[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	England and Wales High Court (Administrative Court) Decisions
You are here: BAILII >> Databases >> England and Wales High Court (Administrative Court) Decisions >> Customs & Excise v Anchor Foods Ltd [1998] EWHC Admin 674 (26 June 1998) URL: http://www.bailii.org/ew/cases/EWHC/Admin/1998/674.html Cite as: [1998] EWHC Admin 674, [1999] V & DR 425

[New search] [Printable RTF version] [Help]

IN THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
CROWN OFFICE

		Royal Courts of Justice Strand, London, WC2A 2LL
		26th June 1998

Mr Justice Dyson:

Introduction

1. This is an appeal by the Commissioners of Customs and Excise ("the Commissioners") against the decision of the VAT Duties Tribunal ("the Tribunal") dated 27 January 1998 that "Spreadable butter" and "Ammix butter" from New Zealand made and imported by Anchor Foods Ltd ("Anchor") are "manufactured directly from milk or cream", and are not "recombined butter", and are therefore subject to a lower rate of tariff duty under Annex I to Council Regulation 1600/95 (as amended).
2. Butter imported into the European Union from New Zealand attracts a preferential rate of customs duty if it satisfies the tariff quota criteria stated in Annex I to Regulation 1600/95 (as amended) , namely :

"Butter, at least six weeks old, of a fat content by weight of not less than 80% but less than 82%, manufactured directly from milk or cream."

3. Annex I identified tariff quota butter under the "CN Code" (Common Nomenclature) as "ex 0405 00 11" , and "ex 0405 00 19". Under Regulation 1600/95, New Zealand is entitled to an annual quota of 76,667 tonnes of butter, meeting the specified description, at a duty of 86.88 ecu per 100 kg, compared with the duty of 260.70 ecu which would otherwise apply.
4. Annex I to Regulation 1600/95 was amended on 27 June 1996 by Regulation 1170/96, to bring the CN Code into line with Regulation 2658/87 as amended by Regulation 2448/95 (which changed the details of the Combined Nomenclature). References in this judgment to "Regulation 1600/95" are to be taken as references to Regulation 1600/95 as amended. The effect of these amendments was that the butter qualifying for the quota tariff in the amended Annex I continued to be identified as 0405 10 11 and 0405 10 19; these sub-divisions were designated as "natural butter" in Regulation 2448/95. Sub-division 0405 10 30 was designated as "recombined butter". Recombined butter does not receive the quota tariff.
5. The issues before the Tribunal were whether, applying Annex I to Regulation 1600/95, Spreadable butter and Ammix butter were:

(i) butter "manufactured directly from milk or cream": it was common ground ( in relation to the other criteria set out in Annex I to Regulation 1600/95 (as amended)) that the test consignments which were the subject of the appeal were butter, of New Zealand origin, over six weeks old, with a fat content of not less than 80%, but less than 82%; and

(ii) natural butter falling within CN Code 0405 10 19, and not recombined butter falling within CN Code 0405 10 30.

There was no dispute before the Tribunal that if (i) and (ii) were satisfied, then the test consignments of Ammix butter and Spreadable butter which are the subject of this appeal were eligible for the tariff quota.

6. In this appeal, the Commissioners contend that the decisions of the Tribunal on these two issues were wrong in law. They say that certain questions raised by the issues should be referred to the Court of Justice of the European Communities ("the ECJ") under Article 177 of the EC Treaty. Anchor does not consent to a reference. For reasons that I explain later in this judgment, I find it necessary to attempt to reach my own conclusions on the two issues raised by this appeal, before deciding whether to make a reference under Article 177.

Butter Manufacturing Processes

7. The traditional method of making butter is by churning. The Fritz process, which is based on churning, has dominated butter-making in Europe for forty years, and is now the most common method world-wide. This system involves the separation of the cream from the milk, crystallisation of the milk fat in the cream, phase inversion to form butter grains and buttermilk, concentration of the milk fat by separating buttermilk and butter grains, and working the grains to butter. A significant feature of all variations of the churning process is that they do not involve dehydration. At no stage in the churning process is the moisture content in the milk fat reduced to 0.2% or 0.1%, so as to produce anhydrous milk fat ("AMF"). There is, however, a point in the process where the product or material being processed has ceased to be milk or cream, and has not yet become butter. At one stage in the Fritz method, the partially processed product was likened by one of the witnesses before the Tribunal to " a bit of a pudding".
8. The first step in the Ammix manufacturing process is the production from fresh milk of skim milk and fresh pasteurised cream with 40% fat. After being held in an intermediate storage for a few hours, the cream is passed through a centrifugal separator, which produces a cream concentrate of 78% fat with buttermilk as a by-product. The concentrate is then subjected to phase inversion and further centrifugal separation, to produce an oil concentrate with 95.5% fat. The moisture content of the concentrate is then reduced to less than 0.2%. The resultant AMF is then cooled, and held briefly in an emulsion tank. The next stage starts with the AMF being mixed with fresh cream and a salt slurry, thereby creating a water in oil emulsion with 80 to 82% fat. The mixture is then subjected to crystallisation, and is then cooled and worked. The worked product is the butter which emerges at the end of the process. Apart from the period of storage at the end of the first step, and any period during which the AMF is held in the emulsion tank for balancing with cream and salt, the manufacturing process is a continuous uninterrupted process. The system is entirely closed, in that there is an entry point for the milk and an exit for the finished butter.
9. The early stages in the process by which Spreadable butter is manufactured are substantially the same as the corresponding stages in the Ammix process. The main difference occurs after AMF has been produced. It is unnecessary for the purposes of this judgment to describe the differences between the Ammix process and the Spreadable butter process, since they are not material to the issues that I have to decide. Suffice it to say that the process for the manufacture of Spreadable butter includes fractionation, which involves a process of filtration to separate the hard and soft fractions. As with Ammix, the manufacture of Spreadable butter is a continuous uninterrupted process automatically controlled.

The History of the Quota and the Legislation

10. The following summary derives in the main from paragraphs 48 to 71 of the Tribunal's decision.
11. It is a matter of public record that, when the accession of the UK to the EEC was negotiated, a major objective of the UK was the safeguarding of the traditional access of New Zealand butter to the UK market. Protocol 18 to the Accession Treaty in 1972 authorised the UK, as a transitional arrangement, to import specified quantities of butter in the five years to 1977. Article 5.2 provided that appropriate measures to ensure the maintenance, after 31 December 1977, of "exceptional arrangements" in respect of imports of butter from New Zealand were to be determined by the Council acting unanimously on a proposal from the Commission. Protocol 18 was implemented by Council Regulation 226/73. Thereafter, other Council regulations were made, providing short term quotas in the periods to March 1981.
12. On 1 April 1981, Council Regulation 858/81 was adopted, providing for the period to 31 December 1983. Its recitals included the following:

"Whereas in order to contribute, in accordance with the common interests to the harmonious development of world trade, arrangements should be made to permit New Zealand to continue to export butter to the United Kingdom on special terms; Whereas the new arrangements must include measures to prevent the balance of the United Kingdom butter market from being disturbed; whereas therefore the quantities of New Zealand butter imported under these arrangements must be fixed on a degressive scale."

Article 4 provided:

"4. Entry under the special import arrangements shall be conditional upon the presentation of a certificate showing that the butter concerned:

- is of New Zealand origin,

- is at least 6 weeks old,

- has a fat content by weight of at least 80% but less than 82%, and

- has been manufactured directly from milk or milk cream".

13. In respect of the period between 1 January 1984 and 30 June 1995, Council Regulations were introduced from time to time, all of which included a provision in the same terms as Article 4 of Council Regulation 858/81. Accordingly, throughout the period between 1 April 1981 and 30 June 1995, in order to meet the quota criteria, New Zealand butter had to satisfy the description set out in Article 4 of Council Regulation 858/81.
14. Meanwhile, between 1986 and 1994, the Uruguay "Round" of GATT negotiations had been taking place. These culminated in the Final Act, which was signed in Marrakesh, Morocco on 15 April 1994. The Agreement on Agriculture in the Final Act (the "Marrakesh Agreement") recited that Members were "committed to achieving specific binding commitments in each of the following areas: market access .......".

Article 4.1 provided:

"1. Market access concessions contained in Schedules relate to bindings and reductions of tariffs, and to other market access commitments as specified therein".

15. Schedule CXL - European Communities covers Most-Favoured-Nation Tariff Quotas for agricultural products in Section 1B of Part One. Under Current Access Quotas there appears New Zealand butter. Under " other terms and conditions" appears this description:

"Butter of New Zealand origin

- at least six weeks old

- with a fat content of not less than 82% by weight

- manufactured directly from milk or cream

Qualification for the quota is subject to conditions laid down in the relevant Community provisions".

16. The Uruguay "Round" agreement was implemented by Regulation 1660/95. Its recitals include the following:

"Whereas the Agreement on Agriculture concluded during the Uruguay Round multilateral trade negotiations (hereinafter referred to as the "Agreement") provides for certain tariff quotas for milk and milk products under the "current access" and " minimum access" arrangements; whereas those quotas should be open for an initial annual period ending on 30 June 1996; whereas rules should be laid down for the management of those quotas;

.......

whereas certain special conditions previously applied to imports authorized under special arrangements should be applied to imports of New Zealand butter under the quota provided for in the Agreement in order to monitor their origin and destination".

17. Article 6 provides that the tariff quotas should be as set out in Annex 1. Annex 1 is headed "Tariff Quotas Pursuant to the GATT/WTO Agreements specified by country of origin". Under "CN Code" appear "ex 0405 00 11, ex 0405 00 19". Under "description" appears:

"Butter, at least six weeks old, of a fat content by weight of not less than 80% but less than 82%, manufactured directly from milk or cream".

18. The quota was set at 76,667 tonnes, as agreed in the Marrakesh Agreement, based on average supplies between 1986 and 1988. Regulation 1600/95 contained the first reference in the community quota arrangements to a CN Code. Under Regulation 1359/95, which amended the community tariff from 1 July 1995, the two CN Code sub-divisions stated in Regulation 1600/95 covered "butter and other fats and oils derived from milk: of a fat content, by weight, not exceeding 85%". It is of note that no distinction was drawn in this Regulation between natural and recombined butter.
19. Regulation 2448/95 (which took effect from 1 January 1996) recited that it was necessary to amend the combined nomenclature, to take account of changes in the harmonised system nomenclature, pursuant to the Recommendation of 6 July 1993 of the Customs Co-operation Council, and to changes in requirements relating to statistics and commercial policy. The regulation designated sub-divisions 0405 10 11 and 0405 10 19 as "natural butter", and introduced (inter alia) sub-division 0405 10 30 "recombined butter". At the foot of the page on which the CN Codes for butter appear in Regulation 2448/95 is a note "WTO quota: see Annex 7". Annex 7 is a schedule, item number 28 of which relates to New Zealand butter. It shows " 0405 00 11, 0405 00 19" under the column headed "CN Code". Under the column headed "description" appear the words:

"butter, at least six weeks old, of a fat content by weight of not less than 80% but less than 82%, manufactured directly from milk or cream".

20. The remaining columns give the quota quantity, rate of duty and other applicable terms and conditions.
21. As I have already said, Annex 1 to Regulation 1600/95 was amended (with effect from 1 July 1996) by Regulation 1170/96 to bring the CN Code into line with Regulation 2448/95 as amended.

The First Issue: was the butter directly manufactured from milk or cream?

The correct approach to interpretation

22. The Tribunal started from the principle laid down by the Court of Justice in EC Commission v Germany [1996] ECR 1-3989, at paragraph 52 on page 4021:

"the primacy of international agreements concluded by the Community over provisions of secondary Community legislation means that such provisions must, so far as is possible, be interpreted in a manner that is consistent with those agreements".

23. The Tribunal said that, since the Marrakesh Agreement was based on current access, it was necessary to consider the legal basis of that access prior to that agreement, namely the successive Council regulations enacted pursuant to Article 5.2 of the Accession Treaty. At paragraph 93 the Tribunal said:

"93. Since the agreement was based on the "current access" principle it seems that the correct interpretation in Community Law of Reg. 858/81 is a relevant factor in interpreting the agreement but that it is not determinative. The agreement was an international multi-lateral agreement and cannot be assumed to have adopted the interpretation of the domestic courts of any party. Indeed we understand that the Marrakesh Agreement includes a disputes procedure. Considerable caution must therefore be exercised in relation to travaux preparatories for the 1981 regulation".

24. At paragraph 94, the Tribunal said that the travaux preparatories did not assist. It was, therefore, not possible to impute any intention beyond what appeared from the words used in Regulation 858/81. This led them to say:

"In our judgment we are thrown back to the natural interpretation of the actual words used. We do not consider that Reg. 858/81 had any special interpretation which governed current access and thereby the Marrakesh Agreement and Reg. 1600/95".

25. The Tribunal were invited by counsel to decide the meaning of the words "directly manufactured" by applying certain principles or alleged principles of Community law. Thus, on behalf of the Commissioners, Mr Lasok submitted to the Tribunal (as he does to me) that the tariff is an exception or derogation, and that the words should therefore be construed strictly or narrowly. The Tribunal refused to do so not because they did not consider that the tariff is an exception or derogation, nor because they thought that there was no such principle of Community law; rather because they did not consider the principle to be relevant in the context of a regulation implementing an international agreement. Mr Pannick submitted to the Tribunal (as he does to me) that the Commissioners' approach to the meaning of the words offended certain basic principles of Community law, for example proportionality and certainty. The Tribunal did not "derive much assistance" from these basic principles, although they said that the agreement should be construed so as to reduce uncertainty (paragraph 96).
26. They arrived at what they considered to be the natural interpretation of the actual words used in the passage of the decision which appears at paragraphs 97 to 99, which I need to set out in full:

97. Having heard extensive evidence and argument about the manufacturing processes used, which is available on transcript in the event of an appeal, and having considered the documentary evidence to which we were referred, we have concluded that on a normal use of language the butter in the consignments was manufactured directly from milk. It seems to us that, given that the milk was subjected to an uninterrupted process at the end of which butter was produced without the introduction of any further raw material except salt, the fact that during that process a product described as anhydrous milk fat was produced all (or, in the case of Spreadable Butter, part) of which formed part of the final butter does not mean that the butter was not manufactured directly from milk. We do not consider that the words "manufactured directly" require that at one moment there is milk or cream and that at the next it is butter, with a direct transformation. If that was the test it seems to us that butter made by the traditional Fritz process would not qualify; it is to be noted that the Fritz process takes much longer than the Ammix and that for a considerable time the substance being processed is neither milk, nor cream nor butter.

98. The submissions of Mr Lasok seem to us to involve disregarding part of what was a continuous manufacturing process, excluding the entire part which resulted in the production of AMF. We do not consider that to be a legitimate approach. No doubt butter with similarities to Ammix or Spreadable Butter could be produced by a process starting with AMF. It seems to us to be irrelevant that AMF produced elsewhere or earlier could theoretically have been used. The fact is that it was not; furthermore it would have involved considerable adaptation of the plants. At paragraph 58 of his statement, Mr Arph said this,

"You cannot have an intermediate product in an uninterrupted manufacturing process."

We see considerable force in this observation. Put in another way, we do not consider that it is appropriate when considering whether a continuous manufacturing process is direct to examine the state of the materials in the middle of that process.

99. Although English is only one of 15 Community languages, it is one of the two official languages of the WTO. In those circumstances the New Shorter Oxford English Dictionary is helpful. Under "directly" there appears (inter alia),

"1. In a straight line; without deviation; 2. ...3. Completely, absolutely, exactly ...4. Without an intermediary; by a direct process ...5. ...6. At once, immediately; ...".

It seems to us that all of these meanings are consistent with Mr Pannick's submissions. We find it difficult to conceive manufacturing processes more direct that those in this case".

27. The first question that I need to address is whether, in interpreting the words "manufactured directly", I should apply the principles of Community law. The correctness of the principle enunciated in EC Commission v Germany [1996] ECR 1-3989 at paragraph 52 is not in issue before me. In the present case, however, the relevant regulation (1600/95) was implemented in order to give effect to an international treaty (the Marrakesh Agreement), the parties to which intended to maintain "current access", thereby providing for the continuation of quotas for milk products. Mr Pannick rightly concedes that "current access" refers to de jure access, and not de facto access. Thus, the current access which the treaty was designed to preserve was, in the case of New Zealand butter, access permitted by the applicable Community rules. I see no basis for determining what access was allowed by these rules other than by reference to Community law. It follows in my view that, if any of the principles of Community law are relevant to the question of what "directly manufactured" means in Regulation 858/81 and its successor regulations, they must be applied. I do not regard this conclusion as doing violence to the principle stated in EC Commission v Germany, or as being inconsistent with Article 31.1 of the Vienna Convention on the Law of Treaties, which provides that:

"A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose".

28. It is necessary to apply Community law in order to determine what was the extent of the permitted current access for New Zealand butter, which the parties to the Marrakesh Agreement agreed should be preserved.
29. I turn, therefore, to consider the principles of Community law which it is submitted shed light on the meaning of "directly manufactured" as it appears in Regulation 858/81 and the succeeding regulations made prior to the Marrakesh Agreement. I start with Mr Lasok's derogation/exception point. I accept his submission that the tariff quota is an exception or derogation from the normal rate of duty provided for in what is now Annex 1 to Regulation 2658/87 as amended. Mr Pannick submits that it is not an exception or derogation, because Community Law contains a specific, substantive rule that provides for a specific tariff rate for a set quantity of New Zealand butter meeting set criteria. He relies on the fact that the entitlement to a preferential tariff quota has been a well-established part of Community law since the UK's accession to the Community, and submits that the authorities relied on by Mr Lasok concerned a "general" exception, from which a large number of persons may benefit, not (as here), a specific provision creating a tariff quota to benefit one other party.
30. It seems to me that the fact that the preferential tariff has existed for a long time is irrelevant to the question whether it is an exception to the normal rate of duty payable in respect of butter imported by third countries outside the Community. If an exception could cease to be an exception merely by the passage of time, and for no other reason, this would give rise to real uncertainty. When would the exception lose its status as an exception? Nor do I see how the number of persons who do (or can?) benefit from the provision is relevant in determining whether it is an exception or not.
31. Mr Lasok relies on Ethicon GmbH v Hauptzollant Itzehoe [1986] ECR 1131 (a customs tariff case), in particular the opinion of Advocate General Mischo at 1135, and the judgment at page 1145 where at paragraph 13, the court said:

"In the view of the Council and the Commission, such exceptions to the Common Customs Tariff must be interpreted strictly. To interpret those provisions contrary to their clear wording in the light of a subsequent amendment of the regulations is contrary to the requirements of legal certainty".

32. I accept, therefore, that if there is any relevant ambiguity in the words "directly manufactured", then they should be construed strictly, and against the party seeking to invoke the preferential tariff quota. If, however, the meaning of the words is clear, they should be given their full effect.
33. I shall postpone a consideration of Mr Pannick's arguments that the Commissioners' approach offends a number of Community law principles until I have examined the meaning of the words "directly manufactured" in detail.

"Directly Manufactured"

34. Mr Lasok submits that the Ammix butter and Spreadable butter are not directly manufactured from milk or cream because (i) the raw material is not made into butter without first having been made into something else, namely AMF, and/or (ii) the manufacture of the butter involves a "detour", since butter containing approximately 80% fat could be made by excluding from the manufacturing process the stage in which the emulsion with a fat content of 78% is converted into AMF, before it is converted into an emulsion with a fat content of 80-82%.
35. Before the Tribunal, it seems that Mr Lasok concentrated largely, if not entirely, on the first point. No doubt, this explains why there is no discussion of the second point in the decision. I shall start with the first point. Mr Lasok submits that the Ammix and fractionation processes necessarily involve the production of a different product (AMF) before butter can be produced. The AMF is not a by-product of the process, but the product from which the butter is manufactured. Accordingly, the butter is not directly manufactured from milk; it is directly manufactured from AMF. As we have seen, the Tribunal's view was that "manufactured directly" means manufactured by "an uninterrupted process" or by a "continuous process". Mr Lasok argues that, although the phrase "directly manufactured from milk or cream" is capable of being read as referring to an uninterrupted process, the word "directly" is not necessary in order to convey the idea of an interrupted process. The addition of the word "directly" introduces a requirement that is different from the requirement that the process be uninterrupted. It refers to what happens in the course of the process.
36. Mr Pannick supports the reasoning and conclusion of the Tribunal. There was a good deal of evidence before the Tribunal from experts eminent in the field of butter manufacture. It was agreed, however, that the phrase "directly manufactured" is not a term of art, or a technical expression which has a clearly understood meaning throughout the industry. In my view, the legitimate role of the experts, in relation to the directly manufactured issue, was to explain to the Tribunal the manufacturing processes, and to give evidence about the state of the materials during the processes, and technical matters of that kind. I do not, therefore, consider that the Tribunal's decision can be supported solely because it accords with the opinion of Anchor's experts as to the meaning of the words.
37. In my judgment, it was a necessary, but not sufficient condition of direct manufacture from milk or cream that the butter be manufactured by a continuous and uninterrupted process. Thus, if the process were interrupted at the AMF stage, and the AMF taken out of the manufacturing unit, placed into store, and then later used in order to complete the process, the butter emerging at the end would not have been "directly" manufactured from the milk with which the process started. I believe that this much would be common ground. The concept of direct manufacture, therefore, clearly connotes temporal continuity. The other essential element, in my view, is that the butter be manufactured without deviation or detour, i.e. without adding unnecessary steps to the manufacturing process. In fact, it is most unlikely that a manufacturer of any product would make an unnecessary detour in the manufacturing process, since, ex hypothesi, there would be no point in doing so, and it would be likely to increase the cost of manufacture needlessly.
38. I cannot accept Mr Lasok's argument that it is a necessary condition of direct manufacture that the butter should be transformed from milk or cream without the intervening creation of an identifiable product. As the Tribunal pointed out, the words "manufactured directly" do not require that at one moment there is milk or cream, and that at the next it is butter, with a direct and immediate transformation from one to the other; if that were the test, churned butter made by the traditional Fritz process would not qualify, since in that process, for a considerable period of time there is a pudding-like substance which is neither milk, cream nor butter. That substance is a product. There is no suggestion in the material that I have seen that it is marketable. On the other hand, it is true that AMF is a marketable product. But Mr Lasok does not base his argument on the fact that AMF is marketable. No doubt this is because the evidence before the Tribunal was that it would be "commercially unrealistic" to buy in AMF in order to manufacture Ammix butter or Spreadable butter. The fact is that neither of these butters is made in this way; the processing plants would require considerable adaptation if they were to be manufactured in this manner. These butters are made by a process which starts with milk, and proceeds continuously via AMF to butter. Since the marketability of the intermediate product is not relevant to the question whether its existence prevents the manufacture from being direct from milk, it is difficult to see why the AMF makes the manufacture indirect in the two processes with which this appeal is concerned, and yet the pudding-like substance does not have that effect in the traditional Fritz process. The AMF and the pudding-like substance share the common features that they are both (i) products created from milk or cream and from which butter is produced, and (ii) part of a continuous process which leads from milk or cream to butter.
39. In my view, the mere fact that AMF is created during the process is not a reason for deciding that the butter is not directly manufactured from milk or cream. I agree with the reasoning and conclusions of the Tribunal. I also think that they were correct in saying that it was irrelevant that AMF produced earlier, or even elsewhere, could have been used to manufacture butter by the two processes under consideration. As they said, the fact is that it was not, and if this had been done, it would have required considerable adaptation of the processing plants.
40. I can deal briefly with Mr Lasok's "detour" point. As Mr Pannick points out, in order to determine whether there has been a detour, it is first necessary to identify the destination. In this case, the destination is not butter in the general sense, but the specific butter in question. It would be very surprising if it were otherwise, since implicit in the fact that the word "directly" appears in the description of butter which qualifies for the special tariff is an acknowledgement that some butters are directly manufactured from milk or cream, and others are not. Moreover, Article 4 of Regulation 858/81 (and all the succeeding regulations made prior to the Marrakesh Agreement) referred to "a certificate showing that the butter concerned ...... has been manufactured directly from milk or cream" (my emphasis). Thus, the question is whether the manufacture of the butters concerned in this appeal involved a detour. They clearly did not. Each of the steps in the processes was a necessary stage in the manufacture of these particular butters. Accordingly, although the Tribunal did not deal with the detour point, I am satisfied that it does not cast doubt on the correctness of their decision.
41. I return to the principles of Community law. In my view, the Tribunal were clearly correct in deciding that these butters are directly manufactured from milk or cream. I do not consider that there is any ambiguity in the word "directly", and do not, therefore, consider that it is appropriate to apply a strict or narrow interpretation of the word. It is a plain and ordinary word, whose meaning is clear. In the result, it is unnecessary for me to consider Mr Pannick's arguments that the Commissioners' approach offends a number of Community law principles. Accordingly, I determine the first issue in favour of Anchor.

The Second Issue: Was the butter recombined?

42. Mr Lasok submits that, even if the butter was directly manufactured from milk or cream, it did not qualify for the quota because it was "recombined" within the meaning of the tariff sub-division 0405 10 30 in Regulation 2448/95. The phrase "recombined butter" is not defined in any of the relevant EC regulations. The Tribunal heard a substantial amount of evidence as to the meaning of the phrase. At paragraph 46 of their decision, they said:

"46. The most common use of reconstituted or recombined products is supplying milk products to countries with no indigenous milk supply or an insufficient supply, especially tropical countries. Powdered milk is reconstituted by the addition of water. AMF is recombined to produce milk fat. Within the dairy industry the phrase "recombined butter" is used to describe butter produced using a recombined product or products although we observe that it is not in fact the butter which is re-established or recombined but the intermediate product, for example powdered milk. Normally the process in relation to butter will start by recombining a stored product which has been dehydrated at another time and place. It is not usual within the dairy industry to describe butter produced in a single process from fresh milk or cream and not from a stored product as "recombined".".

43. Two conditions which the Tribunal accepted were important features of recombined butter were first that it involved the use of at least one main ingredient which was stored, and secondly that recombination took place separated in time and/or place from where at least one of the two main ingredients was produced. The Tribunal were also referred to a number of publications which were relied on by the Commissioners as showing that there was an internationally accepted definition of recombined butter, which differed materially from the definition which Anchor's witnesses said was the one understood by those in the dairy industry. One of these was the Codex Alimentarius published by the Joint Commission of the Food and Agriculture Organisation of the World Health Organisation in 1984. This included the following decision of the Committee of Government Experts on the Code:

"Decision No.5

The Committee decided that standards adopted under the Code should apply to products so defined, whether made from milk, reconstituted milk or recombined milk or by reconstitution or recombining milk constituents unless the provisions of the standards provide otherwise. 'Reconstituted (product)' is the milk product resulting from the addition of water to the dried or condensed form of (product) in the amount necessary to re-establish the specified water solids ratio. 'Recombined (product)' is the milk product resulting from the combining of milk fat and milk-solids-non-fat in one or more of their various forms with or without water. This combination must be made so as to re-establish the product's specified fat to solids-non-fat ratio and solids to water ratio."

The Tribunal's conclusion on this issue appears at paragraphs 102-104 in these terms:

" 102. We can deal with the "recombined" argument shortly. If the consignments had been entered between 1 July 1995 and the amendment to the CN Code, there could have been no argument based on whether or not the butter was recombined. Unless we are driven to the conclusion that the alteration in the CN Code altered the quota criterion in a way not envisaged in the Marrakesh Agreement, we should not so interpret the amendment to Regulation 1600/95.

103. In fact the term "recombined" is used in a confusing sense. The Codex referred to "re-establish the product's specified fat to solids-non-fat ratio and solids to water ratio" (paragraph 45 above). Since there was here no butter before the recombination process on which Mr Lasok relies it does not fall within the literal meaning of the Codex definition. There was no evidence as to whether butter itself can be recombined as opposed to its constituents. It would appear that "recombined" as used in the CN Code is a transferred epithet referring not to the butter but to the product from which it is made. We accept Mr Arph's evidence that the normal use of the term is in relation to products manufactured from stored products. We would be very slow to accept Mr Lasok's submission that there is some international usage which conflicts with trade usage.

104. We hold that the introduction of a sub-division of the CN Code for recombined butter is consistent with the quota criteria the Marrakesh Agreement because if the butter is manufactured directly from milk or cream it is not recombined."

44. Mr Lasok makes a number of criticisms of this. It will be convenient if I start with his submission that they should not have accepted the evidence of Anchor's witnesses as to the meaning of recombined butter. He argues that the Tribunal was not entitled to take into account the evidence of Mr Arph as to the meaning of recombined butter, since his evidence was as to the trade usage, which the Tribunal could not take into account for the purpose of tariff classification. He relies on Ludwig Post GmbH v Oberfinanzdirektion Munchen [1991] ECR 1-2391 at paragraph 24, where the court said:

" Similarly, neither the alleged usage nor any divergent application of the rules in certain Member States can influence the interpretation of the CCT which is based on the wording of the tariff headings."

45. That was a case concerning the tariff classification of a product defined as "75% whey protein concentrate". It contained 5% lactose. The evidence was that the product whey "normally" contained 70% lactose. The court held that, by reason of the low lactose content, the product in question "no longer displays the essential characteristics of the basic product whey". Post relied on the terminology allegedly used in trade circles to argue that the product should be classified as whey, and this argument was rejected in paragraph 24 of the judgment.
46. It will be seen, therefore, that in Post the court was able to reach a clear conclusion as to the meaning of the tariff heading without reference to any facts or evidence other than that whey normally contained 70% lactose. The question for the court (applying the principles stated in Weber [1989] ECR 1395) was whether the essential constituents of the product under consideration differed substantially from those of the basic product. It seems that the court was able to determine this issue without the benefit of expert evidence. It declared that it was immaterial whether a product like the one at issue was regarded in trade circles as whey or not. The context of paragraph 24 of the decision is important. What the court was saying was that where (as in that case) the meaning of a tariff sub-division is clear, it cannot be influenced by a different meaning given to it by trade usage.
47. In the present case, the task facing the Tribunal was rather different from that which the court was confronted in Post. Here there was no agreed starting point or yardstick of what is "normal". The Tribunal had to determine what "recombined butter" means in the tariff sub-division. The tariff itself gives no help. In those circumstances, it is difficult to see why the opinion of experts in the dairy industry as to its meaning should be inadmissible. In my view, the Tribunal was entitled to hear evidence from experts as to what they thought the word means, the interpretation of references to "recombined" in various international publications, as well as how the word is used and understood in the industry. The Tribunal was entitled to take all this into account in determining the true meaning of the word.
48. Mr Lasok submits that the Tribunal was bound to find that the word bears what he describes as the "internationally accepted meaning". By this he means the meaning given in the Codex Alimentarius to which I have already referred. There are two difficulties with this. First, put at its highest from the Commissioners' point of view, Post is authority for the proposition that trade usage which conflicts with the true meaning of a tariff sub-division cannot be invoked to give it a different meaning; it is not authority for the proposition that a definition contained in an international document like the Codex must determine the meaning.
49. Secondly, I do not consider that the evidence before the Tribunal did clearly show that there was an internationally accepted meaning, or that, if there was, it was clearly inconsistent with the trade usage contended for by Anchor. A number of documents were put by Mr Lasok to Mr Arph during his cross examination: see transcript pages 333 to 339. Mr Arph was an expert witness whom the Tribunal described as "wholly independent" and "highly experienced". They said that they preferred his evidence insofar as he differed from Dr Rajah (the Commissioners' expert). Mr Arph agreed that there was no express reference in any of these documents to the 2 features of recombined butter to which I have referred, although he said that there was an implicit reference in the most recent document, the "Dairy processing handbook"(1995). Mr Arph said of this at page 338E:

"Q .... My point is in this attempt at a definition we once again do not see the two features that you say are fundamental to the concept of recombination? A. If you take it out of context like this, then I agree with you, it can be difficult to see that. But, for people who read this book and who are familiar with -- I mean, this is a text book more or less for the dairy industry. I understand your problem. You do not have that familiarity with the terminology and the whole situation. But, for dairy people, this is completely clear, in my opinion, that it refers to products which are stored or a commodity on the world market or whatever. The other definitions you have referred to is, of course, undeniable. I am not opposing that.".

He then went on to make a more general observation at page 338G:

"If I may, however, add to that, to justify a little bit my position on this point, which I understand you find peculiar and difficult? It is that in the industry this is really the accepted practice and I think it should be known to the Tribunal also that the International Dairy Federation, which is the body from which most of these regulations emanate, it is always worked out there in the beginning and they are referenced to the Commission in Brussels, they have recently taken up this definition on a programme for re-examination because they find it too -- it is not clear enough what we mean. In due course probably there will be a definition which clearly states what I have said all the time in my statement."

50. What I take him to have been saying to the Tribunal was that the definitions in the documents put to him by Mr Lasok were not clear. They required to be interpreted by those in the Industry who understood the meaning of technical words like "recombined". On a fair interpretation of his evidence, he was not saying that there was a conflict between the meaning of "recombined" in these documents and its meaning as understood by the Industry. He was saying that the meaning in the documents needed to be clarified. That is why he told the Tribunal that the meaning had been taken up with the Commission in Brussels: note his evidence that "it is not clear enough". The Tribunal did not understand the evidence to be that there was a clear international usage which conflicted with trade usage (paragraph 103 of the decision). In my judgment, they were fully entitled to take that view of the evidence.
51. It follows that the foundation of Mr Lasok's attack on the Tribunal's decision on the second issue falls away. For the reasons given earlier, I rather doubt whether Post prevented the Tribunal from having regard to the meaning accorded to "recombined" by those in the Industry. But even if I am wrong as to that, the Tribunal decided (as they were entitled to do) that there was no conflict between international usage and trade usage. Accordingly, there was no infringement of the principle in Post even if it applies.
52. Mr Lasok's other main criticism of the Tribunal's approach on this issue is that they seem to have believed in relation to this part of the case that they were still engaged in the interpretation of the tariff quota set out in Regulation 1600/95, which they believed had to be read subject to the Marrakesh Agreement. The question for the Tribunal was one of the true interpretation of a tariff sub-division in a completely different regulation, namely Regulation 2658/87 as amended, which did not fall to be construed by reference to Regulation 1600/95 or the Marrakesh Agreement. If the Tribunal were faced in a subsequent case with a consignment of butter manufactured by the Ammix or fractionation process, but imported from a country not benefiting from a tariff quota under the Marrakesh Agreement, how, asks Mr Lasok, would the Tribunal classify the butter? If they allowed themselves to be influenced by the existence of the New Zealand tariff quota, they would be using an irrelevant criterion to determine the true meaning of the tariff. If they ignored the New Zealand tariff quota and concluded that the butter was recombined and fell within sub-division 0405 10 30, that would produce an irreconcilable conflict with their approach in the present case.
53. I see considerable force in this argument. Mr Pannick seeks to counter it by saying that, since the aim of Annex 1 to Regulation 1600/95 was to implement the Marrakesh agreement, it cannot have been intended to introduce a separate criterion (ie that recombined butter be excluded) which was not part of the Marrakesh Agreement. He relies on the principle stated in The Commission v Germany. Consistently with the Marrakesh Agreement, Annex 1 of Regulation 1600/95 did not suggest any distinction between natural and recombined butter, save for any that was already implicit in the substantive criterion "manufactured directly from milk or cream". If the amended Annex 1 were intended to introduce a new substantive criterion, it is highly unlikely that this would have been done by means of a tariff classification, especially when Annex 1 itself already states the substantive criteria for eligibility for the tariff quota. Mr Pannick submits that the obvious intention of the amended classification was simply to bring the CN reference into line with the substantive criteria. Thus, because recombined butter is not manufactured directly from milk or cream (because the butter is not manufactured in a continuous uninterrupted process), the CN Code excludes recombined butter from the scope of the butter eligible for the tariff quota.
54. I readily accept that it is unlikely that, if it were intended to amend the substantive criteria stated in Annex 1 of Regulation 1600/95, this should have been done by means of a tariff classification. It is also clear that, if the effect of the amendment of the tariff classification is to make Annex 1 of Regulation 1600/95 not conform to the current access which the Marrakesh Agreement was intended to preserve, then there may be a question as to the validity of the amended Annex. The difficulty that I have with Mr Pannick's arguments is that they do not seem to me to meet the main thrust of Mr Lasok's argument that the meaning of the tariff classification cannot be determined by reference to Regulation 1600/95 or the Marrakesh Agreement, since the meaning of the classification cannot vary according to whether the butter is imported from New Zealand or some other country.
55. Although I am inclined to accept this part of Mr Lasok's argument, it cannot avail him to impugn the Tribunal's decision on the second issue. This is because, although the Tribunal may have been in error in the approach set out at paragraph 102 of the decision, it is quite clear that they decided the meaning of "recombined" on the basis of the evidence of Mr Arph as to how the word was understood by the Industry, and their finding that this was not in conflict with any international usage. They did not decide the issue by applying some principle that required them to give a special meaning to the word (different from its meaning as normally understood in the Industry), in order to arrive at a meaning which did not involve any alteration to the quota criteria stated in Annex 1 to Regulation 1600/95.
56. In the result, I decide the second issue in favour of Anchor.

The Third Issue: Should there be a reference under Article 177?

57. I am asked by the Commissioners to make a reference to the ECJ of questions inter-alia as to the meaning of (i) the words "manufactured directly from milk or cream" in Annex 1 to Regulation 1600/95, and (ii) the phrase "recombined butter" in tariff sub-division 0405 10 30 in Annex 1 to Regulation 2658/87 as amended.
58. Since I am not sitting as a national court of last resort, I have to apply the Article 177(2) which provides:

"Where such a question is raised before any court or tribunal of a Member State, that court or tribunal may, if it considers that a decision on the question is necessary to enable it to give judgment, request the Court of Justice to give a ruling thereon".

"Any such question" includes questions as to the interpretation of Council or Commission Regulations.

59. Mr Pannick submits that I should not refer either of the 2 questions for the following reasons:

(i) The questions are not ones of interpretation of law at all, but rather of the application of law to the facts. Accordingly, no relevant questions are raised.

(ii) Even if the questions are ones of interpretation of law, a reference is not "necessary", since the points of law are clear and can be determined by me with "complete confidence".

(iii) Even if a reference is "necessary" in the sense that I am unable to decide the points with complete confidence, I should exercise my discretion against making a reference because:

(a) there would be a delay of at least 18 months before a ruling could be obtained from the ECJ; this would cause great hardship to Anchor;

(b) the ECJ is already overloaded with references, and national courts (at any rate other than courts of last resort) should exercise restraint in applying Article 177;

(c) the points are not ones of especial difficulty, nor are they of particular importanceto the Community as a whole, as distinct from the parties to this dispute;

(d) I should be slow to make a reference against the wishes of one of the parties.

60. I start with the question whether the issues raised are ones of interpretation or application. In my view, Mr Lasok is right when he submits that they are ones of interpretation. The question raised by the first issue is what the words "manufactured directly" mean in their context in Annex 1 of Regulation 1600/95. Once that meaning has been determined, it is applied to the facts of the case. What is in dispute in this case is the meaning of the phrase, and not whether, in the light of the meaning ascribed to the words, the butter does or does not meet the ascribed meaning. I bear in mind the cautionary words of Advocate-General Jacobs in Wiener S.I. GmbH v Hauptzollamt Case C-338/95, where at paragraph 17 of his Opinion, he said:

"I do not consider that it is appropriate, or indeed possible, for the Court to continue to respond fully to all references which, through the creativity of lawyers and judges, are couched in terms of interpretation, even though the reference might in a particular case be better characterised as concerning the application of the law rather than its interpretation".

61. I acknowledge that there are cases where (in English law) it has been held that the meaning of an "ordinary word of the English language" is a question of fact, and involves no question of law: see for example, Cozens v Brutus [1973] AC 854. But as Staughton LJ said in Her Majesty's Treasury v Lane [unreported decision of 19 March 1993], for that doctrine to apply, it must first be established that the words in question are used in their ordinary English sense, and that the context is of no assistance in deciding what they mean; otherwise, the problem is one of interpretation, and thus a question of law. It is true that the word "directly" is an ordinary English word, but the issue in this dispute is what it means in the context of the manufacture of butter. When I turn to "recombined butter", I rather doubt whether the meaning of that phrase can sensibly be determined simply as a matter of ordinary English language. The matter is even more open to doubt if the EC dimension is added.
62. The next question is whether a reference is "necessary". I have heard submissions as to the applicability of the doctrine of acte clair. I gratefully adopt the observations of David Anderson at page 127 of his book, References to the European Court. The doctrine applies when:

"....the correct application of Community law may be so obvious as to leave no scope for any reasonable doubt as to the manner in which the question raised is to be resolved. Before it comes to the conclusion that such is the case, the national court or tribunal must be convinced that the matter is equally obvious to the courts of the other Member States and to the Court of Justice. Only if these conditions are satisfied, may the national court or tribunal refrain from submitting the question to the Court of Justice and take upon itself the responsibility for resolving it:" (see Srl CILFIT v Ministry of Health [1982] ECR 3415, 3430, para 1).

63. As Mr Anderson points out, there is no indication in the CILFIT judgment that the doctrine of acte clair has any application other than to references by a court of last instance under Article 177(3). English and Scottish courts have sometimes used the phrase "acte clair" to express the degree of confidence which they feel about the correct answer to the question. I agree that it is preferable for a court falling outside Article 177(3) not to use this phrase. In my view, the test of "necessity" is satisfied unless I can resolve the questions with "complete confidence": see per Sir Thomas Bingham MR in R v Stock Exchange, ex parte Else [1993] QB 534, 545E. The converse of "complete confidence" is "real doubt" (page 545F). Mr Lasok has presented his arguments on issues 1 and 2 with great vigour, but after careful consideration, I have reached my conclusions on both points without any real doubt. I find, therefore, that the condition of necessity is not made out in relation to either issue.
64. In case I am wrong as to necessity, I turn to the question whether I ought in any event to exercise my discretion to make a reference. Mr Pannick relies on Bulmer Ltd v Bollinger SA [1974] 1Ch 401,422D-425C, where Lord Denning MR set out the guidelines for the exercise of the discretion. The length of time which may elapse before a ruling can be obtained from the ECJ is undoubtedly an important consideration: see Bulmer, p423G. It is not in dispute that it will take at least 18 months, and possibly considerably longer. It seems that this is because the court is overloaded with work. Fear of possible collapse under its caseload moved Advocate General Jacobs in Wiener to ask national courts to exercise restraint. The consequences for Anchor of a delay in the order of 18 to 24 months in obtaining a ruling from the ECJ are very serious indeed. They are described in some detail in affidavits sworn on 29 April 1998 by Mr Milne, and by Mr Larsen on 10 June 1998. It is unnecessary to go into these affidavits; what they say has not been challenged by the Commissioners. Mr Lasok does not contend that the delay that is likely to be occasioned by a reference under Article 177 will be other than very serious for Anchor.
65. It is true that in Bulmer, one of the factors listed by Lord Denning as relevant to the exercise of discretion was "Do not overload the court". I rather doubt whether, as a discrete point, this is a legitimate consideration for refusing a reference. It seems to me that the better view may be that a reluctance to overload the court is the rationale for some of the other considerations to which a national court should have regard, when deciding how to exercise its discretion. I do not, therefore, propose to take this concern into account as a discrete consideration.
66. In Bulmer, Lord Denning did say (p.424F) that unless the point "is really difficult and important, it would seem better for the English judge to decide it himself". In my judgment, this is plainly a relevant factor, and one which militates against a reference in this case. The issues raised by this appeal are obviously of importance to the parties, but I also have to take into account their importance generally throughout the Member States. It seems to me that the issues are primarily of concern to the parties to the appeal, although I accept that the case is of interest to Anchor's competitors in the manufacture of products similar to Ammix butter and Spreadable butter. This case, however, has less importance than, for example, a case concerning the interpretation of legislation whose aim is to achieve common or harmonised practices in all the Member States.
67. Lord Denning also said in Bulmer (p.425B) that the English court should hesitate before making a reference against the wishes of one of the parties. I think it right to give a little (but only a little) weight to the fact that Anchor opposes a reference.
68. Weighing these various considerations, even if I had concluded that a reference was "necessary", I would not have exercised my discretion to refer the questions to the ECJ. In my view, the combined effect of the points relied on by Mr Pannick would tip the scales heavily against a reference.

Conclusion

69. In the event, for the reasons that I have given, Anchor succeeds on the first and second issues, and I refuse to make a reference under Article 177. The appeal is, therefore, dismissed.
70. MR JUSTICE DYSON: The parties have a copy of the draft. The two clarifications which Mr Pannick pointed out are to incorporated in it, that is to say, lines 3 on page 22, should read Counsel or Commission Regulations and line three page 23, 1600/95 not 1660/95. The appeal is dismissed with costs.

(Submissions re: House of Lords certified point.)

71. MR JUSTICE DYSON: I think this is a case in which I ought to grant leave to appeal. I have expressed my views in clear, I hope, and strong terms but others may take a different view. I grant leave to appeal because I recognise that the issues here are important and are not confined to the parties before me in their implications.