstone that a move to an “as consumed” approach would reduce the effectiveness of the NPM in discerning the healthiness of a given food relative to another and/or substitute a fundamentally different aim and approach. It would entail the comparison no longer being on a like-for-like basis and it would exchange one point of certainty, the nutritional profile of 100g of the food, with three points of uncertainty: the amount of cereal consumed, the volume of liquid consumed with it and the type of liquid consumed with it.

I accept that breakfast cereals are typically consumed with milk and are marketed and presented on this basis. But this is not invariably the case, as Mr Silcock’s letter of 21 June 2021 and the accompanying technical dossier implicitly recognised. They may be eaten dry or with liquids other than milk - e.g. yoghurt, juice or water - or with non-dairy milk or dairy milk other than semi skimmed milk. Breakfast cereals clearly are a food and they can be, and are, consumed in a variety of ways and portion sizes. This is another respect in which it is not accurate to claim that Kellogg’s approach measures the actual health impact of a breakfast cereal: much depends on how much cereal is eaten on a given occasion by a given person and with what liquid or other food, if any.

The ratio based approach advocated by Mr Hickman does not dispose of these difficulties, which were at the heart of the rejection of the various arguments for an “as consumed” approach put forward in 2004-2009. It begs the question what the ratio should be if it is accurately to reflect how breakfast cereals are consumed in practice having regard to differences in eating habits and appetite amongst consumers. Dr Hoyland’s suggested answers have the same defect. She said that the amount of semi skimmed milk should be 125ml, or a cereal to milk ratio of 1:2.2 should be applied. She said that the proposed ratio was a conservative one based on information as to the nutritional values of notional portions of breakfast cereal displayed on breakfast cereal packaging. I was shown Kellogg packaging which displays nutritional values for portions of, for example, 30g or 45g of the cereal itself, and Dr Hoyland referred to packaging used by other manufacturers which displayed nutritional information on alternative bases – with or without milk – albeit in quantities (e.g. 100ml per 45g serving, and 125ml per 30g serving) which differed according to the cereal. Not only was the approach on the packaging not invariable, however; the standard measure proposed by Dr Hoyland remained effectively a notional one rather than one which reflected actual consumption habits, which are not uniform amongst consumers.

Eighth, as to the argument based on so called reconstituted foods, I note that although Kellogg’s pleaded case asserted that there is “no logic from a public health perspective” in adopting a different approach to breakfast cereals ([72] ASFG) a fair reading of its pleading is that its four grounds of challenge did not include a contention that the difference in treatment is irrational in itself. Even if that was its case, I reject it. The foodstuffs to which the relevant approach applies can be rationally distinguished from breakfast cereals. They cannot sensibly be eaten without the addition or use of liquid and/or other ingredients and/or being cooked or prepared. It is for this reason that they have manufacturer’s instructions as to how they should be prepared. If they are to be regarded for present purposes as “foods” when in their “un-reconstituted” form, which is debatable, they are an exception to the general approach under the NPM. The fact that they have instructions as to preparation not only reflects the different nature of these foodstuffs; it also provides a practical and predictable basis for a standardised approach to assessment. Mr McGowan’s argument and evidence that, like breakfast cereals, these foodstuffs will not necessarily be prepared in accordance with the manufacturer’s instructions and/or may have other ingredients added to them and/or may be consumed with other foods does not materially undermine the Defendant’s approach. This is because the point remains that, unlike breakfast cereals, the so called reconstituted foods cannot sensibly be eaten without additional preparatory steps being taken, or “from the packet”.

Breakfast cereals may have liquids added to them or be mixed with other products before they are eaten, but this is not necessarily the case. Many of them can be eaten in dry form. They do not come with instructions for preparation which say that they should be consumed with milk, although I accept that their packaging and messaging will often depict or refer to milk with the cereal and some manufacturers include information as to nutritional values in the event that they are consumed with a specified quantity of milk. The fact that manufacturers provide information as to nutritional values, rather than manufacturer’s instructions, is not just a reflection of the different nature of breakfast cereals; it also illustrates a practical difference in terms of the availability of an effective standardised approach.

Dr Hoyland and Mr McGowan’s arguments therefore do not lead to the conclusion that a further exception should be made to the general approach under the NPM, or a different approach adopted in the case of breakfast cereals. Nor do they come close to showing that it is disproportionate or irrational to apply the underlying approach of the NPM to breakfast cereals. Still less do they establish that this approach is outwith the margin of discretion which it is appropriate for me to apply in this case.

Mr Hickman did not focus his fire on the fourth Bank Mellat question per se but it is worth noting that the restrictions are only to particular types of promotion, and in the larger outlets and online. They also affect all breakfast cereal manufacturers who sell their products through these outlets, and not just Kellogg. Any competitive disadvantage from which Kellogg suffers as a result of the 2021 Regulations is therefore a reflection of the fact that a relatively high proportion of its products and its sales by volume are HFSS, and it is consistent with the aims of the 2021 Regulations that this is so. Given that the restrictions on promotions only apply to breakfast cereals which are HFSS, the central options for Kellogg include using location and price volume promotions to promote its non HFSS products and/or reformulating its HFSS products so that they pass the test under the NPM. Again, this is consistent with the aims of the 2021 Regulations. Moreover, breakfast cereals as a category have significant scope for reformulation to reduce levels of fat, sugar and/or salt and the evidence is that Kellogg’s competitors, including Nestle, have indeed reformulated their products to reduce levels of sugar.

Mr Simpson - Revenue and Channel Director of the First Claimant - gave estimates of the impact of the 2021 Regulations on Kellogg’s business in terms of its likely loss of profit, which he said was substantial, and its loss of marketable goodwill. However his evidence assumes that Kellogg will not reformulate its products and he does not explain why Kellogg is unable or unwilling to do so. The irony of Mr Simpson’s evidence was also that the adverse impact on Kellogg which he predicted directly correlates with what the Government is trying to achieve. For example, his evidence is that Kellogg makes substantial use of location promotions, and his statistic that 30% of its sales of HFSS products are achieved in this way, is telling. His estimate that 2.5 million kilogrammes of sales will be lost as a result of the restrictions on location promotions, i.e. approximately £5 million of annual profits, shows that the measures will have the effect of reducing sales of breakfast cereals which are HFSS.

Looking at Kellogg’s arguments and evidence more broadly, there was nothing which caused me concern that there may be unfairness to Kellogg or arbitrariness in the effects of the NPM on its business. On the contrary, the public health case for the approach under the 2021 Regulations is compelling and I am quite satisfied that it is both proportionate and rational.

I therefore reject Ground 3B and dismiss the Claim.