Horvath, R (on the application of) v Secretary of State for Environment, Food and Rural Affairs

A glance at a large scale map of England shows that in many areas the countryside has a very large number of public footpaths and bridleways. Many of these run through woods or cross commons or moorland. But many others cross or run along the edge of agricultural land. Some such footpaths cross pasture. Others cross arable land. In spring and summer, you can see trodden public footpaths across the middle of cornfields.

Section 146 of the Highways Act 1980 obliges owners of land to maintain stiles, gates and similar structures across footpaths and bridleways in a safe condition and to the standard of repair required to prevent unreasonable interference with the rights of persons using the footpaths or bridleways. If the owner does not do so, the appropriate authority may carry out the necessary work and recover the reasonable cost from the owner. By section 134 of the 1980 Act, farmers are allowed to plough across footpaths or bridleways, if it is not reasonably convenient to plough their fields without doing so, but they have to restore the footpaths and bridleways afterwards to make them reasonably convenient for people to exercise the right of way. It is the duty of a highway authority to enforce these provisions.

The Highways Act 1980 applies to England and Wales, but not to Scotland or Northern Ireland – see section 345.

The Common Agricultural Policy of the European Communities has as one of its central features that farmers receive a minimum guaranteed income. They may do so whether they use their land productively or not. They may become entitled to payments under the Single Payment Scheme. The present Scheme is established under Council Regulation (EC) No. 1782/2003, the whole of which is directly applicable in all Member States. One aim is that entitlement to full payment requires compliance with rules relating to agricultural land, agricultural production and activity. The rules intend to incorporate basic standards which include basic standards of “good agricultural and environmental condition”. By Article 3, in order to qualify for full direct payments, farmers have to respect statutory management requirements and “the good agricultural and environmental condition established under Article 5”. If they do not, their payments may be reduced or, in extreme cases, they may receive no payment at all.

Article 5 of the 2003 Regulation includes the following:

“Member States shall ensure that all agricultural land, especially land which is no longer used for production purposes, is maintained in good agricultural and environmental condition. Member States shall define, at national or regional level, minimum requirements for good agricultural and environmental condition on the basis of the framework set up in Annex IV, taking into account the specific characteristics of the areas concerned, including soil and climatic condition, existing farming systems, land use, crop rotation, farming practices, and farm structures. This is without prejudice to the standards governing good agricultural practices as applied in the context of Council Regulation (EC) No. 1257/1999 and to agri-environment measures applied above the reference level of good agricultural practices.”

There is reference in Annex IV to avoiding the deterioration of habitats and to retaining landscape features.

The definition of minimum requirements for good agricultural and environmental condition was effected for England by The Common Agricultural Policy Single Payment and Support Schemes (Cross Compliance) (England) Regulations 2004 – a less than snappy title. These have been updated by 2005 Regulations, but nothing turns on that. The standards of good agricultural and environmental condition required by Article 5(1) of the Council Regulation are set out in the Schedule to the 2004 Regulations. Paragraphs 26 to 29 of the Schedule are the subject of these proceedings and of this appeal by the Secretary of State from part of a decision of Crane J in the Administrative Court on 21 July 2006.

Paragraphs 26 to 28 are as follows:

“Public rights of way

26.

A farmer must not—

(a)

without lawful authority or excuse, disturb the surface of a visible footpath, a visible bridleway, or any other visible highway which consists or comprises a carriageway other than a made-up carriageway, so as to render it inconvenient for the exercise of a public right of way; or

(b)

without lawful authority or excuse, in any way wilfully obstructed the free passage along a visible highway.

27.

A farmer must maintain any stile, gate or similar structure, other than a structure to which section 146(5) of the Highways Act 1980 applies, across a visible footpath or bridleway in a safe condition, and to the standard of repair required to prevent unreasonable interference with the rights of persons using the footpath or bridleway.

28.(1) Where a farmer has disturbed the surface of a visible footpath or bridleway (other than a field-edge path) as permitted under section 134 of the Highways Act 1980, he must, within the relevant period under section 134(7) of that Act, or within an extension of that period granted under section 134(8) of that Act—

(a)

so make good the surface of the path or bridleway to not less than its minimum width as to make it reasonably convenient for the exercise of a right of way; and

(b)

so indicate the line of the path or bridleway on the ground to not less than its minimum width that it is apparent to members of the public wishing to use it.

(2)

In this paragraph, “minimum width”, in relation to a highway, has the same meaning as in Schedule 12A to the Highways Act 1980.”

Paragraph 29 contains definitions with reference to sections of the Highways Act 1980.

To a large extent, these paragraphs do no more than restate provisions already to be found in the 1980 Act. But their potential effect is different. Under the 1980 Act, an offending landowner is subject to the sanctions available under that Act. Under the regulations, an offending farmer may lose some or all of his direct payments under the Common Agricultural Policy. Furthermore, as I understand it, a farmer who incurs expense in order to comply with these requirements has to bear the cost. By contrast, it is suggested that, if the regulations did not contain these requirements, the farmer might be able to recover the cost of maintaining footpaths etc under the Common Agricultural Policy as an “agri-environment measure… above the reference level of good agricultural practices” such as is referred to in the final sentence of Article 5(1) of the Council Regulation.

Mr Horvath is a farmer in Suffolk who challenges by judicial review the lawfulness of these paragraphs of the 2004 Regulations. He, and I dare say many other farmers, would prefer not to be at risk of losing payments under the Common Agricultural Policy if they interfere with or do not restore visible footpaths and bridleways, or if they do not maintain stiles and gates across visible footpaths and bridleways. Conversely, he would like to be potentially eligible for payments which might include the cost he incurs.

Mr Horvath has two grounds of challenge. He says, first, that Article 5(1) and Annex IV of the 2003 Council Regulation do not enable the United Kingdom to include requirements relating to footpaths and bridleways in the 2004 Regulations because the expression “requirements for good agricultural and environmental condition” does not extend that far. Crane J thought that there were cogent arguments for the validity of these paragraphs in the 2004 Regulations. But he considered that the question was sufficiently open to argument to justify a reference to the European Court of Justice. He had noted that a Regulatory Impact Assessment in July 2004 had stated that the decision to include measures protecting footpaths and rights of way could be justified legally, but that there was a high risk of a challenge and perhaps fifty per cent chance of winning such a case.

Crane J accordingly ordered a reference to the European Court of Justice under Article 234 of the Treaty and Part 68 of the Civil Procedure Rules. The first referred question was:

“Can a Member State include requirements relating to the maintenance of visible public rights of way in its standards of good agricultural and environmental condition under Article 5 and Annex IV to Council Regulation 1792/2003?”

There is no appeal against the making of that part of the reference, although Buxton LJ, in giving permission to appeal to this court against the second question referred, to which I shall come in a moment, suggested that there was a high level of artificiality in the proceedings. The paragraphs in the 2004 Regulations did not, he observed, do more than require farmers to comply with existing domestic legislation under the 1980 Act; and he could not understand how these paragraphs were beyond the scope empowered by a Council Regulation which emphasised the need to achieve good environmental condition.

I pause to say that an enlarged version of the facts and the relevant European and domestic legislation, which it is not necessary to rehearse for the purposes of this appeal, may be found in paragraphs 1 to 58 of Crane J’s judgment which is at [2006] EWHC Admin 1833.

The 2004 Regulations apply only to England. There are separate regulations implementing the 2003 Council Regulation for Wales, Scotland and Northern Ireland respectively. This is because the making of implementing regulations such as these is among those matters devolved to the Welsh, Scottish and Northern Irish legislators respectively. The scheme of such devolution is that the Scottish Parliament, for instance, is responsible for effecting compliance with European legislation, but Westminster retains a residual power and obligation to see that this is achieved if Scotland were to fail to do so and it was necessary to comply with section 2(2) of the European Communities Act 1972. The details of the devolution arrangements are elaborated in Arden LJ’s judgment in this appeal.

The Welsh, Scottish and Northern Irish regulations do not contain anything equivalent to paragraphs 26 to 29 of the English regulations; and they are not the same as each other. There are, therefore, four different regulations implementing the 2003 Council Regulation within the United Kingdom.

The consequence of the difference between the English regulations and those for Wales, Scotland and Northern Ireland is that United Kingdom farmers outside England are not at risk of reduction or loss of their Common Agricultural Policy payments if they do not maintain footpaths and bridleways, as are farmers in England; and that a Welsh farmer, for instance, may be able to recover the cost of maintaining footpaths as an agri-environmental measure when an English farmer is not. There is the further anomaly that a Welsh farmer is nevertheless subject to the parallel provisions of the Highways Act 1980. Mr Horvath says that these differences constitute unlawful discrimination under EC law and that they offend Article 14 of the European Convention on Human Rights with reference to Article 1 of Protocol 1. They are differences which, if the paragraphs in the English regulations are otherwise legitimate, would in principle be capable of being justified as lawful, proportionate and necessary by reference to regional differences in soil, climate, farming systems, land use, crop rotation, farming practices and farm structures (see the terms of Article 5(1) of the 2003 Council Regulation), but the Secretary of State does not attempt to do this, and on Mr Horvath’s case cannot do so.

The Secretary of State’s case is that differences such as these, resulting from different decisions of the Secretary of State for England and of the devolved administrations for Wales, Scotland and Northern Ireland, do not offend EC law, since the defining EC regulation gives a discretion to impose minimum requirements and expressly enables the definition of those requirements to be made “at national or regional level”. On the assumption for this purpose that the English regulations are, taken alone, legitimate, there is no suggestion that the Welsh, Scottish or Northern Irish regulations are other than legitimate in themselves. These are expressly permitted differences “at regional level”, which do not need to be justified by a comparative process – no more than would equivalent differences between (say) United Kingdom regulations and French regulations, provided that each was individually legitimate.

Crane J rehearsed the parties’ submissions on this topic at length in paragraphs 59 to 86 of his judgment, where he considered a number of authorities. His conclusion in paragraph 86 was as follows:

“I have come to the conclusion that it is necessary to refer the discrimination issue to the European Court of Justice. There is authority in Klensch, Mulligan, and Romeu, that tends to support the propositions that a Member State may not discriminate in such circumstances as this and that it is not necessarily an answer to say that a devolved authority has taken the decision and was entitled to do so. The fact that the devolution legislation appears to reserve a right to make regulations to the United Kingdom government may be an additional consideration.”

The three authorities referred to in that passage are Marthe Klensch v Secretaire D’État à L’Agriculture et à La Viticulture (1986) C-201 and 202/85; Mulligan v Minister for Agriculture and Food, Ireland (2002) C-313/99; and Romeu v Commission of the European Communities (2005) T-298/02.

Crane J accordingly ordered the reference of a second question as follows:

“Where a Member State’s internal constitutional arrangements provide that different devolved administrations shall have legislative competence in relation to different constituent parts of that Member State, does it give rise to impermissible discrimination for constituent parts to have different standards of good agriculture and environmental condition under Article 5 of Annex IV to the Council Regulation?”

It is suggested that this question on reflection would be more appropriately expressed if it were introduced by the word “can” rather than “does”? The point here is that Mr Fordham does not say that these legislative differences must give rise to unlawful discrimination; but that they may do so, if they are not justified in the way that I have indicated.

The Secretary of State appeals against the part of Crane J’s order which refers this second question. Mr Eicke says that this court can and should resolve the issue in favour of the Secretary of State with complete confidence; that Crane J’s decision was plainly wrong; that this court is as well placed as he was to decide a matter which does not admit of any spectrum of judicial discretion; and that this court should be confident that no assistance from the European Court of Justice is necessary – see for this synthesis R v International Stock Exchange ex parte Else (1982) Ltd. [1993] QB 534; H.P. Bulmer Ltd v J. Bollinger S.A. [1974] Ch. 401; R (A) v Secretary of State for the Home Department [2002] EWCA Civ. 1008 and [2002] 3 CMLR 353. As the test to be applied by this court on this appeal, that is uncontroversial.

Mr Fordham says that this court should be extremely slow to entertain an appeal against the form and content of a reference, when there is going to be a reference anyway. Mr Eicke, for his part, says that the second question would open up a very wide ranging investigation into the nature and content of devolved legislation within the framework of EC law generally which is not justified by a measure which plainly authorises the particular different treatment under discussion.

Buxton LJ noted that the second question only arises if the first question is answered in the affirmative. In granting permission to appeal, he wrote:

“The proposition inherent in the second question is therefore that it may be discriminatory for one region in a Member State to impose an otherwise legitimate requirement when other regions choose not to impose that requirement. That seems so unlikely an outcome of a provision that envisages decision-making at regional level that, even noting all warnings as to the caution with which the court should proceed, it is arguable that the issue does not justify a reference. A reference will also threaten to open up very wide issues, going well beyond the facts of this case, as to whether different decisions by different regimes at federal, regional or devolved levels within a particular Member State by that fact alone attract the jurisprudence on discrimination. I am not persuaded that what is likely to be a very protracted inquiry, attracting intervention from other Member States, is necessary for the decision of a case that properly turns on the terms of one particular regulation.”

The case has proceeded and the appeal has been presented at two levels, one of general principle, the other with reference to the words of Article 5(1) of the 2003 Council Regulation. Each is informed by what Mr Fordham submits is the basic position in EC law.

A basic principle to be found in paragraphs 8 to 11 of the judgment of the European Court of Justice in Klensch is that Member States must adhere to the fundamental principle of equality. This requires that similar situations should not be treated differently unless differentiation is objectively justified. Member States must comply with this principle where Community rules leave them to choose between various methods of implementation. Member States may not choose an option whose implementation would be liable to create, directly or indirectly, discrimination between [in this instance] farmers having regard to the structure of the agricultural activities carried out in its territory. The United Kingdom is the relevant Member State, and none of England, Wales, Scotland or Northern Ireland is a separate Member State. Accordingly, says Mr Fordham, discrimination between English and Welsh farmers, for instance, is not permissible if it is not objectively justified.

Paragraphs 34 and 35 of the judgment of the European Court in Mulligan restate the principle in Klensch and say that, where Member States lay down or apply measures of this kind, they must do so on the basis of objective criteria. In paragraph 29 of the judgment in Romeu, the court said that the term “Member State” for the purposes of institutional provisions, refers only to the government authorities of the Member States and cannot include the governments of regions or autonomous communities, irrespective of the powers they may have. In paragraph 35, the court said that an action following which the Court of Justice can declare that a Member State has failed to fulfil one of its obligations can only be brought against the government of that state, even if the failure to act is the result of the action or omission of the authorities of a federal state, a region or an autonomous community.

At the level of principle, Mr Fordham submits that a Member State can devolve the authority to implement Community requirements as it sees fit, but the Member State remains responsible if what is implemented is not in accordance with Community law. It is not sufficient to show only that what is implemented by a federated or devolved authority is compliant within the confines of its own territory. It must also be shown to be compliant within the territory of the Member State as a whole. Mr Fordham directed us to paragraph 39 of the judgment of the European Court of Justice in Gattoussi v Stadt Russelsheim Case-97/05 to the effect that it would be quite unacceptable for Member States, in that instance Germany, to deal with the principle of non-discrimination laid down in Article 64(1) of the Euro-Mediterranean Agreement by using provisions of national law to limit its effectiveness. To do so would undermine the provisions of an agreement entered into by the Community and its Member States and would jeopardise the uniform application of that principle. Article 64(1) provides for non-discrimination against workers of Tunisian nationality. In this context, paragraph 39 of the judgment of the court seems to me to shed little light on the effect and consequences of a Member State devolving a power to implement Community measures to regional assemblies.

Mr Eicke’s submission was, or tended to be, that, where a Member State has federal or formally devolved regional authority, implementation of Community measures such as those under consideration in this case by the federal or devolved authority will not offend the principle of equality and non-discrimination if it does not do so within the confines of the region concerned. Thus implementation which is inoffensive in England taken alone will not offend only because that produces inequality between England and Wales. That is so in Community law as between two neighbouring Member States, and it is, or should be, the position as between formally devolved regions of a single Member State where the devolved regions have relevant legislative powers.

The Community principle as between Member States can be found in J. van Dam en Zonan and others 185/78-204/78, ECR 1979 02345, where Netherlands fishermen complained that restrictive Netherlands fishing quotas offended the principle of equal treatment between the nationals of all Member States. The European Court of Justice rejected this contention, saying, at paragraph 10 of the judgment, that it cannot be held contrary to the principle of non-discrimination to apply national legislation, the compatibility of which with Community law is not contested, because other Member States allegedly apply less strict rules. If there are inequalities of this kind, they must be eliminated by consultation. They cannot be the foundation of a charge of discrimination with regard to the provisions made by a Member State which apply to any person under its jurisdiction. National provisions cannot be considered as discriminatory as long as they are applied uniformly to all under the jurisdiction of the Member State concerned.

Mr Eicke would accept, I think, that material regional differences within England alone which are not objectively justified, as for instance between one county and another, could offend the principle of equality. But that is not a particularly helpful comparison, because there is at present no relevant regionally devolved legislative power within England. One version of his submission was, or perhaps had to be, that devolved regions with legislative powers, such as Wales, Scotland and Northern Ireland, should be treated for discrimination purposes as if they were Member States. Mr Fordham says firmly that this is not the correct approach. The responsible body for the purposes of the principle of equality is the Member State, not devolved regional parts of it. Another version, perhaps, of Mr Eicke’s submission was that choices made by regionally devolved legislative authorities, legitimate in themselves, should be treated as policy decisions, so that the reasons for their adoption should not be amenable to judicial enquiry.

Mr Eicke relied on the fact that, with the four sets of different regulations within the United Kingdom, there was no single discriminator. Conceptually this seems to me to be unpersuasive, since, if Mr Fordham is right, the single entity responsible, if there is unjustified discrimination, is the United Kingdom, the Member State. But in practical terms there are obvious difficulties if devolved regions each have legislative power to implement, but the combined product of their respective implementations cannot have differences which would constitute unlawful discrimination unless they are objectively justified. That would seem to require a degree of coercive supervision from Westminster which would tend to undermine the concept of devolved autonomy. True it is that the United Kingdom Parliament retains the obligation and formal power to ensure that the devolved assemblies do not individually or collectively contravene Community law. But using this power routinely and other than in extreme circumstances would be politically unacceptable. Mr Eicke asserted that the problem could be even more acute in Member States such as Germany, with its numerous federal authorities whose existence antedates the establishment of the European Community. He said that referring the judge’s second question to the European Court of Justice would engender unnecessary and troublesome uncertainty, since the point could extend well beyond footpaths, as for instance to the absence of prescription charges in Wales and of student fees in Scotland.

It is surprising that there is scarcely any authority directly relevant on this difficult topic. You would have thought that the question would have arisen and been answered before now, but apparently not. It might possibly be said that the absence of authority shows that Mr Fordham’s point must be a bad one. But that is not an analytical reason for reaching that conclusion with confidence.

In Commission v Germany (1988) Case-8/88 ECR 1990 page 1-02321, three German federal regions were said to have perpetrated irregularities in relation to premiums in favour of animal producers. It was alleged that the federal authorities did not give detailed instructions to the regions for appropriate checks. In this context, the European Court of Justice said at paragraph 13:

“In that connection, it should be observed that it is for all the authorities of the Member States, whether it be the central authorities of the State or the authorities of a federated State, or other territorial authorities, to ensure observance of the rules of Community law within the sphere of their competence. However, it is not for the Commission to rule on the division of competences by the institutional rules proper to each Member State, or on the obligations which may be imposed on federal or länder authorities. It may only verify whether the supervisory and inspection procedures established according to the arrangements within the national legal system are in their entirety sufficiently effective to enable the Community requirements to be correctly applied.”

Mr Eicke says that this case shows that the question of compliance was investigated at regional level. But he accepted that the case did not deal with discrimination between regions.

In Commission v Germany Case-301/95, there had been alleged incomplete transposition of Article 5(2) of Council Directive 85/337/EEC of June 1985. Germany said that stricter domestic legislation which took precedence covered the position. The ECJ said at paragraph 52:

“If, for reasons which may be linked to the federal structure of the State, other specific provisions of the Federation or the länder lay down particular requirements corresponding, possibly, to the individual needs of the various fields of activity covered by the Directive, it must be noted that Article 13 of the Directive allows the Member States to lay down stricter rules than those contained in the Directive. Furthermore, the general rule in Paragraph 4 of the UVPG ensures, as has been explained by the German Government, that the provisions of Paragraph 6(3) and (4) are to apply when the requirements contained in specific provisions fall short of those laid down in the corresponding provision of the UVPG.”

Mr Eicke says that this implies that there might be legitimate differing requirements within the regions. I find the implication at best tenuous and certainly insufficient to sustain Mr Eicke’s main submission with confidence.

In An Application by Joseph McParland for Judicial Review (2002) NICA 22, the applicant had been refused a road service licence because he had previous convictions which under Northern Irish legislation were not spent. The issue was whether this complied with the requirements of Council Directive 89/438/EEC when the Directive had been implemented in England in a more flexible manner. In Northern Ireland, the applicant could not get a road service licence whatever the merits of his case. In England, the Traffic Commissioners would have had a discretion. It was argued that Northern Ireland could not adopt a different standard from another constituent part of the United Kingdom. Carswell LCJ in the Northern Ireland Court of Appeal said:

“We consider that Directive 89/438 left it open to Member States to apply either their own domestic rehabilitation legislation or some other measure having an equivalent effect. As Mr Barling pointed out, the Community has consistently left matters relating to the criminal law to Member States and they are free to adopt any type of rehabilitation legislation. We do not think that if a Member State has a formal rehabilitation statute it is bound to use that exclusively as a standard in the present context; if it chooses to allow the application of a more lenient standard, as was the case in England, it is in our view free to do so.

It is, we agree, surprising, and it might be regarded as undesirable, that the law should differ markedly between two different constituent parts of the United Kingdom. But we would regard that as a matter of policy for the Government to address, determining whether differences in social or other conditions justify the maintenance of a different provision in each jurisdiction. It is sufficient for present purposes that the harmonising imperative contained in the Directive does not in our view invalidate the law of either merely because they differ.”

This, so far as it goes, tends to support the Secretary of State’s case. But Mr Eicke accepts that it was in a very different field from the present case. It was in the field of criminal law, and Mr Fordham correctly submits that the question was not addressed analytically by reference to authority, no doubt because none was cited. In a sense, all conscious legislative choice may be said to be a choice of policy. But some choices of policy which differ may more readily be seen to be outside the proper sphere of judicial examination than others. Conversely, some choices which differ may more obviously cry out for articulated objective justification than others.

Mr Eicke referred to two Strasbourg cases, Dudgeon v United Kingdom (1981) 4 EHRR 149 and Magee v United Kingdom (2001) 31 EHRR 35. In Dudgeon, legislation in England, which relaxed the criminal law so that private acts of what had formerly been offences of buggery and gross indecency between consenting males over 21 were no longer criminal offences, had not been matched by equivalent relaxing legislation in Northern Ireland. The European Court of Human Rights held that this constituted in Northern Ireland a breach of Article 8 of the European Convention on Human Rights. The majority of the court held that it was not necessary to examine a separate case of discrimination under Article 14 of the Convention in conjunction with Article 8. Judge Matscher, however, in a dissenting opinion alone of the nineteen judges considered the substance of the case for infringement of Article 14. In his opinion, the court could not avoid expressing an opinion by using phrases which risked restricting excessively the scope of Article 14 even so far as to deprive it of all practical value. He said at paragraph 20 of his judgment, with reference to the allegation that different legislation was in force in different parts of the United Kingdom:

“The diversity of internal legislation inherent in a federal State can never, in itself, constitute discrimination, and it is unnecessary to justify it. To claim the contrary would be to mistake totally the very essence of federalism.”

Mr Fordham accepts that this supports the Secretary of State’s case. But he says that this lone voice is insufficient for complete confidence on an issue which in any event has to look to the proper meaning and extent of Article 5(1) of the 2003 Council Regulation.

In Magee, the applicant, who had been arrested in connection with an attempted bomb attack, complained of ill-treatment after arrest when he was being interviewed and that he had for a time been refused access to a lawyer. He also complained that under the law in Northern Ireland adverse inferences could be drawn from his silence in interview. The European Court of Human Rights held as its main decision that there had been a violation of Article 6 of the European Convention on Human Rights. They dismissed a subsidiary complaint of discrimination in violation of Article 14 in conjunction with Article 6, because suspects arrested and detained in England and Wales under prevention of terrorism legislation could have access to a lawyer immediately and were entitled to his presence during interview; and because in England and Wales incriminating inferences could not at the time be drawn from an arrested person’s silence during interview. The complaint of a violation of Article 14 failed because, under that Article, the discrimination relevantly had to be on the ground of national origin, association with a national minority or other status. The differences complained of were not explained in terms of personal characteristics, but on the geographical location where the individual was arrested and detained. Article 14 permitted “legislation to take account of regional differences and characteristics of an objective and reasonable nature”. This decision seems to me to be of little assistance in the present case, because it depended on the ambit and extent of the words of Article 14. But on one view, the passage which I have quoted predicates that regional differences need to be objectively justified.

At the particular level, the parties’ submissions address the construction of Article 5(1) of the 2003 Council Regulation. Accepting that the authorities do not by themselves provide resounding support for the Secretary of State’s case, Mr Eicke nevertheless submits that none of them was decided with the benefit of the words “at national or regional level”. He submits that, by these words, Article 5(1) expressly permits minimum requirements for good agricultural and environmental condition to be implemented at regional level. With devolved or federated assemblies, regional provisions, legitimate in themselves, which are not discriminatory internally within the region, do not have to be justified in comparison with different but equally legitimate provisions in other regions of the same Member State. Mr Fordham submits that this is not the correct reading of the Article; or at least not unarguably so. It is not entirely clear whether it is the process of defining or the resulting definition which is permitted at national or regional level. But, if that may be something of a pedantic quibble, the definition has to take account of the specific characteristics of the areas concerned including those listed, and unjustified regional differences do not achieve this. This, says Mr Fordham, supports his submission that devolved regional differences are permitted provided they are objectively justified. In any event, Article 5(1) places the obligation which it imposes on Member States. Achieving the obligation can be done in any way that is legitimate within the Member State, but the result has to be judged at the level of the Member State.

Mr Fordham finds support for his construction from other parts of the Council Regulation. He points to the use of the expression “the competent national authority” in Article 3(2), but that seems to me to be ambivalent. Somewhat more persuasive is Article 58. This concerns regional allocation of the ceiling referred to in Article 41, and enables Member States to decide to apply the single payment scheme at regional level and requires Member States to define the regions “according to objective criteria”. Article 74 concerns national base areas in the traditional production zones listed in Annex X. A Member State may sub-divide its base areas into sub-base areas “in accordance with objective criteria”. Annex X has lists of traditional production zones for durum wheat as referred to in Article 74. For France, Italy and Hungary, the zones are referred to as “Regions”. These, admittedly straws in the wind, suggest that “regions” and “regional level” do not refer to regions to which implementation of the requirements of Article 5 are devolved, but to geographical sub-divisions made by Member States for the purpose of implementing the Regulation. The references to objective criteria suggest that it is a requirement that differential regional definition for the purposes of Article 5(1) should be objectively justified.

In summary, the Secretary of State’s case depends on one or both of the following:

I am not completely confident that the construction of Article 5(1) in paragraph 38(a) above is correct. Indeed, I am provisionally inclined to think that, taken alone, it may be wrong. I am provisionally inclined to think that “regional level” refers to geographical regions; that it remains the Member States who have to do the defining; that an objective justification for differences is necessary with reference to the specific characteristics of the areas concerned; but, importantly, that the Regulation may not have been drafted with devolved assemblies in mind. It may in these circumstances be possible for the European Court of Justice to do some creative construction here, or to see the words of the Article as operating within a general framework by which regulations such as this may be implemented by devolved assemblies. But I am not confident that this court should do so without their assistance.

As to paragraph 38(b) above, I am not completely confident that the European Court of Justice would so decide and I am not confident that assistance from the European Court of Justice is unnecessary. I do not think that complete confidence can be found in the single voice, dissenting on the main issue, of one judge of the European Court of Human Rights in 1981. None of the other authorities gives more than indications, not all of which point in the same direction. It is strange indeed that the issue has not arisen for clear decision before. But, as I have said, that is not an analytical reason for complete confidence. I would rather expect that the European Court of Justice would find a way, perhaps with reference to the principles of subsidiarity, of concluding that the Secretary of State is correct. It would certainly be constitutionally unsatisfactory within the United Kingdom if this were not so; but that again is not a ground for complete confidence. It would, I think, involve developing an analysis or principle which no authority to which we have been referred has yet developed. I am not completely confident that this court can or should develop that analysis or principle, as it would have to consider doing if this were purely a domestic matter. If the Secretary of State succeeds on the first issue referred, the European Court of Justice may have to face up to the fact that, under individually legitimate existing regulations, neighbouring English and Welsh farmers are treated differently within the same Member State.

For these reasons I would dismiss this appeal.

I likewise conclude that this appeal must fail for the reasons explained in this judgment.

We have to determine whether the judge, Crane J, erred in making a reference of an issue as to art 5 of Council Regulation 1782/2003 (“the Council Regulation”) to the Court of Justice of the European Communities ("the Court of Justice"). As May LJ has explained, the issue involves the application of the Community law principle of non-discrimination ("the non-discrimination principle") to the minimum standards defined by member states for the purpose of art 5 of the Council Regulation. Art 5 has been implemented by the United Kingdom. In the case of Scotland, Northern Ireland and Wales implementation has been by regional bodies pursuant to the United Kingdom’s devolution arrangements. In this judgment I use the expression “regional body” to mean an organ within a member state that has shared or exclusive competence under national law to implement Community law on a regional basis. The legislation which implements the Council Regulation in England, Wales, Scotland and Northern Ireland contains material differences.

The question at issue is whether regional bodies can discharge the member state’s obligations to implement art 5 differently for their different regions without having to justify their departure from some legislative norm applying to the member state as a whole, that is, minimum standards determined for the whole of the United Kingdom. The United Kingdom has no process for establishing common standards applying to the United Kingdom as a whole where a Community or other measure is implemented by the devolved administrations under the United Kingdom’s devolution arrangements (explained below).

The judge’s order provides for another question to be referred to the Court of Justice at the same time, but there is no appeal from that part of his order.

For the purposes of the question with which we are concerned, there is no suggestion that the legislative acts of the United Kingdom implementing the Council Regulation, or the United Kingdom's constitutional arrangements, involve any other non-compliance with Community obligations.

As to the background, I gratefully adopt what has already been said by May LJ and I respectfully agree with his careful analysis of the cases cited by him.

I reach the conclusion expressed in the first sentence of this judgment for the following reasons which I amplify below:

In para. 39 of his judgment, May LJ has expressed certain provisional views with which Scott Baker LJ agrees. It would be convenient to state at an early point in this judgment where I have come to different conclusions, although those differences are not material to the result of this appeal. I agree with them that this appeal should be dismissed. However, I do not share their provisional view that the critical sentence of art 5 of the Council Regulation does not permit regional implementation. Moreover, I express doubts as to whether it is correct to say as a matter of Community law that (where member states have a discretion as to the manner of implementation) differential implementation by regional bodies needs to be objectively justified for the non-discrimination principle, and accordingly I do not share the tentative provisional view of May and Scott Baker LJ rejecting the Secretary of State’s submission on this. In addition, I consider that on its true interpretation art 5 permits implementation by regional bodies. However, since we are all agreed for other reasons that the judge was right to make a reference, those differences of view are not determinative of this appeal, though in another case the existence of those differences could well have been a reason for making a reference.

I now take each of the reasons given above in turn in order to develop them. At the end of my judgment, I deal with three final matters, namely art 1 of the First Protocol to the European Convention on Human Rights, the approach on appeal to an order for reference and directions.

The devolution arrangements in the United Kingdom are unusual, but the devolved administrations have competence to implement Community obligations for their respective territories in devolved subject areas, such as the environment and agriculture, should they choose to exercise them, and therefore they fall to be treated under Community law like states with fuller federal status

The devolution arrangements were introduced by Acts of the United Kingdom Parliament (“the Westminster Parliament”) in 1998. It is not necessary for me to describe all the elements of the devolution arrangements, and the description that follows incorporates the helpful analysis in the skeleton argument of Mr Tim Eicke, for the Secretary of State, with which analysis Mr Michael Fordham QC, for Mr Horvath, largely agrees. These Acts were passed as part of a “programme of constitutional legislation …with a view to improving the government of the United Kingdom as a whole.” (per Lord President Rodger in Whaley v Lord Watson of Invergowrie and the Scottish Parliament 2000 SC 340 at 353). The Acts set up new legislative bodies in Scotland, Northern Ireland and Wales.

I start with Scotland. S 28 of the Scotland Act 1998 (“the 1998 Act”) establishes a new Scottish Parliament. This is not an independent sovereign body, but a statutory body which has to work within the powers conferred by the Westminster Parliament and is subject to the courts in various respects. As Lord Rodger said in the Whaley case, the Scottish Parliament in that respect joined the family of Parliaments to which the Westminster Parliament had in the past transferred sovereignty throughout the Commonwealth. It has competence to legislate for Scotland save as provided in s 29. The Scottish Parliament cannot legislate for “reserved matters”, which includes relations with the European Union, but not the obligation to implement Community law (schedule 5, para. 7). Neither agriculture nor the environment is a reserved matter. The 1998 Act in effect transfers to the Scottish Ministers (who are the members who head the executive in Scotland: see s 44(1)) within their devolved competence the powers to implement Community law by regulations conferred by s 2(2) of the United Kingdom’s European Communities Act 1972: see per Lord Rodger in Beggs v Scottish Ministers [2007] 1 WLR 455 at [32]. The United Kingdom government has retained power to make regulations for implementing Community law (s 57), but, as I explain below, it has agreed to exercise the power in particular circumstances only. The 1998 Act also provides that it is outside the competence of the Scottish Ministers to act or legislate in a way that would be incompatible with EC law (s 29(2) and 57(2)).

The devolution arrangements for implementing Community obligations are so far as relevant the same in Wales and Northern Ireland as in Scotland, but they are governed by separate legislation. The legislative competence of the Welsh Assembly is less than that of the Scottish Parliament, although the powers of the Welsh Assembly have been enhanced by the Government of Wales Act 2006. The Northern Ireland Act 1998 established devolved government in Northern Ireland and the new Northern Ireland Assembly. Its provisions were to take effect once sufficient progress had been made in completing an agreement concerning the future of Northern Ireland known as the Belfast Agreement and they were brought into force in December 1999. Later, the Northern Ireland Act 2000 was passed to give the Secretary of State the powers to suspend the Northern Ireland Assembly by order in council and to revoke the suspension (thus reviving the power of suspension). These powers of suspension and revocation were each exercised on four occasions, but they were repealed by legislation restoring devolved government in Northern Ireland on 8 May 2007 (see the Northern Ireland (Restoration of Devolved Government) Order 2007 (SI 2007/1397)).

The statutory relationship between the devolved administrations and the United Kingdom government in relation to obligations to implement Community law is supplemented by a Devolution Memorandum of Understanding (“Devolution MoU”) entered into in December 2001 between the government of the United Kingdom, Scottish Ministers, the Cabinet of the National Assembly for Wales and the Northern Ireland Executive Committee (Cm. 5240). The Devolution MoU contains a series of non-legally binding agreements between the United Kingdom government and the devolved administrations and sets out the principles which underlie their intergovernmental relations. Para. 2 of the Devolution MoU states that:

“This Memorandum is a statement of political intent, and should not be interpreted as a binding agreement. It does not create legal obligations between the parties. It is intended to be binding in honour only... ”

Para 20 of the Devolution MoU provides that:

“The devolved administrations are responsible for implementing international, ECHR and EU obligations which concern devolved matters. In law, UK ministers have powers to intervene in order to ensure the implementation of these obligations. If the devolved administrations wish, it is open to them to ask the UK government to extend UK legislation to cover their EU obligations. The devolved administrations are directly accountable through the domestic courts, in the same way as the UK government is, for shortcomings in their implementation or application of EC law. It is agreed by all four administrations that, to the extent that financial penalties are imposed on the UK as a result of a failure of implementation or enforcement, or any damages or costs arise as a result, responsibility for meeting them will be borne by the administration(s) responsible for the failure.”

This makes it clear that, in terms of the division of functions and responsibilities between central government and devolved administrations, the primary responsibility of implementing obligations arising under EC law in relation to devolved matters lies within the devolved administrations. In practice, United Kingdom ministers would only act either to ensure that such obligations are in fact implemented or on the invitation of the devolved administration in question. Thus, the power of the United Kingdom government to make implementing regulations for the whole of the United Kingdom is shared, but by agreement the powers are not co-extensive. The United Kingdom government has accepted that its power arises only upon a request from the devolved administration (which might simply be made to avoid any debate as to competence) or a failure by a devolved administration duly to implement a Community measure. Accordingly a devolved administration can pre-empt the exercise by the United Kingdom government of its powers by implementing the Community law.

The United Kingdom devolution arrangements lack some of the characteristics of a federal system. The Westminster Parliament has not given up its sovereignty over the devolved administrations and that means that in theory, subject to constitutional conventions, it could restrict or revoke the powers that it has given to the devolved administrations. Furthermore, there is no provision for judicial review of legislation passed by the Westminster Parliament on the grounds that it deals with devolved matters. The only qualification to that principle is if the court decides that the legislation of the Westminster Parliament violates Community law. If any such question arises, the courts of any part of the United Kingdom can refer a question to the Court of Justice for a preliminary ruling. In addition, there is no separate legislative body for England as opposed to Wales, Scotland or Northern Ireland. The judicial systems for England and Wales are not separate. There is no dual system of courts in any part of the United Kingdom. Moreover, the United Kingdom ministers have, as I have described, a reserve power with respect to the implementation of Community law.

However, the important point is that for Scotland, Wales and Northern Ireland there are political structures now in place that have stability. In addition, the legislation of the devolved administrations is subject to judicial review. The stability of the devolution arrangements will be enhanced over time by the operation of constitutional conventions. Constitutional conventions play a large part in the United Kingdom where the constitution is uncodified. For example, conventions developed in the nineteenth and early twentieth century so that the Westminster Parliament would take no step to amend the Canadian constitution except at the request of the Canadian government approved by a resolution of the Canadian Parliament, even though the Canadian Parliament, like the Scottish Parliament, was a body to which the Westminster Parliament had transferred powers in the past (see Monaghan, Constitutional Law, 3 ed. 2006, page 162-4). Irrespective of the question of whether a convention exists at the present time, the Devolution MoU affirms the intention of the United Kingdom government to “proceed in accordance with the convention that the UK Parliament would not normally legislate with regard to devolved matters except with the agreement of the devolved legislature."(para.13)

It should be noted that in the United Kingdom there is no assembly in which all the devolved administrations are represented. Thus there is no systematised process for the collective representation of the devolved administrations or the formulation of their common view on devolved matters. But this is not a necessary feature of a federal system either. This point is, however, important because Mr Fordham submits that art 5 of the Council Regulation can only be implemented by the devolved administrations differentially if there are United Kingdom minimum requirements from which departures can be justified. As I have said there is no formal process for determining United Kingdom minimum requirements in those circumstances, although the Devolution MoU contemplates that there will be intergovernmental consultation (paras. B4.16 to B4.19). The Westminster Parliament may be unable, consistently with the devolution arrangements as explained above, to carry out the formulation of those requirements on its own where the subject matter falls within the remit of the devolved bodies. In addition, the costs of implementation will fall on the devolved administrations (unless borne by the European Union), and thus in principle it should be for them to choose how implementation in their area of responsibility should take place. While Germany may have appropriate arrangements (see art 35 of its Basic Law), they do not exist in the United Kingdom.

Community law clearly permits implementation by regional bodies, although the treaty obligation to ensure compliance with Community obligations remains with the member state

This is demonstrated for instance by the extract from the judgment of the Court of Justice in Commission v Germany C-8/88, which is considered by May LJ in para. 31 above and by Commission v Germany (case 103/10) referred to below.

However, there is no decision of the Court of Justice on the question whether the non-discrimination principle applies to prevent differential implementation by regional bodies without objective justification, where, as in art 5, member states are given a discretion as to the form of implementation or standards to be imposed

Mr Tim Eicke, for the Secretary of State, accepts that there is no decision directly in point on the issue of Community law that has arisen in this case. We have not been shown, nor would we expect to see, a decision on art 5 itself, but there is also no authority on the principle of Community law involved here in any other context. Thus there is no authoritative decision which makes it clear that, where Community legislation is in fact implemented at a regional level, the principle of non-discrimination is not engaged by differences in the implementing legislation. It is essential to the Secretary of State’s case to demonstrate that regional bodies may exercise the power of implementation differentially without any need to justify those differences. Mr Eicke has to draw a distinction between national agencies or bodies in a purely unitary state and regional bodies in a member state with devolution arrangements or with a fully federal system.

We have not been taken to any decision dealing expressly with the effect for the purposes of the non-discrimination principle of a division of responsibility for implementing Community obligations in this situation. Regional legislative bodies exist in several other member states, such as Austria, Belgium, Spain and Germany (for more details, see Legislative, Executive and Judicial Governance in Federal Countries (ed. Le Roy and Saunders), and A Global Dialogue on Federalism, vols1, 2, 3 and 5, McGill-Queens University, 2006).

I find insufficient support for the Secretary of State’s submission in either of the two cases involving Germany and the Commission cited by Mr Eicke (see above, paras, 31 and 32). In Commission v Germany C-8/88, the länder (that is, the states which form the German federation) had to enact rules which fulfilled certain criteria. The member state had discretion as to the form of the rules and different länder exercised it differently. Neither the Court of Justice nor the Commission suggested that the rules had to be the same throughout Germany, unless the differences were objectively justified, and that is a curious omission if Mr Fordham is right. But the point is not addressed directly. In Germany v Commission C-301/95, the question was whether a member state could impose requirements additional to those required by Community law and the Court of Justice accepted that the länder or the federal government could impose different requirements. These additional requirements were not therefore subject to Community law and the Court of Justice did not deal with the question whether different länder could legislate for different requirements.

Similar points arose in the later case of Commission v Germany (Case 103/01) [2003] ECR 1-5269 (not cited). In this case, two of the länder in Germany had decided to implement a Community harmonisation measure (the PPE Directive) on the personal protective equipment (“PPE”) by imposing additional conditions. This then is a case of differential legislation by regional bodies within a single member state. The Court of Justice decided the issue as a matter of the interpretation of the directive in question. It concluded that additional conditions could not be attached because that would be an impediment to the free movement of goods. The Court of Justice held:

“46.

By harmonising the national provisions relating to PPE intended for the protection of fire fighters in the performance of their usual duties, the PPE Directive does not infringe either the principle of subsidiarity or that of proportionality…

49.

Since the derogation provided for by point 1 of annex one to the PPE Directive does not apply in this case, the länder were not entitled, by virtue of article 4 (1) thereof, to impose additional conditions on PPE which satisfies the provisions of that directive and bears the EC marking."

When it gave relief at the end of its judgment, the Court of Justice declared "that by subjecting, by means of the legislation of certain länder, [PPE] for firefighters to additional requirements despite the fact that it complies with the requirements of Council Directive 89/66/EEC … the Federal Republic of Germany has failed to fulfil its obligations under… that Directive”.

Effectively, the conclusion was that the Community measure pre-empted any member state measure in the same field. The Court attached significance to the fact that the measure was a harmonisation measure to remove a barrier to trade. For that reason, the länder could not adopt additional requirements. By implication, if the measure had permitted additional requirements to be attached, the länder could have adopted differential legislation. One interpretation of this decision is that, where Community law does not require member states to impose the same requirements (as under a harmonisation measure), then regional bodies can pass legislation, and in that event , just as member states are not bound to adopt the same requirements as are adopted by other member states, so too regional bodies can impose different requirements. However, again, that point is not addressed in terms by the Court of Justice.

Mr Eicke referred briefly to subsidiarity at the start of his oral submissions, and I refer to it again under (v) below. Commission v Germany (Case 103/01) provides some support for Mr Eicke’s reliance on this principle in connection with the issue of differential implementation by regional bodies. But, as I have explained, we find no clear answer in Community law and we therefore have to look to see whether the Council Regulation has expressly dealt with the point. In the absence of a clear provision in the Council Regulation, I consider the point of principle is so fundamental that this court cannot be confident of the answer to it without guidance from the Court of Justice.

Art 5 of the Council Regulation expressly permits implementation by regional bodies, but it is not clear whether the non-discrimination principle applies to prevent differential implementation by regional bodies under art 5 without objective justification

Relevant extracts of the Council Regulation are set out in the judgment of May LJ and the judge. The critical second sentence of art 5(1) provides that:

“Member States shall define, at national or regional level, minimum requirements of good agricultural and environmental conditions on the basis of the framework set up in Annex IV, taking into account the specific characteristics of the areas concerned, including soil and climatic condition, existing farming systems, land use, crop rotation, farming practices, and farm structures.”

This provision is clearly not self-executing. It specifically requires further measures by member states. There is no one method by which the measures are to be achieved. There can therefore properly be differences in the requirements laid down by different member states.

In any devolved or federal system, differences between implementing measures (if some latitude is permitted) are bound to occur. This was one of the points made by Judge Matscher in Dudgeon v United Kingdom (above, para.34). Diversity is even more likely in the United Kingdom, which has three different legal systems. The Scottish system is distinctive, having civilian origins not shared by the rest of the United Kingdom.

In the context of the Council Regulation, the term “regional" (which appears near the start of the critical sentence) will bear the meaning assigned to it by Community law. May LJ has referred to the possibility that the words "at national or regional level” in that sentence may refer either to the process of defining the requirements for agricultural and environmental condition or the results of the process of definition (judgment [36]). (We have not been shown any other language versions of the Council Regulation. I have independently considered the French version of art 5, but that does not clarify this point). In my judgment, if there are specific characteristics of regions within the member states, those characteristics must be taken into account under the last 27 words of the critical sentence in art 5. The words "at national or regional level" at the start of the sentence in my judgment refer to the political structures for defining the minimum requirements. Those political structures can under the wording of art 5 be regional or national, depending on the internal constitutional arrangements of the member state. The words "at national or regional level", read naturally, qualify the word "define". There is no point in including those words, unless they have a different meaning from the last 27 words of the critical sentence. My interpretation is both in accordance with the language used and logical. A process of legislative definition has inevitably to use political structures. The Council Regulation could reasonably have been adopted on the view that it was more effective and more appropriate to use regional political structures, where they exist, than national ones, particularly in relation to the environment. Mr Fordham submits that a regional political structure may not match the geographical region. In my judgment, this is no answer. That is always going to be a potential problem in relation to the environment, and it is not eradicated by requiring legislation at the level of the member state, which may have territorial jurisdiction only over a part of a region, such as a mountain range or river.

The judge explained in his judgment that references to the environment to the Council Regulation were added during the drafting process. Neither party on this appeal has sought to rely on those matters, nor have we been concerned with the legal basis in Community law for the Council Regulation.

Accordingly, I conclude that art 5 permits regional bodies to lay down the minimum requirements, but that still leaves the question whether differential implementation within a member state by regional bodies requires objective justification under Community law for departures from some national set of minimum standards. The wording of art. 5 does not provide a clear answer to that point.

Further reasons supporting the judge’s order for a reference include (a) the developing nature of the concept of subsidiarity and (b) the fact that the issue raises constitutional questions for the Community

A further reason supporting the judge's order for a reference is that the issue of interpretation will need to be considered at the level of fundamental principle (on this, see generally Meltzer, Member State Liability in Europe and the United States, (2006) Int’l J Con Law 39, 59-67). If the rationale of the non-discrimination principle is to prevent unjust action by a member state, that rationale has no application where the difference in treatment arises only because of implementation by a regional body acting properly and proportionately in a manner different from that of another regional body in the same member state.

Likewise, the principle of subsidiarity has to be considered. This principle was introduced by the Maastricht treaty. (An account of some of the negotiations over subsidiarity can be found in Gunlicks, The Länder and German Federalism (Manchester University) 2003, pages 360-373). The principle is thus relatively new and there has been little judicial consideration of it (see Tridimas, The General Principles of EU Law (Oxford) (2006) pages 183 to 192, Craig, EU Adminstrative Law, (Oxford) (2006) pages 422 to 433). It is moreover arguable that the Community has no legitimate interest in enforcing uniformity across the constituent elements of a federal state, contrary to the internal arrangements of that state, where, as in the critical sentence of art 5, it has chosen not to impose uniformity on the member states themselves. Uniformity in those circumstances is, in the eyes of the Community institutions responsible for the measure, not required for the purpose of ensuring the effectiveness of Community rights or for the purposes of the advancement or coherence of the European Union. On the contrary, implementation might be greatly impeded if the formulation of national minimum standards were required. Moreover, if the principle of subsidiarity does not apply, then any increase in the activities of the Community institutions is liable over time adversely to affect internal constitutional arrangements involving regional bodies by reducing the autonomy of regional bodies and augmenting the powers of the national body. The issues arising here are clearly better dealt with by the Court of Justice than a national court since the Court of Justice will be able to consider the issues from the perspective of all the member states.

Mr Fordham makes the point that the observations of the Strasbourg Court in Magee v United Kingdom (see above, para.35) provide support for his submission. There is force in that point, but the Strasbourg court would not be concerned in issues of the kind mentioned in the preceding paragraph.

Mr Fordham also relies on the right to property in art 1 of the First Protocol to the European Convention on Human Rights. Mr Eicke did not address us on this in his submissions, and accordingly I do not refer to it further in this judgment, other than to say that it does not seem to me, provisionally, that it raises any additional issue which requires to be considered by this court on this appeal.

I must finally deal with Mr Fordham's submission that the test on this appeal is whether the judge was plainly wrong in concluding the second question in a reference. He submitted that this court should not set aside the exercise by the judge of his discretion to make the reference of the second question unless it was satisfied that he had gone outside the reasonable ambit of his discretion. Mr Fordham also suggested that there was some distinction between making a reference and including a question in a reference, but I do not understand what the difference can be, other than the lessening of any argument about cost and delay. In this case, the question has been whether the judge correctly directed himself as to Community law when he decided to make the reference. This is not an appeal against the weighing up by the judge of considerations relevant to the exercise of discretion, but an appeal on the point of law. If he was wrong on the view of Community law he took and the answer is clear, the appeal would be allowed without reference to any discretionary factor relevant to making a reference. The issue in this case therefore is whether the judge was correct in law and not whether he was perverse in the way he exercised his discretion.

Finally, I turn to Mr Eicke’s submission that if this matter is referred, there will be uncertainty in the execution of devolution arrangements. It is impossible to avoid the conclusion that this case raises important constitutional issues for the United Kingdom that will affect the executive in the Westminster government and the devolved administrations in their day to day activities. The reference may also cast doubt on what has happened in the past. I accept that these are serious matters. However, they may provide an additional argument for a reference rather than an argument against it. That said, this reference is likely to make it difficult for the devolved administrations to operate and may give rise to the risk of state liability, which under the arrangements described above will fall on the devolved administrations responsible. I do not underestimate all these difficulties. The situation is likely to call for a considerable amount of cooperation between the devolved administrations, as is regularly required in a federal situation: see generally Intergovernmental Relations in a Devolved United Kingdom: Making Devolution Work, Cornes, in Constitutional Futures ed. Hazell (The Constitution Unit) (1999). The court asked the parties to consider whether they would wish this court to make a request to the President of the Court of Justice pursuant to rule 104a of the rules of the Court of Justice for an accelerated procedure to be applied in this case, but the parties have not asked the court to make this request as they do not consider that this would be a suitable case for such a request to be made. In those circumstances I would not make a request to the President of the Court of Justice for an accelerated procedure.

The judge gave a direction under CPR 68 PD1.4 that a copy of his judgment should be annexed to the reference. The judge’s judgment is, however, some ninety paragraphs. In addition, parts of the judgments of this court will be relevant. The practice direction draws attention in para. 1.2 to the need to translate documents into many other languages and they should always therefore be as succinct as possible. In those circumstances, I would direct the parties to consult the Registry of the Court of Justice to see whether it would assist the Court of Justice to have a summary of the judgments prepared by counsel (either instead of or in addition to the judgments of this court and the High Court). If so, and subject to any further submissions from counsel which may be submitted in writing after this judgment is handed down, I would direct the parties to provide to the Senior Master within a period to be fixed by the court (if not agreed) an agreed summary of the various judgments of the national court (insofar as relevant to the Court of Justice), that this summary must not exceed fifteen A4 pages (single-spaced) and that it must be clearly headed “summary prepared by the parties”. It is important that the summary should be complete (as regards any matter relevant to the proceedings before the Court of Justice) and accurate, as this document would then be used by the members of the Court of Justice and circulated to the other member states to enable them to decide whether to make representations to the Court of Justice. I would make this direction in case the officials of Court of Justice should themselves decide to omit parts of the judgments. This process may lead to the omission of passages which the parties wish to be before the Court or to be included in the communication about the reference from the Court to the other member states. It may also add to the delay.

In conclusion, I would dismiss this appeal and give the direction set out in the preceding paragraph.

I agree that this appeal should be dismissed for the reasons given by May LJ.