Recall Support Services Ltd & Ors v Secretary of State for Culture, Media and Sport

Having found that it is only the Authorisation Directive that applies in this case, I now turn to the issue as to what grounds of justification are available to a Member State under that Directive.

As I have already explained in outline, the main thrust of the Authorisation Directive is to provide that Member States can still control the use of telecoms equipment by the grant of authorisations but that the State should wherever possible issue general authorisations covering all users rather than requiring users to apply for an individual licence. Thus Article 3 of the Authorisation Directive provided (emphasis added):

“Article 3

General authorisation of electronic communications networks and services

1.

Member States shall ensure the freedom to provide electronic communications networks and services, subject to the conditions set out in this Directive. To this end, Member States shall not prevent an undertaking from providing electronic communications networks or services, except where this is necessary for the reasons set out in Article [52 TFEU].

2.

The provision of electronic communications networks or the provision of electronic communications services may, without prejudice to the specific obligations referred to in Article 6(2) or rights of use referred to in Article 5, only be subject to a general authorisation. The undertaking concerned may be required to submit a notification but may not be required to obtain an explicit decision or any other administrative act by the national regulatory authority before exercising the rights stemming from the authorisation. Upon notification, when required, an undertaking may begin activity, where necessary subject to the provisions on rights of use in Articles 5, 6 and 7.

3.

The notification referred to in paragraph 2 shall not entail more than a declaration by a legal or natural person to the national regulatory authority of the intention to commence the provision of electronic communication networks or services and the submission of the minimal information which is required to allow the national regulatory authority to keep a register or list of providers of electronic communications networks and services. ….”

The scheme therefore required Member States to grant general authorisations to users, subject to the ability to establish ‘rights of use’ in accordance with later articles, and to an obligation on the user to notify. As is clear from Article 3, a key provision for our purposes is Article 5 of the Authorisation Directive. That sets out the circumstances in which a Member State is entitled to insist on a scheme which grants rights of use, or individual licences, rather than conferring a general right or authorisation to use the equipment.

When the Authorisation Directive was first adopted in 2002, Article 5 read as follows:

‘Article 5

Rights of use for radio frequencies and numbers

1.

Member States shall, where possible, in particular where the risk of harmful interference is negligible, not make the use of radio frequencies subject to the grant of individual rights of use but shall include the conditions for usage of such radio frequencies in the general authorisation.

2.

Where it is necessary to grant individual rights of use for radio frequencies and numbers, Member States shall grant such rights, upon request, to any undertaking providing or using networks or services under the general authorisation, subject to the provisions of Articles 6, 7 and 11(1)(c) of this Directive and any other rules ensuring the efficient use of those resources in accordance with Directive 2002/21/EC (Framework Directive). Without prejudice to specific criteria and procedures adopted by Member States to grant rights of use of radio frequencies to providers of radio or television broadcast content services with a view to pursuing general interest objectives in conformity with Community law, such rights of use shall be granted through open, transparent and non-discriminatory procedures. …’

I shall refer to that version of Article 5 of the Authorisation Directive as ‘the Original Article 5’.

The Better Regulation Directive adopted in 2009, by Article 3(3), substituted a new Article 5 of the Authorisation Directive which reads as follows:

‘Article 5

Rights of use for radio frequencies and numbers

1.

Member States shall facilitate the use of radio frequencies under general authorisations. Where necessary, Member States may grant individual rights of use in order to:

— avoid harmful interference;

— ensure technical quality of service,

— safeguard efficient use of spectrum, or

— fulfil other objectives of general interest as defined by Member States in conformity with Community law.

2.

Where it is necessary to grant individual rights of use for radio frequencies and numbers, Member States shall grant such rights, upon request, to any undertaking for the provision of networks or services under the general authorisation referred to in Article 3, subject to the provisions of Articles 6, 7 and 11(1)(c) of this Directive and any other rules ensuring the efficient use of those resources in accordance with Directive 2002/21/EC (Framework Directive).

Without prejudice to specific criteria and procedures adopted by Member States to grant rights of use of radio frequencies to providers of radio or television broadcast content services with a view to pursuing general interest objectives in conformity with Community law, the rights of use for radio frequencies and numbers shall be granted through open, objective, transparent, non-discriminatory and proportionate procedures, and, in the case of radio frequencies, in accordance with the provisions of Article 9 of Directive 2002/21/EC (Framework Directive). An exception to the requirement of open procedures may apply in cases where the granting of individual rights of use of radio frequencies to the providers of radio or television broadcast content services is necessary to achieve a general interest objective as defined by Member States in conformity with Community law.’

I shall refer to that version of Article 5 as ‘the Amended Article 5’.

It is accepted by the Claimants that Article 5 is not entirely the exhaustive list of reasons why a Member State can impose restrictions. A Member State may also rely on the reasons set out in Article 52 TFEU (formerly Article 46 EC). Article 52 TFEU provides (emphasis added):

‘Article 52

1.

The provisions of this Chapter and measures taken in pursuance thereof shall not prejudice the applicability of provisions laid down by law, regulation or administrative action providing for special treatment for foreign nationals on grounds of public policy, public security or public health.’

As I have said, DCMS seek to justify the restriction on the use of GSM Gateways by arguments relating to public security, avoidance of harmful interference and the efficient use of spectrum. The avoidance of harmful interference is expressly mentioned in both the Original and Amended Article 5(1) and the Claimants accept that it is available to DCMS as a matter of law as a possible justification for the Commercial Use Restriction. The Claimants do not accept that the other two grounds, public security and efficient use of the spectrum, are available to DCMS as justifications. For each of these grounds therefore I need to consider:

As regards some of these questions, the Claimants conceded that the ground of justification was available to DCMS although they did not accept that the ground was made out on the facts.

The first question is whether it is open to a Member State to seek to rely on public security arguments in order to justify a regime which requires an individual licence to be granted for the use of GSM Gateways under the Original Article 5. On this point, the Claimants concede that it is open in theory for a Member State to rely on the grounds set out in Article 52 TFEU (that is public policy, public security or public health). But, the Claimants say, DCMS cannot in this case rely on that ground to justify the Commercial Use Restriction because the Government failed to make the necessary notification to the European Commission under what is now Article 114 TFEU. In the absence of such a notification, DCMS cannot rely on a public security justification which would otherwise be available as a matter of EU law.

Article 114 TFEU (formerly Article 95 EC) provides:

“Article 114

1.

Save where otherwise provided in the Treaties, the following provisions shall apply for the achievement of the objectives set out in Article 26. The European Parliament and the Council shall, acting in accordance with the ordinary legislative procedure and after consulting the Economic and Social Committee, adopt the measures for the approximation of the provisions laid down by law, regulation or administrative action in Member States which have as their object the establishment and functioning of the internal market.

…

4.

If, after the adoption of a harmonisation measure by the [EU], a Member State deems it necessary to maintain national provisions on grounds of major needs referred to in Article 36, … it shall notify the Commission of these provisions as well as the grounds for maintaining them.

5.

Moreover, without prejudice to paragraph 4, if, after the adoption of a harmonisation measure by the European Parliament and the Council, by the Council or by the Commission, a Member State deems it necessary to introduce national provisions based on new scientific evidence relating to the protection of the environment or the working environment on grounds of a problem specific to that Member State arising after the adoption of the harmonisation measure, it shall notify the Commission of the envisaged provisions as well as the grounds for introducing them.

6.

The Commission shall, within six months of the notifications as referred to in paragraphs 4 and 5, approve or reject the national provisions involved after having verified whether or not they are a means of arbitrary discrimination or a disguised restriction on trade between Member States and whether or not they shall constitute an obstacle to the functioning of the internal market.

In the absence of a decision by the Commission within this period the national provisions referred to in paragraphs 4 and 5 shall be deemed to have been approved.

7.

When, pursuant to paragraph 6, a Member State is authorised to maintain or introduce national provisions derogating from a harmonisation measure, the Commission shall immediately examine whether to propose an adaptation to that measure.”

Thus the Article provides that where the Union adopts a harmonisation measure setting a standard for the whole of the EU, a Member State which wants still to retain its own different standard must notify the Commission of that intention and seek the Commission’s approval for it to derogate from the otherwise binding EU norm. It is accepted by DCMS that the Government did not notify the Commercial Use Restriction to the Commission under Article 114 or its predecessor after the Authorisation Directive came into effect.

The Claimants accept that once the Amended Article 5 came into force, there was no need to notify the restriction. Once the text of the EU measure itself empowers the Member State to rely on a particular ground as a basis for adopting one regime rather than another, there is no need for the Member State to notify the Commission under Article 114 if it exercises that power. The Amended Article 5, in referring to ‘other objectives of general interest as defined by Member States in conformity with Community law’ sufficiently incorporates a public security justification into the Directive for this to be capable of being relied on without the need to comply with Article 114. But under the Original Article 5 where there is no mention of public security, the Member State is relying purely on the application of Article 52 TFEU.

DCMS deny that there was any need to notify the Commercial Use Restriction to the Commission under Article 114 TFEU or its predecessor provision. They say that the public security ground is incorporated into the Original Article 5 because the wording makes clear that the risk of harmful interference is not the only possible reason for making use of radio frequencies subject to an individual rights regime rather than a general authorisation. The Original Article 5 uses the words ‘in particular’ when referring to the risk of harmful interference as a factor, indicating that other factors can also be relevant. These other factors include, DCMS submit, public security.

In my judgment, public security was a factor that a Member State was entitled to take into account when deciding whether to issue a general authorisation or adopt an individual licensing regime under the Original Article 5. The drafting of that Article incorporating the words ‘in particular’ makes clear that harmful interference was not intended to be the only relevant factor. As to what other factors a Member State can take into account, one must look to the other provisions of the Directive and to EU law more generally. There are references in the Authorisation Directive to public security grounds: see Article 3(1) which refers to Article 52 TFEU.

The Amended Article 5 is also of assistance here. Although it may be unusual to use an amendment to a provision as an aid to interpreting the original version, one intention of the amendment to Article 5 in 2009 was to spell out the content of the previously vague ‘in particular’ of the Original Article 5. That clarification – omitting ‘in particular’ and listing so far as possible the reasons why a Member State can adopt an individual licensing regime – is indeed ‘better regulation’ as promised by the 2009 amendment. I note that Recital (25) of the Better Regulation Directive states that:

“(25)

Radio spectrum policy activities in the Community should be without prejudice to measures taken at Community or national level, in accordance with Community law, to pursue general interest objectives, in particular with regard to content regulation and audiovisual and media policies, and the right of Member States to organise and use their radio spectrum for the purposes of public order, public security and defence.”

In my judgment, therefore, a possible public security justification was available to Member States under the Original Article 5.

Further, my finding that the Original Article 5 itself contemplates that a Member State can rely on public security means that the obligation to notify the Commission under Article 114 TFEU is not triggered. That notification procedure is designed to forestall infraction proceedings where the national legislation is prima facie inconsistent with a harmonisation measure adopted by the Union. This is apparent from the case law on Article 114 to which the parties referred me; Case C-41/93 Re Pentachlorophenol: France v Commission [1994] ECR I-1829, Case C-319/97 Kortas [1999] ECR I-3160 and Cases 439&454/05P Land Oberösterreich and Republic of Austria v Commission [2007] ECR I-7141. Those cases are different from the present case because in those, the EU harmonising measure itself represented the Union’s decision on what was necessary to protect public health in terms of the acceptable levels of pentachlorophenol in the France case, the safety of a particular food colourant in Kortas and the need to control GM crops in the Austria case. The more restrictive domestic legislation that the defendant Member State wanted to adopt or retain would have been a clear infringement of that state’s obligations to implement the relevant directive, unless the Commission had approved the derogation under Article 114. If, instead, the EU measures had allowed the Member States to legislate after making their own assessment of public health needs, there would be no need to notify that decision under Article 114 because the resulting domestic law would not be an infringement of the EU instrument.

In this case, since Article 5 in both its original and amended versions does contemplate that the Member State can base its decision on its own assessment of public security factors, the Member State does not need to comply with the procedure under Article 114 TFEU if it does so. It does not follow, of course, that the Member State’s assessment of public security needs in the context of implementing Article 5 is immune from challenge on the facts or on grounds, for example, that it is disproportionate. However, the Authorisation Directive does not purport to make an EU-wide assessment of public security concerns and leaves it to the Member State to make that judgement.

I therefore conclude that the public security justification was available to the United Kingdom under both the Original and Amended Article 5 and that DCMS are not precluded from relying on that justification by a failure to notify the Commercial Use Restriction to the Commission pursuant to Article 114 TFEU.

The next question is whether the way in which Article 5 of the Authorisation Directive was transposed into UK law made a public security justification available to DCMS as a matter of domestic law. Curiously, the way in which this was done only came to light once the trial of this claim had started. The CA 2003, as well as transposing many of the provisions of the CRF in its own provisions, made many amendments to the WTA 1949. Section 166 of the CA 2003 inserted section 1AA into the WTA 1949 as follows:

“166 Exemption from need for wireless telegraphy licence

After section 1 of the Wireless Telegraphy Act 1949 there shall be inserted—

“1AA Exemption from need for wireless telegraphy licence

(1)

If OFCOM are satisfied that the condition in subsection (2) is satisfied as respects the use of stations or apparatus of any particular description, they shall make regulations under section 1 of this Act exempting the establishment, installation and use of any station or apparatus of that description from the prohibition in that section.

(2)

That condition is that the use of stations or apparatus of that description is not likely to involve any undue interference with wireless telegraphy.”

That new section 1AA was in force from 25 July 2003 until 7 Feb 2007. On that day it was replaced by section 8 of the WTA 2006. That provided:

‘8 Licences and exemptions

(1)

It is unlawful—

(a)

to establish or use a wireless telegraphy station, or

(b)

to instal or use wireless telegraphy apparatus,

except under and in accordance with a licence (a “wireless telegraphy licence”) granted under this section by OFCOM.

…

(3)

OFCOM may by regulations exempt from subsection (1) the establishment, installation or use of wireless telegraphy stations or wireless telegraphy apparatus of such classes or descriptions as may be specified in the regulations, either absolutely or subject to such terms, provisions and limitations as may be so specified.

(4)

If OFCOM are satisfied that the condition in subsection (5) is satisfied as respects the use of stations or apparatus of a particular description, they must make regulations under subsection (3) exempting the establishment, installation and use of a station or apparatus of that description from subsection (1).

(5)

The condition is that the use of stations or apparatus of that description is not likely to involve undue interference with wireless telegraphy.’

Section 8 of the WTA 2006 was then amended by the 2011 Regulations (Schedule 2, paragraph 4(b) and (c)) to reflect the Amended Article 5(1) of the Authorisation Directive as substituted by the Better Regulation Directive. Section 8 (4) and (5) of the WTA 2006 then read (amendments in italics):

‘8 (4) If OFCOM are satisfied that [the conditions in subsection (5) are] satisfied as respects the use of stations or apparatus of a particular description, they must make regulations under subsection (3) exempting the establishment, installation and use of a station or apparatus of that description from subsection (1).

[(5) The conditions are that the use of stations or apparatus of that description is not likely to —

(a)

involve undue interference with wireless telegraphy;

(b)

have an adverse effect on technical quality of service;

(c)

lead to inefficient use of the part of the electromagnetic spectrum available for wireless telegraphy;

(d)

endanger safety of life;

(e)

prejudice the promotion of social, regional or territorial cohesion; or

(f)

prejudice the promotion of cultural and linguistic diversity and media pluralism.]’

The Claimants point out that there is no mention of public security in any of this. They say that the upshot of this transposition is that even if the Original or Amended Article 5(1) did allow a Member State as a matter of EU law to rely on public security as a reason for imposing an individual licensing regime, the UK never took advantage of that option by incorporating it into domestic law. It was therefore not open to DCMS to rely on that when deciding to maintain in force the Commercial Use Restriction.

DCMS objected to the Claimants’ reliance on section 1AA of the WTA 1949 on the basis that the point was not pleaded and arose very late in the day. A note was produced for me by the Claimants seeking to show how the point did fit in their pleadings. DCMS did not contest that section 1AA was in fact enacted and remained in force for several years. I regard it as unattractive for the Government to raise a pleading point asserting that it is taken by surprise when the Claimant unearths a statutory provision that the Government itself enacted in purported compliance with its EU obligations. Section 1AA WTA 1949 is a relevant piece of this complex statutory jigsaw and it seems to me unfair for DCMS, as the department responsible for this legislation, to contend that I should ignore its existence in considering these issues because DCMS were unaware of this section.

DCMS put forward two arguments of substance as to why they can rely on a public security justification as a matter of domestic law. First they rely on the principle derived from Case C-106/89 Marleasing v La Comercial Internacional de Alimentacion [1990] ECR I-4135 to interpret section 1AA WTA 1949/ section 8 WTA 2006 as including that justification, since I have found that it was available under Article 5 of the Authorisation Directive as a matter of EU law. It is necessary, they say, to give a purposive interpretation to section 1AA WTA 1949 since it was intended to implement Article 5(1) of the Authorisation Directive and should be read as if it covered the same ground. DCMS also referred to HMRC v IDT Card Services Ireland [2006] EWCA 29 where Arden LJ considered in detail the scope of the Marleasing principle.

Secondly, DCMS rely on section 5 of the CA 2003 as an additional string to the UK Government’s bow as regards the transposition of both versions of Article 5 of the Authorisation Directive. Section 5 of the CA 2003 provides:

‘5 Directions in respect of networks and spectrum functions

(1)

This section applies to the following functions of OFCOM —

(a)

their functions under Part 2; and

(b)

their functions under the enactments relating to the management of the radio spectrum that are not contained in that Part.

(2)

It shall be the duty of OFCOM to carry out those functions in accordance with such general or specific directions as may be given to them by the Secretary of State.

(3)

The Secretary of State’s power to give directions under this section shall be confined to a power to give directions for one or more of the following purposes —

(a)

in the interests of national security;

…

(4)

The Secretary of State is not entitled by virtue of any provision of this section to direct OFCOM to suspend or restrict—

(a)

a person’s entitlement to provide an electronic communications network or electronic communications service; or

(b)

a person’s entitlement to make available associated facilities.

(5)

The Secretary of State must publish a direction under this section in such manner as appears to him to be appropriate for bringing it to the attention of the persons who, in his opinion, are likely to be affected by it.’

The term ‘enactments relating to the management of the radio spectrum’ in section 5(1)(b) is defined in section 405(1) of the CA 2003 as including the WTA 1949. If the public security justification is not included in section 1AA/section 8, then, DCSM contend, that aspect of Article 5 of the Authorisation Directive was transposed as part of the power conferred on the Secretary of State under section 5 of the CA 2003.

The Marleasing principle does not assist DCMS on this issue. That principle is relied on to enable the court to interpret domestic legislation in a way which avoids that legislation being incompatible with the EU rules it is intended to implement. Here it could not be suggested that it was incompatible with Article 5(1) of the Authorisation Directive to transpose the provision in a way which precluded reliance on grounds other than the risk of harmful interference as reasons for imposing an individual licensing scheme.

I do not consider that the Marleasing principle requires or entitles me to read the words ‘for example’ or ‘in particular’ into section 1AA(1) or section 8(4) so as to make the conditions set out in the following provision non-exhaustive. This would do considerable violence to the provision and introduce a lack of precision that was not intended. I note that when transposing the Amended Article 5 into section 8 of the WTA 2006, the draftsman did not transpose the rather uncertain wording of the final bullet point of that Article (‘other objectives of general interest as defined by Member States in conformity with Community law’) into domestic law but spelled out what was meant in section 8(5)(d) – (f).

However, I accept that section 5 of the CA 2003 does have the effect contended for by DCMS and does allow the UK to rely on a public security justification in relation to the making of exemptions under section 1AA/section 8. Indeed, it is also more likely, in my judgment, that the legislative intention was that decisions about the needs of national security would be placed in the hands of the Secretary of State rather than of OFCOM.

The Claimants countered with three points on the potential application of section 5 of the CA 2003. They argue that a direction made under section 5 could not be used to override the duty imposed on OFCOM in primary legislation such as the duty to issue an exemption imposed in section 1AA/section 8. I do not see why this should be the case. Sections 5 and 405(1) of the CA 2003 (which defines which OFCOM functions the power relates to) contain no such limitation.

The Claimants refer to section 5(4) which, they say, precludes the Secretary of State from exercising the section 5 power to restrict use of electronic communications services. However, I read section 5(4) as preventing the Secretary of State from directing OFCOM to restrict or suspend the rights of a specific individual or company, not as preventing him or her from directing that a restriction be imposed generally.

The Claimants also submit that even if the Secretary of State could have given a direction under section 5 of the CA 2003 to OFCOM to exercise its power under section 1AA/section 8 to impose an individual licensing regime on GSM Gateways, the Secretary of State did not in fact do so. DCMS accept that there is no evidence that the Secretary of State ever made a direction under section 5 to this effect.

On this point I agree with the submissions of DCMS that such a direction would only have been necessary if and when the 2003 Exemption Regulations were re-made in exercise of the powers under section 1AA/section 8. The 2003 Exemption Regulations were made on 20 January 2003 under the power in section 1(1) WTA 1949 before section 1AA was inserted into the WTA 1949 on 25 July 2003 (by section 166 of the CA 2003). It is not suggested that the Regulations were ultra vires that power (i.e. the power in section 1(1) WTA 1949) when they were made.

The 2003 Exemption Regulations were then continued in force by the transitional provisions set out in Schedule 18 to the CA 2003 which provides in paragraph 1 as follows:

‘(1) This paragraph applies where, at any time before the coming into force of a transfer made by virtue of section 2 —

(a)

any subordinate legislation has been made in the carrying out of the transferred functions by the person from whom the transfer is made; or

(b)

any other thing has been done by or in relation to that person for the purposes of or in connection with the carrying out of those functions.

(2)

The subordinate legislation or other thing—

(a)

is to have effect, on and after the coming into force of the transfer, and so far as necessary for its purposes, as if it had been made or done by or in relation to OFCOM; and

(b)

in the case of subordinate legislation to which section 403 applies when it is made by OFCOM, shall so have effect as if made in accordance with the requirements of that section.

(3)

Where any subordinate legislation, direction, authorisation or notice has effect in accordance with this paragraph—

(a)

so much of it as authorises or requires anything to be done by or in relation to the person from whom the transfer is made is to have effect in relation to times after the coming into force of the transfer as if it authorised or required that thing to be done by or in relation to OFCOM; and

(b)

other references in the subordinate legislation, direction, authorisation or notice to the person from whom the transfer is made are to have effect, in relation to such times, as references to OFCOM.’

The 2003 Exemption Regulations were validly made by the Secretary of State under the power in section 1(1) WTA 1949 and are treated by virtue of paragraph 1 of Schedule 18 to the CA 2003 as having been made under that power by OFCOM. In so far as the Commercial Use Restriction depended for its validity on a public security justification, it was capable of being made under section 1(1) WTA 1949 despite the insertion of section 1AA in 2003 because the CA 2003 also conferred a power on the Secretary of State under section 5 to give a direction to OFCOM to make the exemption subject to that restriction. The 2003 Exemption Regulations could therefore have been validly granted in their current form at any point in the relevant legislative history. There was no need to remake the 2003 Exemption Regulations after the coming into force of the CA 2003 and so no need for the Secretary of State actually to make a direction under section 5 of the CA 2003.

I can therefore summarise my findings on the question of whether it is open to DCMS to justify the restriction on the use of GSM Gateways in the 2003 Exemption on the grounds of public security as follows.

This is rather more straightforward than the arguments over public security grounds. The Original Article 5(1) of the Authorisation Directive does not mention efficient use of spectrum as a reason why a Member State can impose an individual licensing regime rather than issuing a general authorisation. Is the efficient use of spectrum nonetheless intended to be a possible ground, given that the words ‘in particular’ in the Original Article 5(1) indicate that avoiding harmful interference is not intended to be the only ground? In my judgment it is for the following reasons.

First, the Original Article 5(2) refers to rules ensuring the efficient use of radio frequencies as being part of the scheme of individual licensing that a Member State is entitled to adopt under Article 5(1):

‘2. Where it is necessary to grant individual rights of use for radio frequencies …, Member States shall grant such rights, upon request, to any undertaking …, subject to … rules ensuring the efficient use of those resources in accordance with Directive 2002/21/EC (Framework Directive).’

Original Article 5(5) provides that Member States can only limit the number of rights to be granted where this is necessary to ensure the efficient use of radio frequencies having considered the various factors set out in Article 7.

It would be odd to enable the grant of individual rights to be subject to rules to ensure efficient use of spectrum but not to enable a Member State to adopt individual licensing in order to be able to apply those rules. It also does not make sense to provide that the efficient use of spectrum provides a justification for limiting the number of individual licences but does not provide a justification for adopting an individual licensing scheme in the first place.

Secondly, the need efficiently to manage spectrum is at the centre of the CRF more generally. Recital (19) of the Framework Directive states that “It is important that the allocation and assignment of radio frequencies is managed as efficiently as possible”. Recital (11) of the Authorisation Directive which appears by its position within the recitals to relate particularly to Article 5 states:

‘(11) The granting of specific rights may continue to be necessary for the use of radio frequencies and numbers…. Those rights of use should not be restricted except where this is unavoidable in view of the scarcity of radio frequencies and the need to ensure the efficient use thereof.’

I therefore find that it is open to a Member State under both the Original and Amended Article 5 to decide to impose an individual licensing regime on the grounds that it is necessary to ensure the efficient use of spectrum.

The same question arises now as to whether the way that the Original Article 5 was transposed by the insertion of section 1AA into the WTA 1949 by the CA 2003 made ensuring the efficient use of spectrum available to DCMS as a justification for imposing the Commercial Use Restriction as a matter of domestic law.

I have already explained why I reject DCMS’ submission that one can achieve this result by relying on the Marleasing principle of statutory construction to read more conditions into section 1AA(2) WTA 1949 (or into the original version of section 8 of the WTA 2006) than the single condition that appears there.

For DCMS, there is no silver bullet here akin to section 5 of the CA 2003 on the public security ground. Instead, DCMS point to other provisions of domestic law which emphasise the importance of efficient use of spectrum in the exercise of OFCOM’s functions. For example, section 3(2) of the CA 2003 stipulates that among the things which OFCOM are required to secure in the carrying out of their functions is “the optimal use for wireless telegraphy of the electro-magnetic spectrum”. I doubt, however, that this duty can override the clear wording of section 1AA WTA 1949 which obliges OFCOM to issue an exemption unless there is a risk of harmful interference.

DCMS also point to sections 154 and 164 of the CA 2003 which provide:

‘154 Duties of OFCOM when carrying out spectrum functions

(1)

It shall be the duty of OFCOM, in carrying out their functions under the enactments relating to the management of the radio spectrum, to have regard, in particular, to–

(a)

the extent to which the electro-magnetic spectrum is available for use, or further use, for wireless telegraphy;

(b)

the demand for use of that spectrum for wireless telegraphy; and

(c)

the demand that is likely to arise in future for the use of that spectrum for wireless telegraphy.

(2)

It shall also be their duty, in carrying out their functions under those enactments to have regard, in particular, to the desirability of promoting–

(a)

the efficient management and use of the part of the electro-magnetic spectrum available for wireless telegraphy;

(b)

the economic and other benefits that may arise from the use of wireless telegraphy;

(c)

the development of innovative services; and

(d)

competition in the provision of electronic communications services.

(3)

In the application of this section to the functions of OFCOM under the enactments relating to the management of the radio spectrum … OFCOM may disregard such of the matters mentioned in the preceding subsections as appear to them–

(a)

to be matters to which they are not required to have regard apart from this section; and

(b)

to have no application to the case in question.

164 Limitations on authorised spectrum use

(1)

If they consider it appropriate, for the purpose of securing the efficient use of the electro-magnetic spectrum, to impose limitations on the use of particular frequencies, OFCOM must make an order imposing the limitations.

…’

Although I accept that these provisions reflect the importance generally placed on the efficient use of spectrum, I doubt that they could be relied on as overriding the duty in section 1AA WTA 1949, particularly having regard to section 154(3)(a).

DCMS make the point that once section 8(4) and (5) WTA 2006 had been amended by the 2011 Regulations to reflect the Amended Article 5(1), an efficient use of spectrum justification was clearly available as a ground for declining to issue an exemption under section 8(3). It would be very unsatisfactory, they say, for section 1AA WTA 1949 and the original version of section 8 of the WTA 2006 to be interpreted in such as way as to create a lacuna as between 25 July 2003 and 26 May 2011 when the amendments introduced by the 2011 Regulations came into effect.

Although I see the force of these arguments, given my findings later in this judgment on the facts of this case, I do not have to decide whether the 2003 Exemption Regulations could have been re-enacted using the powers in section 1AA WTA 1949 or section 8 WTA 2006 prior to 26 May 2011 or, if not, how this affects the validity of the 2003 Exemption Regulations in the light of Schedule 18 to the CA 2003. I do not therefore come to a decision on this point.

I can therefore sum up this part of my judgment as follows.

I now turn to the question whether the justifications put forward by DCMS and the Home Office for imposing the Commercial Use Restriction are made out on the facts. Both parties accepted that I should approach this question on the basis of the evidence as to the situation as it currently stands, rather than attempt to assess whether the justification was or was not made out during any particular period when the Commercial Use Restriction was in force. The evidence of the witnesses was accordingly directed at the current position.

In relation to each ground relied on by DCMS, my task is to consider whether I accept that the concerns expressed by DCMS on each ground are legitimate concerns and then to consider, if they are, whether the Commercial Use Restriction is a proportionate response to those concerns.

The case as regards public security was presented at the trial by Mr Beard QC appearing on behalf of the Home Office. Much of the evidence was heard in private during the trial and it was agreed at the end of the hearing that I would put the confidential parts of my analysis of the public security issues in a confidential annex to this judgment. Some preliminary matters can be set out here without trespassing into confidential material.

Mr Beard submitted that I should approach the evidence about public security concerns bearing two points in mind. The first is that EU law generally confers a margin of discretion on Member States in such matters. Where, as here, the question whether or not a Member State is in breach of EU law depends on that State’s assessment of its public security interests, a court should be slow to second guess the State’s assessment. DCMS relied on Case C-36/02 Omega [2004] ECR I-9609 where the Court of Justice considered a ban imposed by Germany on the sale of games which simulated acts of homicide. Although recognising that some Member States did not prohibit the sale of such games and that derogations from the freedom to provide services must be interpreted strictly, the Court upheld the Member State’s power to impose such a ban on public policy grounds. The Court said (omitting citation of authorities):

“31.

The fact remains, however, that the specific circumstances which may justify recourse to the concept of public policy may vary from one country to another and from one era to another. The competent national authorities must therefore be allowed a margin of discretion within the limits imposed by the Treaty”

I accept that that is the right way to approach this case though this does not mean that the court should not look closely at the evidence presented to see if the public security concerns are real and if the Government’s response to them is appropriate.

Allied to this margin of discretion is what Mr Beard called the ‘precautionary principle’ which is that a Member State does not have to wait until some harm arises to the public before adopting measures to prevent a recurrence of that harm. He referred to Case T-392/02 Solvay [2003] ECR II-4555 as setting out both principles:

‘121. The precautionary principle constitutes a general principle of Community law requiring the authorities in question, in the particular context of the exercise of the powers conferred on them by the relevant rules, to take appropriate measures to prevent specific potential risks to public health, safety and the environment, by giving precedence to the requirements related to the protection of those interests over economic interests. …

122.

It is settled case-law that, in the field of public health, the precautionary principle implies that, where there is uncertainty as to the existence or extent of risks to human health, the institutions may take precautionary measures without having to wait until the reality and seriousness of those risks become fully apparent.

…

125.

As regards the scope of the discretion of the relevant institution it should be noted that, where scientific evaluation does not make it possible to determine the existence of a risk with sufficient certainty, whether to have recourse to the precautionary principle depends on the level of protection chosen by the competent authority in the exercise of its discretion, taking account of the priorities that it defines in the light of the objectives it pursues in accordance with the relevant rules of the Treaty and of secondary law. That choice must, however, comply with the principle that the protection of public health, safety and the environment is to take precedence over economic interests, as well as with the principles of proportionality and non-discrimination …

126.

In that context, as regards the extent of the review by the courts of the implementation of the precautionary principle, it should be noted that it is settled case-law that where a Community institution is called upon to make complex assessments, it enjoys a wide measure of discretion the exercise of which is subject to a judicial review restricted to verifying that the measure in question is not vitiated by a manifest error or a misuse of powers and that the competent authority did not clearly exceed the bounds of its discretion. …’

I agree that the same principles apply to the issue of whether DCMS was justified in maintaining in force the Commercial Use Restriction on the grounds of public security. I have applied those principles in my consideration of the issues raised by the parties as regards the public security concerns raised by DCMS about the operation of GSM Gateways.

My conclusions on those issues are set out in the Confidential Annex to this judgment. Here I say only that my finding is that the public security concerns justify the imposition of the Commercial Use Restriction in so far as that restricts the provision of COMUGs but not in so far as it restricts the provision of COSUGs.

The Claimants accept that the risk of harmful interference is a ground available to a Member State for imposing an individual licensing regime rather than granting a general authorisation pursuant to both the Original and the Amended Article 5 of the Authorisation Directive. Article 2(2)(b) of the Authorisation Directive defines harmful interference as follows:

“‘harmful interference’ means interference which endangers the functioning of a radio navigation service or of other safety services or which otherwise seriously degrades, obstructs or repeatedly interrupts a radio communications service operating in accordance with the applicable Community or national regulations.”

The Claimants also accept that the way that the EU legislation was transposed into UK law enable OFCOM to rely on this justification when exercising its functions under the WTA. To recap: section 1AA of the WTA 1949 inserted by the CA 2003, and then section 8 of the WTA 2006, provided that OFCOM must grant an exemption from the need for a licence where they are satisfied that the relevant apparatus is not likely to involve ‘undue interference’ with wireless telegraphy.

Before July 2003, the domestic definition of “interference” and of when interference was ‘undue’ was set out in section 19(4)-(6) of the WTA 1949 as follows:

“(4)

In this Act, the expression "interference," in relation to wireless telegraphy, means the prejudicing by any emission or reflection of electro-magnetic energy of the fulfilment of the purposes of the telegraphy (either generally or in part, and, without prejudice to the generality of the preceding words, as respects all, or as respects any, of the recipients or intended recipients of any message, sound or visual image intended to be conveyed by the telegraphy), and the expression "interfere" shall be construed accordingly.

(5)

In considering for any of the purposes of this Act, whether, in any particular case, any interference with any wireless telegraphy caused or likely to be caused by the use of any apparatus, is or is not undue interference, regard shall be had to all the known circumstances of the case and the interference shall not be regarded as undue interference if so to regard it would unreasonably cause hardship to the person using or desiring to use the apparatus.

(6)

Any reference in this Act to the sending or the conveying of messages includes a reference to the making of any signal or the sending or conveying of any warning or information, and any reference to the reception of messages shall be construed accordingly.”

Sub-section 19(5) was replaced on 25 July 2003 by section 183 of the CA 2003 which modified the definition of “undue” in “undue interference” and provided:

“For subsection (5) of section 19 of the Wireless Telegraphy Act 1949 (c. 54) (meaning of undue interference) there shall be substituted—

“(5)

Interference with any wireless telegraphy is not to be regarded as undue for the purposes of this Act unless it is also harmful.

(5A) For the purposes of this Act interference is harmful if—

(a)

it creates dangers, or risks of danger, in relation to the functioning of any service provided by means of wireless telegraphy for the purposes of navigation or otherwise for safety purposes; or

(b)

it degrades, obstructs or repeatedly interrupts anything which is being broadcast or otherwise transmitted—

(i)

by means of wireless telegraphy; and

(ii)

in accordance with a licence under this Act, regulations under the proviso to section 1(1) of this Act or a grant of recognised spectrum access under Chapter 2 of Part 2 of the Communications Act 2003 or otherwise lawfully.”

Under section 115 (1) of the WTA 2006, “interfere” and “interference”, in relation to wireless telegraphy were to be construed in accordance with subsection (3) which provided:

“(3)

For the purposes of this Act, wireless telegraphy is interfered with if the fulfilment of the purposes of the telegraphy is prejudiced (either generally or in part and, in particular, as respects all, or as respects any, of the recipients or intended recipients of a message, sound or visual image intended to be conveyed by the telegraphy) by an emission or reflection of electromagnetic energy.

(4)

Interference with any wireless telegraphy is not to be regarded as undue for the purposes of this Act unless it is also harmful.

(5)

For the purposes of this Act interference is harmful if—

(a)

it creates dangers, or risks of danger, in relation to the functioning of any service provided by means of wireless telegraphy for the purposes of navigation or otherwise for safety purposes; or

(b)

it degrades, obstructs or repeatedly interrupts anything which is being broadcast or otherwise transmitted—

(i)

by means of wireless telegraphy; and

(ii)

in accordance with a wireless telegraphy licence, regulations under section 8(3) or a grant of recognised spectrum access or otherwise lawfully.”

The issues between the parties here are first what kinds of problems do GSM Gateways cause to the operation of the MNOs’ networks and secondly, do those problems fall within the definition of ‘harmful interference’ for the purpose of the Authorisation Directive and the domestic legislation?

Both parties submitted expert evidence on these issues. The Claimants instructed Professor William Webb. He has a doctorate in electronic engineering and is currently a director of Neul, a company developing machine to machine technologies and networks. From 2003 to January 2011, Professor Webb was a Director at OFCOM managing a team providing technical advice and conducting research across OFCOM’s regulatory remit. Before that he worked for various communications consultancies in the UK in the fields of hardware design, computer simulations and spectrum management. He is also now Visiting Professor at Surrey University and a member of OFCOM’s Spectrum Advisory Board.

DCMS relied on the expert evidence of Professor Simon Saunders. He also has a PhD in electronic engineering and worked throughout his career on wireless communications systems in industry, also as a Visiting Professor at Surrey University and as a design consultant.

Both men are clearly expert in their field and did their best to explain the inevitably complex technical position. The word ‘interference’ suggests to a lay person a crackling on the line which makes it difficult to hear what the other person is saying. Both experts defined ‘interference’ as covering a wider range of problems than that, including when calls are dropped part way through or cannot be made because of interference on the network. The key, both experts agreed, is that interference arises where the presence of another radio signal disrupts the reception of the wanted radio signal.

Without wishing in any way to minimise the complexity of the matters on which Professor Webb and Professor Saunders were giving evidence, I can summarise the facts which I understand are common ground between them.

First, there is no risk that the use of, say, a T-Mobile SIM card in a GSM Gateway will cause problems on the network of an MNO other than T-Mobile. The SIM cards make use of the bands of spectrum allocated to the MNO which issued the cards and cannot send signals on any other band of the spectrum allocated to another MNO. All interference caused by the use of a T-Mobile SIM card in a GSM Gateway arises on the T-Mobile network.

Secondly, it is clear to me on the evidence that the use of SIM cards in a GSM Gateway is likely to create problems for the network on which that card operates. This is because the MNOs configure their network equipment on the assumption that SIM cards will be used in the way in which they are typically used by the owner of a mobile phone. Such a user makes mixed use of the phone, with incoming and outgoing calls; calls or text messages to and from landlines or the networks of the other MNOs. Calls to or from mobiles tend to be shorter than land line calls and a mobile phone user does not, typically, use the phone constantly, making one call immediately after another all day long. The usage pattern of a SIM card installed in a GSM Gateway is very different. The SIM card only makes outgoing calls; the vast majority of these will be on-net calls (unless the minutes bundle on the SIM card does not distinguish in cost between on and off net calls), the SIM cards makes calls which originate on a landline and are typically longer in duration and, most important, the SIM card will make calls continuously, back to back over an extended period. A base station (Footnote: ⁷) operated by an MNO in a particular geographic cell is designed to cope with typical mobile phone use by a typical number of people at the same time. The presence of a large GSM Gateway in a cell can quickly occupy the majority of channels available for calls on that network in that location. This causes disruption for other people trying to use their mobile phones in the vicinity of a GSM Gateway because they are more likely to find that their call is dropped or blocked because there is no free channel available on the MNO’s base station – all the channels are occupied by calls from the SIM cards in the GSM Gateway.

Calls from a GSM Gateway differ from normal mobile phone use in other ways too. A GSM Gateway is always stationary, making all the calls from the same location. Most mobile phone calls are made at about 1.5 metres above the ground, by people walking around. GSM Gateways are often situated on the upper storeys of an office building and so make a large number of calls at a height that the MNO’s network is not designed to accommodate. The GSM Gateway may also be attached to what is called a ‘high gain antenna’, that is an antenna which focuses the signal emitted in a particular direction, towards the MNO’s base station rather than the usual multidirectional antenna incorporated into mobile phone handset. Not only does this also mean that the calls are made at a height above ground different from that of ordinary mobile phone calls, but the signal sent to the base station is much stronger and might overshoot the base station receiver and trespass into the territory of an adjacent cell and base station. Professor Webb said, and Professor Saunders accepted, that the difference between a high gain antenna and a normal mobile phone was like the difference between a torch and a light bulb.

There is plenty of evidence in the contemporaneous documents that at the time that GSM Gateways were operated by the GGOs prior to the Secretary of State’s decision confirming the Commercial Use Restriction, the MNOs complained to the GGOs of significant upset to the networks as a result of GSM Gateway traffic. For example, in the slides for an internal presentation for Vodafone in August 2003 the speaker refers to cell congestion and constant loading resulting from SIM cards in GSM Gateways. The speaker gives two recent examples; one where the user was asking for more capacity because the network could not accommodate more than 24 channels being used at the same time and one where the use by several companies of GSM Gateways in the same office block meant that it was impossible to make a normal mobile phone call in the building. I note also that at a meeting between representatives of Floe and the DTI (the precursor to BIS) on 31 January 2003, it is recorded that it was agreed that irresponsible siting of GSM Gateways was causing ‘hot spot headaches’ for the MNOs and a degradation of service quality to end users.

I have no difficulty in finding that the use of an MNO’s SIM cards in a GSM Gateway has the potential to cause congestion on that MNO’s network and this may result in other subscribers to that network who are trying to make or receive calls suffering dropped or blocked calls or experiencing significant degradation in the quality of the calls they manage to make.

Professor Webb did not really contest that these problems could and did arise from the use of GSM Gateways. He accepted that the term ‘interference’ on a network of, say, Operator A can be caused by activity on a radio system being operated by someone else, or by a non-radio system such as another electronic device which inadvertently emits radio waves, or by other radio devices being operated by Operator A itself as part of its radio system. The last of the three is known as ‘self-interference’. He goes on then to say which of these can qualify as ‘harmful interference’:

“From a radio engineering point of view all of these can be classed as interference and can degrade reception. However, from a regulatory viewpoint, self-interference is excluded. … The reason why this must be so can be seen from the following example. If Operator A decided to implement their system such that one of their transmitters interfered with another of their transmitters this would clearly not be a matter for the regulator but would be the choice of Operator A and their responsibility to resolve if it were problematic. It would be nonsensical for Operator A to cause self-interference then claim regulatory protection from the interference generated. Hence, when a regulator or regulatory document discusses interference, it excludes self-interference and is concerned only with interference generated by other uses of the radio spectrum or by non-radio devices. More specifically, it is the duty of the regulator to ensure that a licence holder does not suffer harmful interference by taking action against (typically illegal) transmission that cause such interference. If Operator A was causing self-interference and complained to the regulator, the regulator could hardly prosecute Operator A for illegally interfering with its own system. Hence the important conclusion that harmful interference cannot include self-interference.

Another important distinction is that interference is the simultaneous reception of both a wanted signal and an unwanted signal. Hence, in the case of, for example, a GSM base station, that has time slots which it dedicates to particular mobile devices, the use of one of these timeslots by the intended device cannot be classified as interference since there is only one (wanted) signal present. It is possible that higher demand for timeslots than availability can lead to congestion (defined below), but this is a completely different phenomenon than interference.

There are many different mechanisms of interference including signals transmitted within the bandwidth used by Operator A from Operator B (co-channel), signals transmitted in a neighbouring band by Operator B that cannot be effectively filtered by Operator A (adjacent-channel) and signals from different geographical regions that propagate into the regions covered by Operator A. For the purposes of the discussion here these distinctions are not important. E1 tab 1 page 5)”

As regards this case, Professor Webb said that the way base stations operate means that it is not possible for two calls to occupy the same bandwidth at the same time. So there can never be two signals from two different SIMs travelling on the same Operator A channel and disrupting each other. There may be congestion, in the sense that there are more SIM cards wanting to send signals than there are base station channels to accommodate them. But that would not be regarded, in his opinion, as ‘interference’ because there are not two signals present at the same time.

More broadly, his point is that none of the problems that the MNOs complain of as resulting from the use of SIM cards by GGOs is a matter for regulatory intervention because this was all “self-interference” and hence not “harmful interference” for the purposes of the Authorisation Directive or the domestic legislation. His argument was that the problem was caused by the way the MNOs have conducted their business. They issue and place on the market hundreds of SIM cards each entitling people to make thousands of minutes of calls on their network. The GGOs buy those SIM cards from the MNOs or from third parties to whom the MNO has sold them. The GGOs then take advantage of the bundle of minutes in a way which the SIM card is configured to permit but which is not the way that the MNO expects the minutes will in fact be used. For example, a SIM card may incorporate a bundle of minutes including several thousand very cheap on-net minutes per month. The MNO does not expect that any mobile user will in fact use more than a fraction of those minutes per month even though he has paid for them when buying the card. He certainly will not use them all up within the first few days of having the card. The GGO, however, uses all the on-net minutes it pays for when buying the SIM card and often uses them up all in one go. If the fact that someone uses the SIM bundle to the full extent allowed by the MNO causes disruption to the MNO’s network, then, says Professor Webb, that is something that the MNO can correct either by expanding its network to accommodate the additional traffic or changing the bundles of minutes included on the SIM cards.

As regards the option of expanding the network to accommodate the extra traffic, Professor Webb points to the example of what happened when smart phones were introduced. Smart phones operate using a different part of the spectrum from GSM Gateways and mobile phones but the example is instructive nonetheless in showing the response of MNOs to unexpected increases in volumes of traffic. A smart phone generates about 24 times more traffic than an ordinary mobile phone. He cites evidence that smart phones have increased overall traffic levels (voice and data) some ten fold (1000 per cent) since they were introduced. This did not lead the MNOs to call for restrictions on the sale of smart phones because of the resulting congestion on their networks. The growing use of smart phones was in the commercial interests of the MNOs and so they changed their networks to accommodate the extra traffic.

Professor Webb says that self-interference is not the concern of a regulator. If a householder complains to the local authority that his neighbour’s radio is constantly playing loud music, he may expect the local authority to take action. If he complains to the local authority that his own radio is constantly playing loud music, he cannot expect the local authority to take action – he should turn down the volume himself.

Professor Saunders takes a different view. He notes that the MNOs configure their networks by carefully balancing the maximum use of the spectrum allocated to them with the need to maintain a high quality of service in terms of call connections and sound quality. He also defines ‘interference’ as a ‘phenomenon which may result in degradation of a “wanted radio signal” due to the presence of an “unwanted radio signal” (emphasis in original). He accepts that the probability of interference occurring is highest when the wanted and unwanted radio signals are carried over the same radio resource: a situation known as ‘co-channel’ interference. He also accepts that some interference is introduced by the MNOs in order to maximise the use of the spectrum allocated to them. In particular, MNOs tend to reuse frequencies between base stations in order to increase the capacity of their networks in a way that they can control and manage in order to achieve a high capacity in the available spectrum consistent with acceptable call quality and network coverage. He accepts that this could not be categorised as harmful interference because “Such interference is predictable to and can be controlled by the MNO”.

Where he differs from Professor Webb is in his characterisation of the use of SIM cards in GSM Gateways as being outside the MNO’s control and unpredictable. He says:

“….interference which is outside the MNO’s control and which has unexpected characteristics risks being harmful, because it creates an increased likelihood that acceptable service quality is degraded below the level which would occur in the absence of such interference and that such quality cannot be restored without significant effort and expenditure by the operator given the resources (e.g. spectrum, base station sites) available to them. This is my understanding of the term harmful interference.”

He describes the use made of SIM cards in a GSM Gateway as unconventional, for the reasons I have already described, in particular because they send a signal through a high gain antenna. He describes a GSM Gateway as operating as a ‘rogue device’ in the MNO network because the MNO cannot predict in which cell a SIM card is going to be operated making continuous on-net calls at full pelt. Even if the MNO wanted to expand its network to accommodate GSM Gateways, it would not know which cells need to be enhanced. It would be difficult for the MNO to plan and target extra capacity given their lack of knowledge or control of the planned deployment of GSM Gateways.

Both Professor Webb and Professor Saunders describe harmful interference as the presence of an ‘unwanted’ signal in the presence of a ‘wanted’ signal. Where they differ is in whether they consider signals sent by a GSM Gateway to the MNO’s base station as wanted or unwanted. Professor Webb considers that this term is a technical term so that an ‘unwanted’ signal is one which is not generated by a piece of equipment emanating from the MNO and/or which the MNO’s receiver equipment is not designed to pick up. So he would describe a signal from a different MNO’s network or from some other item of electronic equipment (such as a nearby laptop) or a flash of lightning as unwanted. He said:

“… the SIM cards within the gateway are under the control of the MNO. Every SIM card which is then inserted into a radio device works in accordance with the GSM Gateway specification and that specification allows the MNO to control the transmission of that particular SIM card. So each specific SIM card in the gateway is indeed under the control of the MNO in terms of whether it is allowed to transmit or not transmit.”

In other words, the SIM card is a genuine item produced and sold by the MNOs and doing what it is designed to do – making outgoing calls to the MNO’s base station. The MNO’s base station is designed to receive calls from that MNO’s active SIM cards. The signal is therefore ‘wanted’.

DCMS say that this is too narrow a definition of ‘wanted’ and ‘unwanted’. They say that while the MNOs have of course issued SIM cards, that cannot be a reason for claiming that the signal generated by a GSM Gateway is wanted in circumstances where such use is contrary to the stated policy of the MNOs; takes place without the knowledge or consent of the MNOs and causes congestion and harmful interference. As Professor Saunders put it in cross-examination:

“I think I am making a very simple point that if an MNO writes down in a policy that they do not want a certain signal, then it is probably unwanted”.

On this point I prefer the opinion and evidence of Professor Webb. The configuration and sale of bundles of minutes on SIM cards is a matter entirely within the control of the MNO. They devise the tariff packages in a way which they consider makes them attractive to mobile phone users. They also control (subject to regulatory constraints) the contractual terms under which those SIM cards are supplied to the market and then operated by the purchaser. For example, a T-Mobile tariff available on a SIM card in 2005 charged £15 per month which price included 3000 minutes (that is 50 hours of phone time). There is currently no contractual or technical requirement that the calls be, say, spread out across the month or that other minutes in the package must be used as well. It may well be that T-Mobile assumes, when deciding to sell that tariff, that no-one will actually make all the calls to which they are entitled from their mobile phone over the course of one or two days of the month. But I do not see that the MNO can complain to the regulator if the mobile phone user in fact does make all those calls to get the best value from the package even if that causes congestion on the MNO’s network because that pattern of use was unexpected. This is in effect what the GGOs are doing. I do not consider that the problems which undoubtedly arise for the MNOs from the use on their network of genuine SIM cards that they have sold can be described as ‘harmful interference’ or that the phone calls made from the SIM cards in a GSM Gateway are ‘unwanted’ in the sense used when describing interference as ‘harmful’.

A witness for VIP drew, for another aspect of this case, an analogy with a supermarket which decides to sell pints of milk as a loss leader. It expects that shoppers will be attracted to do their weekly shopping in the store and will buy a basket of profitably priced goods as well as one or two pints of cut price milk. If, however, the owner of the local convenience store regularly comes into the supermarket and only buys 300 cheap pints of milk to resell at a higher price from his own shop, leaving none for the ordinary customers, the supermarket can respond in a number of ways (subject to any other regulatory constraints). It can greatly increase the stock of cheap pints of milk so as to accommodate the demand from ordinary shoppers as well as the demand from the convenience store; it can limit the offer to two pints per customer; it can refuse to sell milk to the convenience store owner on the grounds that the store is a retail not a wholesale outlet or it can simply abandon the loss leader pricing policy for milk entirely.

In my judgment, a similar situation has been reached here. The MNOs sell SIM cards containing many cheap on-line minutes because they expect that the typical user will not use all those minutes or at least will use them in conjunction with other more profitably priced aspects of the bundle (such as receiving calls that earn the MNO money from MCT charges, making M2F calls or sending texts). When it discovers that a GGO regularly uses all the cheap on-net calls without generating any other revenue it is open to the MNO (subject to any other regulatory or competition law constraints) to respond. It can expand its network capacity to accommodate the additional traffic caused by GSM Gateways or change its distribution chain to ensure that SIM cards are only supplied to normal mobile phones and not to GGOs or change the bundles on the SIM cards either to prevent the kind of use that the GGOs make of the network or to render that use sufficiently profitable. What it should not be able to do is rely on the regulator to prevent someone from making full use of a genuine SIM card that has been placed on the market by the MNO itself and paid for by the GGO.

The difficulty with Professor Saunders’ definition of ‘unwanted’ signals as including phone calls made from GSM Gateways is that it means that the definition of ‘harmful interference’ is dependent on the commercial policies of the MNOs. For as long as the MNOs do not want the GSM Gateways operating on their networks, any problems caused by their operation amount to harmful interference for the purposes of Article 5 of the Authorisation Directive and section 1AA WTA 1949 or section 8 WTA 2006. The regulator is to that extent empowered to impose an individual licensing regime rather than grant a general authorisation or exemption for GSM Gateways. Aside from this being a very curious result, it also creates problems if different MNOs arrive at different commercial conclusions, or change their minds over time as regards their attitude to GSM Gateways on their networks.

I do not accept Professor Saunders’ characterisation of the use of SIM cards in GSM Gateways as being ‘outside the control’ of the MNOs in the same way as, for example, the presence of an entirely extraneous signal from another MNO or different electronic device. The MNOs control the configuration of the SIM card and the terms on which the minutes are used on their networks. If people use the SIM cards in breach of the terms under which they are provided then that is a matter for the MNOs to police by cutting off the SIM card and enforcing their contractual rights against the purchaser. It is not something that the regulator should enforce on the MNOs’ behalf. It may be that the MNOs organised the tariffs and distribution of their SIM cards in a way which meant that there were many more minutes out in the market than they can comfortably accommodate on their networks and that they failed to keep control over to whom they were sold or the way in which they could be used. That is a result of their commercial decisions and the remedy, if there needs to be a remedy, lies with them and not with the regulator.

I reject DCMS’ submission that the use of SIM cards on an MNO’s network causes ‘objective spectrum management issues which it is OFCOM’s duty to investigate’. The only people harmed by an MNO’s failure to control the use of SIM cards on its network are the other subscribers to that MNO’s network. If the MNO considers that it is in its commercial interests to reduce or eliminate the use of certain SIM cards on its network then it has the contractual means to do so. If the MNO does not want to expand its capacity it can change its tariff structure so that, for example, the 3000 inclusive minutes can only be used at the rate of 100 minutes per day. It has the technical means to enforce this in just the same way as it enforces the change in tariff for minutes used in excess of the inclusive limit.

DCMS highlight what they say is an inconsistency in the evidence Professor Webb gave in cross-examination. Professor Webb accepted that some interference arising on the MNO’s own network and generated by equipment using that MNO’s SIM card could amount to harmful interference. This would arise if the device into which the card was inserted was itself illegal because it emitted a signal in excess of the interface standard with which the MNOs’ receiver equipment must comply. Professor Webb accepted that if the equipment was not operating in accordance with the interface standard then it would not be under the full control of the MNO and therefore had the potential to cause harmful interference. He did not accept that this was the effect of using a high gain antenna. His evidence was that if the signal coming from the high gain antenna was so strong that it risked spilling over into an adjacent base station’s cell and causing problems with calls there, that will be remedied automatically because the base station is designed to send a signal instruction back to the antenna to turn down the volume, as it were, so that the signal does not cause co-channel interference. He said it was highly unlikely that a high gain antenna would be used in a situation where its power could not be turned down sufficiently to avoid these problems.

I do not see that Professor Webb’s acceptance that he would regard interference caused by the use of equipment operating illegally outside the interface standard on an MNO’s network as within the definition of ‘harmful interference’ as inconsistent with his rejection of the characterisation of GSM Gateways as causing ‘harmful interference’. No one is suggesting that the any of the equipment used by the GGOs fails to comply with whatever standards are appropriate or that the reason why the MNOs’ networks cannot accommodate with them is that they are emitting signals that the MNOs’ receivers are not designed, technically speaking, to receive.

I therefore find that there is no evidence here that the GSM Gateways cause ‘harmful interference’ for the purposes of Article 5 of the Authorisation Directive or the relevant provisions of the WTA. There is therefore no basis for imposing the Commercial Use Restriction on the ground that it is needed to prevent the risk of harmful interference.

Professor Saunders explained how the use of the spectrum by F2M and M2M calls differs. A F2M call originates from a fixed telephone which is connected to the PSTN (public switched telephone network). The PSTN identifies the called number as being associated with a particular MNO and routes the call through a point of interconnection with that MNO. The call is then established between the fixed line and the intended mobile phone once the SIM has been authenticated and the necessary radio resource allocated to the call. Such a call would use one pair of uplink and downlink time slots. Professor Saunders described this as using one ‘radio resource’. When a GSM Gateway is interposed, the person placing the call on a fixed phone on the PSTN dials the call in a way which identifies that it should go through the gateway (for example by prefixing the number with specific digits) or the device may automatically recognise that this is a F2M call and route it through the gateway. When the GSM Gateway routes the call, the call occupies two radio resources in the same mobile network, one associated with the call as in the normal F2M call and one associated with the call from the GSM Gateway. This call, like any normal M2M call, uses two radio resources. Professor Webb accepted that a F2M call via a gateway uses two spectrum ‘legs’ rather than one.

DCMS say that GSM Gateways therefore make inefficient use of the spectrum. This is of itself a justification for imposing the Commercial Use Restriction on GSM Gateways. As Professor Saunders put it in answer to questions from Ms Carss-Frisk:

“A. I think any reasonable measure of efficiency is going to compare the amount of calls made to the amount of spectrum used. In the world of GSM gateways, the same number of calls are made but more spectrum is used. For me that is less efficient.

Q. I am reminded that of course the number of calls would depend on the price so that might vary?

A. I understand that there are supply and demand questions to be discussed and pricing and so forth. That seems to me an economic question which I am surely not qualified to opine on. But if we are talking about the efficiency of spectrum in engineering terms, the situation seems quite clear to me. I am going to divide something by something else. That is what an efficiency measure will be and good efficiency is delivering a lot of calls and bad efficiency is doing it with too much scarce resource. The scarce resource in question is the spectrum for me, so if we deliver the same number of calls, if we deliver the same number of calls, with twice as much spectrum, that seems to me approximately half as efficient.”

Professor Webb’s evidence was that OFCOM does not approach its task of promoting the efficient use of spectrum in such a simplistic way. I have already set out the provisions in the CA 2003 requiring OFCOM to promote the efficient use of the spectrum when exercising its functions. I note that Professor Webb did not purport to be an expert in economics or on telecoms regulation although he was previously a director of OFCOM. However, I set out what he says not because it is expert evidence on which I rely but because it concisely and helpfully expresses the view that I have come to:

“Efficient use of spectrum: This is an ill-defined term with no numeric parameters. Spectrum is often considered to be used more efficiently when more traffic is accommodated within a given amount of spectrum. However, alternative interpretations are for spectrum to be used more efficiently when its use generates greater economic value. Ofcom has a duty to promote efficient use of spectrum and has generally considered this to mean it should seek to maximise the economic value derived by the UK (often approximated as GDP contribution) from the spectrum. Note that any requirements to use spectrum efficiently are typically only relevant at a spectrum management level and do not cascade down to licence holders. For example, it would be unusual for the conditions in a spectrum licence to require the licence holder themselves to make efficient use of the spectrum. Instead, the regulator is expected to set a framework in place that results in licence holders making efficient use via incentives and similar. Ofcom have concluded that spectrum will be used efficiently (in an economic sense) if it is managed within a “market mechanism” framework where economic incentives apply to its use. Such incentives include auctions, trading and spectrum pricing. Ofcom believes that by effectively attaching a value to the spectrum it will encourage licence holders to treat it as a valuable resource and maximise the profit they can generate from it, which should generally maximise the economic value to the UK. Hence, Ofcom does not seek to measure or control “technical efficiency” of spectrum use.”

In my judgment it would not be rational for OFCOM to impose the Commercial Use Restriction on GSM Gateways, given the way in which it has interpreted its duty to promote the efficient use of spectrum over the course of its regulation of the telecoms industry. This is clearly illustrated by the facts that underlie this case. The MCT charges set by the MNOs have for several years been governed by price controls imposed by OFCOM in exercise of its powers under the CA 2003. Yet the puzzle of this part of DCMS’s case is that the MNOs, who have paid many millions of pounds for the right to use the spectrum, have priced it in the opposite way from the way indicated by the argument put forward by the Government. Given that F2M calls only use one leg of the MNO’s spectrum and M2M calls use two, one would expect the MCT charge for the M2M call to be double that of the F2M charge. In fact this whole case arises because the opposite is true: MNOs have charged more for the use of one leg than for the use of two. They do not price call termination in a way that encourages the least use of their spectrum. No doubt they do this because they want to encourage people to get into the habit of using their mobile phones as much as possible, since of course they earn revenue from those who originate calls on their network as well as from terminating calls. The MNOs have succeeded to the extent that some households now do not install a landline at all and people rely on their mobile phones even when they are sitting at home or at their office desk to make both M2M and M2F calls.

More significantly for this case, OFCOM has not taken steps in the price controls imposed over the years to prevent the MNOs from adopting this pricing practice on the grounds that it encourages an inefficient use of spectrum. They have taken a more sophisticated approach to the concept of efficient use of the spectrum than simply saying that using less is better. OFCOM auction the spectrum off to the highest bidder on the basis that the person bidding the most is likely to have the best plan for optimising the revenue that it can generate from use of the spectrum by selling services to its customers. To impose a restriction on the use of GSM Gateways because they use two legs rather than one would be inconsistent with the approach that OFCOM has taken in its regulatory approach to this issue.

DCMS also argue that if, as Professor Webb asserts, spectrum is used more efficiently when it generates greater economic value because that is likely to encourage licence holders, it makes sense to impose the Commercial Use Restriction because that prevents GGOs from undermining the revenue stream earned by the MNOs from terminating F2M calls. This is a thoroughly bad point. It is not the task of OFCOM to regulate the MNOs in a way which enables them to maximise their revenue in the hope that they will then bid large sums of money in the next spectrum auction. OFCOM has certainly never applied its price control powers to that end. Professor Webb’s response when the point was put to him was right: the spectrum is a resource that belongs to the country and it is OFCOM’s task to maximise the value that the country gains from the use of that natural resource. OFCOM does this by auctioning the spectrum and then trying to ensure that spectrum users either operate in competitive markets or are subject to price controls which mimic competitive markets by limiting charges to cost and incentivising cost reduction.

The Claimants argue that if the duty to promote spectrum efficiency meant encouraging people to use less spectrum, it is surprising that smart phones are not prohibited since they use much more spectrum than other devices. DCMS say that argument is not valid because the service that a user obtains on a smart phone is different from and better than services available on other equipment. GSM Gateways are different, DCMS say, because they use two legs of spectrum to provide exactly the same service as can be provided by one leg, namely an F2M call. I do not agree that the service is ‘the same’. In this context the service that is provided is at a cost and what the GSM Gateway delivers is an F2M call at a lower price. No doubt the customer would describe the GSM Gateway as increasing the efficiency of its business, reducing costs and allowing money saved on phone calls to be used elsewhere in the business.

This is not to say that the way that OFCOM goes about promoting efficient use of spectrum is the only interpretation of its CA 2003 duties that was open to it. But it would be perverse for OFCOM to restrict the use of GSM Gateways on the grounds that they use two legs rather than one leg of radio resource when the very raison d’etre of the GSM Gateway arises from the fact that the sellers of spectrum charge more for terminating a call that uses one leg of their spectrum allocation than for a call that uses two legs.

I therefore find that, even if it is open to OFCOM to decline to grant an exemption under section 8 of the WTA 2006 on the grounds of promoting the efficient use of spectrum, that ground would not be a rational ground for imposing the Commercial Use Restriction, given OFCOM’s regulatory approach to spectrum efficiency hitherto.

My conclusion on this section of the case is therefore:

As a result of these findings, my conclusion is that the Commercial Use Restriction is and was a failure by the UK to implement Article 5 of the Authorisation Directive in so far as it applied to COSUGs.

I have found that there was an infringement of EU law by the imposition of the Commercial Use Restriction albeit a more limited infringement than that alleged by the Claimants. For that reason, and in case I am wrong on my findings on the other issues in Part II of this judgment, I now turn to whether the criteria for State liability are satisfied in this case. The principles were first set out by the Court of Justice in Cases C-6&9/90 Francovich v Italian Republic [1991] ECR I-5357. That case concerned the failure of a Member State to implement a directive where, the Court held, the directive did not confer directly effective rights on individuals. The Court held that the full effectiveness of Community rules would be impaired and the protection of the rights which they grant would be weakened if individuals were unable to obtain redress when their rights are infringed by a breach of Community law for which a Member State can be held responsible. The principle whereby a State must be liable for loss and damage caused to individuals as a result of breaches of Community law is therefore inherent in the system of the Treaty. The Court then set out the conditions for State liability in the following terms:

“40 The content of those conditions is that the result prescribed by the directive should entail the grant of rights to individuals. The second condition is that it should be possible to identify the content of those rights on the basis of the provisions of the directive. Finally, the third condition is the existence of a causal link between the breach of the State’s obligations and the loss and damage suffered by the injured parties”.

In the later judgment of the Court of Justice in Cases C-46/93 & 48/93 Brasserie du Pêcheur and Factortame [1996] ECR I-1131, the question referred to the Court was whether the State’s liability to pay damages was limited to the situation where the breach of Community law involved a rule that was not directly effective. The Court held that there was no such limitation and that, to the contrary, the right to reparation is ‘the necessary corollary of the direct effect of the Community provisions whose breach caused the damage sustained’ (paragraph 22 of the judgment). Since the Treaty provisions relied on by the claimants in the two cases before the Court had direct effect, the Court held that breach of such provisions may give rise to reparation.

The Court also re-examined the conditions for State liability as they applied to a case where the breach was not a failure to take any steps to transpose a directive by the deadline set for transposition as had occurred in Francovich. Referring to the principles that govern the non-contractual liability of the Community institutions themselves, the Court noted that those principles take into account the complexity of the situations to be regulated, difficulties in the application or interpretation of the texts and in particular the margin of discretion available to the author of the act in question. From this source comes the principle that the Community cannot incur liability unless the institution concerned ‘has manifestly and gravely disregarded the limits on the exercise of its powers’ (paragraph 45 of the judgment). The Court then restated the conditions that must be satisfied before State liability is established:

“…Community law confers a right to reparation where three conditions are met: the rule of law infringed must be intended to confer rights on individuals; the breach must be sufficiently serious; and there must be a direct causal link between the breach of the obligation resting on the State and the damage sustained by the injured parties.”

As regards the first condition, the Court in Brasserie du Pêcheur held that it was satisfied in the case of the two free movement Treaty provisions relied on by the claimants.

As to when the second condition that the breach be ‘sufficiently serious’ was satisfied, the Court said at paragraph 56 that:

“The factors which the competent court may take into consideration include the clarity and precision of the rule breached, the measure of discretion left by that rule to the national or Community authorities, whether the infringement and the damage caused was intentional or involuntary, whether any error of law was excusable or inexcusable, the fact that the position taken by a Community institution may have contributed towards the omission, and the adoption and retention of national measures or practices contrary to Community law”.

As for the need for a causal connection, the Court held that it was for the national courts to determine whether there was a direct causal link between the breach and the damage.

I shall refer to the three conditions set out by the Court of Justice in Brasserie du Pêcheur as ‘the Francovich criteria’ since that is how they are generally referred to, and were referred to in the course of the trial.

The first criterion that the Claimants must show is satisfied is that the EU provision infringed in this case was intended ‘to confer rights on individuals’. There is no decision from the Luxembourg courts on the question whether Article 5 of the Authorisation Directive has direct effect. As to what the national court must do in such a case, the parties referred me to the decision of the Court of Appeal in Poole v Treasury [2007] EWCA Civ 1021. In Poole the claimants who were underwriting names at Lloyd’s claimed that the losses they had suffered were in part attributable to HM Treasury’s failure properly to implement the relevant EU Directive on insurance. HM Treasury disputed the claimants’ entitlement on the grounds, amongst others, that the relevant EU Directive would not, if properly implemented, confer on them the right which would have avoided the losses. Buxton LJ (with whom Jacob and Moore-Bick LJJ agreed) held that the judge at first instance had been right to formulate the test in terms of whether it was necessary in order to achieve the objective of the Directive, to confer the asserted rights upon the claimants. The claimants had argued that it was enough to show either (i) that some right, even if not the right asserted in the claim, would have been granted to the claimants if the directive had been properly implemented or (ii) that the Directive was, more generally, intended to protect subjects in relation to their economic welfare rather than simply being directed at the public good. Buxton LJ rejected both those formulations as wrong as a matter of principle and also precluded by EU and domestic authority. He went on to uphold the decision of Langley J that the insurance directive did not entail the grant of rights to the claimants as insurers. Although the claimants were insured as well as insurers themselves - because they took out reinsurance - the Court held that this was not enough to bring them within the class of people who might be entitled to rights under the directive.

The question for me therefore is whether Article 5 of the Authorisation Directive is intended to confer on these Claimants a ‘right’ to operate COSUGs without having to apply for an individual licence. In my judgment, it is. The whole thrust of the Authorisation Directive is carefully to circumscribe the obstacles that the Member State can erect in the path of someone who wants to provide an electronic communications service as defined in the Authorisation Directive. Not only does the Directive set out, in Article 5, the circumstances in which the Member State can insist on an individual licensing scheme rather than a general authorisation but it details what the minimum content of the general authorisation conferred by the Member State must be (Article 4); the way in which individual rights must be granted (Article 5(2)); the procedure to be followed when auctioning a limited number of rights (Article 7); the fees that can be charged (Articles 12 and 13) and the other obligations that can be imposed on users (for example requests for information under Article 11). All these provisions are aimed at ensuring that the Member State cannot undermine by regulation the freedom to provide electronic communications services referred to in Article 3(1) more than is absolutely necessary. The Claimants want to exercise that freedom to provide services and are, they say, prevented from doing so by a failure by DCMS properly to implement Article 5. I agree with the Claimants that the Directive would be robbed of any real effect if it were construed as creating only a regulatory framework.

DCMS argue that the provisions in the Authorisation Directive which permit a limit on the number of licences granted show that the Directive was not intended to confer rights. It cannot be said in advance that anyone had a right to use the radio frequencies under the Directive. I do not accept that that is relevant here. The right that the Claimants are asserting is not the right to obtain a licence in an auction for GSM Gateway licences held in accordance with the provisions of Article 7. The right they are asserting is the right to be able to provide services in accordance with a general authorisation and not to have to apply for an individual licence at all. If the Poole test means that claimants have to show that the EU rule infringed was intended to confer on them the right they are asserting, it must also mean that their claim cannot be defeated by showing that the rule did not confer on them some other, different, right which they do not assert.

I therefore find, applying the test in Poole, that the first criterion in Francovich is satisfied. Article 5 of the Authorisation Directive is intended to confer on the Claimants a right to provide services under a general authorisation if the option of imposing an individual licensing regime is not open, on the facts, to the Member State.

The test as to whether a breach of EU law is sufficiently serious to give rise to State liability to pay damages was considered recently by the Court of Appeal in R (Negassi) v Secretary of State for the Home Department [2013] EWCA Civ 151. Kay LJ reviewed the cases on this element of Francovich and distinguished between the situation where a Member State has failed entirely to implement a directive by the due date and the situation where a Member State “has endeavoured to effect transposition within the required time but has done so imperfectly”. Kay LJ cited the judgment of the Court of Justice in Case C-392/93 R v HM Treasury, ex parte British Telecommunications plc [1996] ECR I-1631. The Court of Justice reiterated the three Francovich conditions and said:

‘40. Those same conditions must be applicable to the situation … in which a Member State incorrectly transposes a Community Directive into national law. A restrictive approach to state liability is justified in such a situation, for the reasons already given by the Court to justify the strict approach to non-contractual liability of Community institutions or Member States when exercising legislative functions in areas covered by Community law where the institution or state has a wide discretion – in particular the concern to ensure that the exercise of legislative functions is not hindered by the prospect of actions for damages whenever the general interest requires the institutions or Member States to adopt measures which may adversely affect individual interests …

….

43.

In the present case, Article 8(1) [of the relevant Directive] is imprecisely worded and was reasonably capable of bearing, as well as the construction applied to it by the Court in this judgment, the interpretation given to it by the United Kingdom in good faith and on the basis of arguments which are not entirely devoid of substance … That interpretation, which was also shared by other Member States, was not manifestly contrary to the working of the Directive or to the objective pursued by it."

Kay LJ in Negassi went on to say:

“13.

Where does all this lead? In my judgment, it demonstrates that, although there will be some cases where a failure to transpose a specific provision at all by a required date may, without more, amount to a sufficiently serious breach, a bona fide attempt at transposition will attract a more nuanced approach. I am entirely satisfied that the breach of EU law with which we are concerned in the present case does not entitle Mr Negassi to say that he is automatically entitled to reparation. On any view, the United Kingdom's breach was unintentional. It arose from a genuine misapprehension of the true legal position. Whatever may be the reach of automatic entitlement, it does not extend to this case.

(2)

The multi-factorial test

14.

It follows that Mr Negassi’s claim for damages must be assessed by reference to the multifactorial test for sufficient seriousness, … In the domestic context, it was the subject of helpful guidance in the speech of Lord Clyde in R v Secretary of State for Transport, ex parte Factortame (No5), [2000] 1 AC 524, at pages 554-556. He identified the following as potential factors: (1) the importance of the principle which has been breached; (2) the clarity and precision of the rule breached; (3) the degree of excusability of an error of law; (4) the existence of any relevant judgment on the point; (5) whether the infringer was acting intentionally or involuntarily or whether there was a deliberate intention to infringe as opposed to an inadvertent breach; (6) the behaviour of the infringer after it has become evident that an infringement has occurred; (7) the persons affected by the breach or whether there has been a complete failure to take account of the specific situation of a defined economic group; (8) the position taken by one of the Community institutions in the matter. He added (at page 554B-D) that the application of the “sufficiently serious” test “comes eventually to be a matter of fact and circumstance”.

“No single factor is necessarily decisive. But one factor by itself might, particularly where there was little or nothing to put in the scales on the other side, be sufficient to justify a conclusion of liability.”

Both parties in this case are agreed that the ‘multi-factorial’ test is the right one. But they approach the application of that test in very different ways. The Claimants focus on the process by which the decision not to amend the 2003 Exemption Regulations but to leave the Commercial Use Restriction in place was arrived at by DCMS. They say that nowhere in the contemporaneous documents does it appear that the obligations of the Secretary of State (Stephen Timms MP at the time) were drawn to his attention. I do not accept that examining the process by which the decision was taken is the only factor that is relevant here. The reference in the authorities to the intentional or involuntary nature of the infringement shows that it is legitimate to consider how the decision to legislate was arrived at, but it is not the whole story.

The factors referred to in the case law that are relevant to the infringement found in this case are the clarity of the provision that is being implemented; the good faith or otherwise of the Secretary of State in arriving at the decision to retain the Commercial Use Restriction; the attitude of the Community institutions to the Member State’s transposition; the persons affected by the breach and whether there was a complete failure to take account of the situation of the GGOs.

I accept that the principle of liberalising the provision of electronic communications services and the free movement of goods and services are important principles for the purposes of the multi-factorial test.

It must be said, however, that Article 5 of the Authorisation Directive is drafted in terms that leave the circumstances in which a Member State can choose to impose an individual licensing regime rather than grant a general authorisation deliberately flexible. The Original Article 5 did not attempt to define exhaustively the grounds on which the Member State could rely, stating that the risk of harmful interference was a ground that could, ‘in particular’ be relied on. Even when the grounds were set out in more detail in the Amended Article 5, the fourth bullet point allowed Member States to grant individual rights to ‘fulfil other objectives of general interest as defined by Member States in conformity with Community law’. The lack of clarity in the Member States’ duties does not, in this case, arise from poor drafting but is introduced deliberately into the provision. Section 8 of the WTA 2006, as amended in 2011 to implement the Better Regulation Directive, interpreted that fourth bullet as covering endangering safety of life, prejudicing the promotion of social, regional or territorial cohesion or prejudicing the promotion of cultural and linguistic diversity and media pluralism (as well as the public security ground I have held applied by virtue of section 5 of the CA 2003).

The flexibility inherent in Article 5 of the Authorisation Directive is not, of course, unlimited. Further, once the Member State has decided to rely on a particular ground, it must do so in a proportionate and rational manner. I have held that DCMS failed to act proportionately or rationally in applying the Commercial Use Restriction to COSUGs while leaving Self-User GSM Gateways exempt. That does not, of itself, mean that the breach was manifest or sufficiently serious to justify the imposition of State liability. I note that the Court of Justice in R v HM Treasury, ex parte British Telecommunications plc said that a restrictive approach to liability should be taken in particular because of ‘the concern to ensure that the exercise of legislative functions is not hindered by the prospect of actions for damages whenever the general interest requires … Member States to adopt measures which may adversely affect individual interests’. That is all the more the case here where the focus of the submissions put to the Secretary of State by the GGOs, as I shall describe below, was not on exempting COSUGs but on lifting the Commercial Use Restriction for COMUGs as well.

I therefore consider that the lack of clarity of Article 5 in both its original and amended form and the nature of the infringement I have found are factors which militate against a finding of State liability.

The history of the Commercial Use Restriction is a long and complicated one. Some of it was set out in the witness statement of Simon Towler on behalf of DCMS. Mr Towler is a senior civil servant within the Home Office who has responsibility for these matters but was not involved in the actual decisions taken in relation to GSM Gateways.

I have set out earlier the origin of the provision enacted in the 2003 Exemption Regulations. As I mentioned there, the original provision was not designed with GSM Gateways in mind since the device only came to the attention of the regulators in 2002. On 31 July 2002 an official at OFCOM wrote to a departmental lawyer seeking advice on the legality of GSM Gateways. He raised a number of questions, in particular whether GSM Gateways were covered by the definition of ‘user stations’ in the 1999 Exemption Regulations; if so, did that mean that they were subject to the Commercial Use Restriction and, if so, what degree of formal consultation would be needed to change the regulations. He said that this last point:

“… is particularly pertinent as we believe that the intention of the current regulations was never to exclude this type of use, and in the case of the possible exclusion of public use was due to a mis-transposition of the [1997 User Station Exemption] during consolidation with other Regs.”

As to the question whether GSM Gateways fell within the definition of mobile user station, he pointed out that the devices are in fact fixed (in the sense of stationary) rather than mobile but that the word ‘mobile’ is used not to connote that the device moves around but that it uses the mobile transmit frequencies to connect to the network. On the point about public and private use, the official states that there is a ‘grey area’ of service providers using a gateway to carry third party customer traffic. He then says:

“We believe that any exclusion on public use was introduced accidentally when the current Exemption Order was created. The aim of the current SI was to combine the several individual Exemption SI’s… into one unified Exemption Order”

The Claimants placed great reliance on the reference here to the imposition of the Commercial Use Restriction on GSM Gateways being ‘accidental’. I do not see that the word, read in context, is open to criticism, at least from GGOs. What was acknowledged as being ‘accidental’ was the application of the carve out in regulation 4(2) to user stations when previously it had applied only to cordless phones (see paragraphs 24 onwards above). The 1999 Exemption Regulations had not been drafted to apply to GSM Gateways and the Claimants did not assert that OFCOM ought to have realised that they applied before 2002. The minute, in my judgment, sets in train precisely the kind of inquiry that the regulator should carry out when a new gadget swims into his ken, to paraphrase Keats. One first looks at how the current regulations as drafted in fact apply to it, albeit that they were not drafted with the new gadget in mind. Then one decides whether that is the right way to deal with the device or if not, how should the regulations be changed to accommodate it.

The response from the DTI lawyer to the 31 July 2002 minute was not before the court. What then occurred in August 2002 was a consultation exercise. An initial letter was sent to interested parties prior to the formal consultation. That set out the view held by OFCOM at the time which was that the 1999 Exemption Regulations did not apply to GSM Gateways at all because they were fixed devices. Hence they were not exempted and operation of the gateway without a licence would be unlawful. On 11 September 2002 a submission was put up to Mr Timms by OFCOM officials asking for his approval for a consultation. The submission indicated that OFCOM’s position was:

“… that the commercial and engineering implications of allowing these applications to connect to any of the mobile networks should be for the licensed Network Operators to consider and any eligible customer of the networks should be exempted from individual licensing. …”

It also recorded that other regulators had indicated that their view also was that ‘the equipment should be exempted because of the potential benefits to consumers in reduced call charges’.

A consultation document was published in November 2002. The submission to Mr Timms again seeking his approval, dated 1 November 2002, recommended that the definition of user station be changed to make clear that it did cover GSM Gateways and that the Commercial Use Restriction be lifted. The consultation document described the public/private use issue as follows:

“… if operators choose to connect customers to the network, does it matter if the traffic carried is a private or a public service? Where large-volume gateway systems might impact on network planning, operators could require users to declare such use before installation to allow for network configuration. In any case, [OFCOM] believes that relaxing the Exemption Regulations to permit public connections would give the operators a choice, and would bring potential benefits for consumers in terms of increased competition and reduction of call costs.”

The consultation therefore recommended the exemption of GSM Gateways with no Commercial Use Restriction.

The Claimants criticise the submissions to the Minister and the text of the consultation document on the basis that they do not refer to the Authorisation Directive or to Article 5 (which had been adopted in March 2002 and would need to be implemented by July 2003). The Minister did not have drawn to his attention, the Claimants say, the fact that the UK could not refuse exemption for GSM Gateways save on the grounds included in Article 5. This argument seems to be a reliance on form over substance. OFCOM were recommending liberalisation here on the grounds that greater competition and cheaper calls would benefit consumers and there appeared to be no good reason to maintain the Commercial Use Restriction in place. The fact that the documents did not expressly refer to the Authorisation Directive does not seem to me to be relevant in this context.

The November consultation document set a deadline of 21 February 2003 for people to respond. 33 responses were received including a response from Floe and a joint response from the MNOs. Officials held meetings with some of the key respondents. In a submission to Mr Timms on 13 March 2003, the officials summed up the responses. They noted that there was overwhelming support for the proposal to include GSM Gateways in the definition of ‘user stations’ in the 1999 Exemption Regulations and a majority of responses were in favour of removing the Commercial Use Restriction. However, the submission noted that ‘these responses must be weighed against Government and Network Operator concerns such as network capacity, network interference, interception and law enforcement issues’. The submission described the concerns about congestion and capacity problems raised by the MNOs when responding to the consultation. It also discusses the security concerns that I have described in the Confidential Annex to this judgment. The submission concludes:

“Notwithstanding the majority of responses being in favour of deregulation, it is felt that the engineering, commercial and security concerns outweigh these and it is recommended that you agree to retain the restriction against provision of third-party telecommunication services over exempted equipment.

Through this action there is a slight possibility that companies who are at present providing a third party service may claim that this action will cause a loss of business (and possibly jobs) by making the service they provide illegal. This point is arguable as such activity was illegal before the consultation exercise under Regulation 4(2). The consultation response simply confirms the situation and it should be noted that some of the Network Operators have already acted to terminate the connections of equipment they deem to be used illegally under the current Regulations.”

The Claimants again criticise this submission on the grounds that it does not draw to the Secretary of State’s attention the soon-to-be implemented Article 5 as a factor weighing in favour of exempting GSM Gateways. On the contrary, by advising the Minister that retaining the Commercial Use Restriction would amount to maintaining the legal status quo ante, the officials failed to draw the Minister’s attention to the imminent change in the legal landscape brought about by the CRF. I agree that it is surprising and perhaps regrettable that the submission did not set the Minister’s decision in the context of the new regulatory regime which places greater emphasis on liberalisation. But the submission does attempt to carry out precisely the balancing exercise that the Minister would be required to undertake pursuant to Article 5, albeit that it does not place the CRF in the scales tilting the balance towards deregulation.

The Minister initially indicated that he was content with the submission. But it appears that some adverse Press comment about the MNOs’ moves to shut down GSM Gateways led to the matter remaining open. On 28 March 2003 an internal Home Office minute urged Bob Ainsworth MP, then a junior Minister in the Home Office to write to Mr Timms because officials had heard that ‘he is changing his mind in response to lobbying’. Mr Ainsworth wrote to Mr Timms on 31 March 2003 praising the post-consultation decision to retain the Commercial Use Restriction and asking for reassurance that ‘those disappointed with the result should not lead to any review’ of that decision. No such assurance was given by Mr Timms. In his letter of 9 May 2003, he noted that the use of gateway services may present an opportunity to drive consumer costs down in the telecoms market. He acknowledged the concerns raised by the Home Office and said that he had asked his officials to examine whether there was any way forward which will also address the Home Officer’s concerns.

Whether Mr Timms had in mind the UK’s obligations under the Authorisation Directive (and DCMS point out that he had recently spent many hours in the House of Commons piloting the CA 2003 through its Parliamentary stages) or whether it was simply his deregulatory instincts that caused him to pause and rethink does not, to my mind, matter. What is clear is that the Minister did not take the objections to GSM Gateways put forward by his officials at face value. He recognised the importance of gateways as a lever for consumer benefit and set his officials to finding a solution which would allow GSM Gateways to operate commercially.

The documents then reveal a flurry of anxious activity, particularly at the Home Office, seeking to reinforce their arguments against deregulation. There is no doubt from this evidence that the Home Office adhered strongly to the view that they did not want GSM Gateways operating and they expressed this view forcefully to the Minister and DTI officials. The further submission was put up to Mr Timms on 7 July 2003. It recorded that unfortunately the further work carried out had not produced a satisfactory way forward. The arguments for and against removing the Commercial Use Restriction were said to be finely balanced. The submission described the spectrum efficiency problems (two legs of spectrum rather than one) and the Home Office’s concerns as the prime reasons for not liberalising GSM Gateways. This submission was comprehensive and detailed. It was followed by a meeting between Mr Timms and his officials on 8 July 2003. The minute of that meeting records that Mr Timms approved the recommendation in the submission, though at the same time asking his officials to carry out more work aimed at finding a way round the restriction for the GGOs.

The result of the consultation was announced by OFCOM on 18 July 2003. It stated that the definition of ‘user stations’ would be amended to cover GSM Gateways. This change, it explained, was not only directed at GSM Gateways but would also legitimise a number of applications that had developed outside the current regulatory framework and which needed to be exempt such as ATM banking machines, vending machines and credit card authorisation terminals. It would also allow self-use of GSM Gateways. As to the second proposal, the press release said:

“Many responses from small businesses also supported the second proposal, to remove the restriction in Regulation 4(2) on the carriage of third party traffic over exempt devices. However the benefits of this are mitigated by the fact that the operators’ ability to comply with their Regulatory requirements with regard to emergency calls and security concerns are impaired and that the resulting use of spectrum is very inefficient. After considerable discussion with manufacturers and users of Gateway equipment and considering technical and other information supplied by them, the Government concludes that the restriction must be retained.

Mobile Network Operators (“MNO’s”) licensed under the Wireless Telegraphy Act 1949 can use their own (or third party) equipment in accordance with their licences in order to provide a telecommunications service. In some circumstances, MNO’s may be able to consider purchasing products or services from Gateway Operators for use under the auspices of MNO licences. Although a commercial matter for the companies concerned, the Government encourages the MNOs and Gateway Operators to consider ways to address pragmatically existing uses of equipment that continue not to meet the requirement for exemption.”

The Press release also noted that the 1999 Exemption Regulations had in fact been replaced during the course of the consultation by the 2003 Exemption Regulations. The Claimants criticised this re-enactment of the exemption as pre-judging the outcome of the consultation. However, it is clear from the documents that the 2003 Exemption Regulations were needed for an entirely different purpose unconnected with GSM Gateways and that the consideration of the responses to the consultation about GSM Gateways was extensive and genuine. What emerges from this account of the consultation and the 18 July 2003 decision is that the Minister who took the decision considered the matter with great care and with the importance of liberalising the market for the benefit of consumers well in mind.

Subsequently, OFCOM decided that the definition of mobile user station in the 2003 Exemption Regulations did catch GSM Gateways so there was no need to amend the Regulations to implement the first recommendation arising from the consultation.

Towards the end of 2003, the Floe litigation commenced. One of the issues raised in that litigation was the incompatibility of the Commercial Use Restriction with the RTTE Directive and the Authorisation Directive. The CAT ruled in August 2006 that the restriction was in breach of both Directives but that the spectrum licences granted to the MNOs should be read as entitling them to grant sub-licences to GGOs enabling the latter to operate GSM Gateways on their networks. OFCOM appealed to the Court of Appeal which overturned the decision of the CAT holding that the MNOs’ licences did not have that effect. The Court of Appeal declined to consider the issue of compatibility with EU law saying that it had been unnecessary for the tribunal to become involved in questions of EU compatibility in order to decide the appeal. Those matters were, the Court said, entirely hypothetical.

The Claimants criticise DCMS for not acting on those parts of the CAT’s judgment which found a breach of EU law and which were not overturned by the Court of Appeal. Since my findings on the law have differed from those of the Tribunal I do not accept that that criticism is justified. In any event, the Court of Appeal’s discussion of the Tribunal’s judgment did not indicate the degree of confidence in the Tribunal’s analysis of the EU instruments that would have justified DCMS relying on those parts of the ruling that were not expressly overturned by the higher Court.

After the Floe litigation had ended, OFCOM revisited the question of the continued application of the Commercial Use Restriction. This appears to have been prompted in part at least by the threat of infraction proceedings by the European Commission which I discuss later. In December 2009 a draft consultation document was prepared by OFCOM. This stated:

“3.37

In response to our 2005 consultations we received confidential responses from Government departments raising significant concerns about the impact of gateway use on public safety and security.

3.38

We take the views of the security services very seriously indeed. Information about phone use is, in our view, of significant and often crucial value in protecting consumers and citizens against terrorist and other criminal activity. We are interested to understand in response to the consultation whether these public safety and security concerns remain. We therefore consider that it is an appropriate time to review the case for regulation of gateways.”

Having set out the options for deregulating the use of GSM Gateways to varying degrees, OFCOM concluded under the heading ‘Ofcom’s preferred approach’ as follow:

“4.259

We believe that the options set out above are finely balanced. We continue to believe that gateways are associated with potentially significant detriments, but we understand that the strength of a key objection, the security concerns, may have diminished over time and the persistence in the medium to long term of other objections were gateways to be deregulated may not be such on their own as to justify maintaining the regulatory burden. We expect that the UK MNOs are likely to be able to deal with the gateway use issues in absence of licensing restrictions in the same way that they do today without Ofcom intervention and in the same way that their counterparts do in certain other Member States.”

OFCOM’s preferred approach was therefore to exempt all gateway uses.

This consultation was, however, never issued. At a meeting between the Home Office and BIS on 18 January 2010 the Home Office officials made clear that their concerns had not diminished and that, indeed, in some respects their concerns have increased. They wanted the current regime to remain in place and reiterated that self-use gateways did not raise the same concerns for them as commercial gateways. Furthermore, they indicated subsequently that they were concerned that the consultation process itself would generate discussion about very sensitive matters and that that, of itself, was undesirable. In a letter to OFCOM on 18 March 2010, BIS noted that OFCOM had indicated that they would not wish to consult on changes to the licensing regime if there were no prospect of the Government implementing those changes. BIS stated that that was the case and, ultimately, OFCOM did not issue the consultation document.

The conclusions I draw from this history is that the balancing of the interests of the GGOs against public security concerns was given prolonged and detailed consideration by those in Government with responsibility for deciding whether to continue the Commercial Use Restriction. There certainly has not been “a complete failure to take account of the specific situation of a defined economic group” to quote Kay LJ in Negassi. The GGOs were able to make submissions to the DCMS officials both in writing and at meetings to put across their arguments as forcefully as they wished. There is nothing to suggest that the Government approached the matter with a closed mind. On the contrary, it is clear that OFCOM and BIS approached the matter on the basis that the default position should be to lift the Commercial Use Restriction unless they could be convinced that it was needed. In the end, they were convinced by the public security arguments put forward by the Home Office. If OFCOM and BIS were in error to the extent that I have found, then that was an excusable error and not an egregious error for the purposes of the Francovich test.

I have considered whether the Home Office was at fault in failing to make the distinction between COMUGs and COSUGs that I have held ought to have been made. I do not consider that it was. The GGOs arguing forcefully for liberalisation did not, so far as I am aware, propose this compromise. Their businesses depended in large part on selling COMUG services and they were intent on getting the Commercial Use Restriction lifted in its entirety. The failure to draw that distinction is not a serious and manifest disregard of the UK’s obligations under the Directive.

It is true that the submissions of officials were not cast in terms of the Government’s obligations under the relevant EU or domestic legislation, but I do not see how the substance of the discussions would have differed if it had been expressly considered within that framework. There is nothing in the history of the Government’s analysis of the issue that indicates an intention to restrict the use of GSM Gateways for some purpose other than that publically expressed and consulted upon or to disregard the UK’s legal obligations. The facts that I have set out militate firmly against any finding that there has been a manifest disregard of the UK’s legal obligations.

One of the factors that Kay LJ referred to in Negassi as relevant in the ‘multi-factorial test’ is the position taken by the Community institutions in the matter. As I have stated earlier, this is not a case where the Government has failed to comply with an earlier court ruling establishing an infringement of EU law.

In January 2004, Floe made a complaint to the European Commission alleging that the Commercial Use Restriction was in breach of the RTTE Directive. The matter was effectively stalled pending the resolution of the Floe litigation and was picked up again by the Commission in August 2009. In a letter of 12 August 2009, the Commission wrote to OFCOM asking for information on the regulatory framework for commercial GSM Gateways. There followed exchanges of letters and meetings between UK and Commission officials. I note that the Commission’s focus by this time was on an alleged breach of the Authorisation Directive rather than the RTTE Directive.

There was debate within Government about how to convince the Commission of the validity of the public security concerns without having to disclose very sensitive material. The Commission officials had told the UK officials that it was not sufficient to cover the material orally in meetings – if the UK wanted to rely on security arguments it would ‘need to spell them out’. A letter was sent by UKRep in Brussels to the Commission in July 2010 setting out in more detail the arguments relied on. There appears to have been no response to this until a letter in July 2011 from the Commission to UKRep asking for clarification of the effect of the UK’s transposition of the Better Regulation Directive’s changes to the Authorisation Directive. In that letter, the Commission noted that the new section 8 of the WTA 2006 which was intended to list what the UK interpreted as the content of the ‘general interest objectives’ referred to in the fourth bullet of the Amended Article 5 did not refer to public security. The Commission wanted to know how this squared with the Government’s insistence, in the context of the threatened infraction proceedings, that public security concerns were the basis for imposing the Commercial Use Restriction. Since it did not appear that the UK was transposing the potential public security grounds referred to in Amended Article 5 into domestic law by the new section 8 WTA 2006, the Commission asked to be pointed to where this would be covered in the domestic provisions.

In reply, in a letter dated 10 August 2011 the UK said that the public security issues in relation to the exercise by OFCOM of its functions are dealt with by the exercise of the power in section 5 of the CA 2003 (see the discussion of section 5 at paragraphs 81 onwards above). The Commission, at that stage, closed its investigation into the complaint and infraction proceedings were not pursued. In accordance with its practice, the Commission did not give reasons for its decision to close the investigation.

The Claimants rely on the Commission’s action as showing that the UK only narrowly escaped infraction proceedings. However, I do not accept that the Claimants can rely on the fact that the Commission threatened but ultimately withdrew infraction proceedings as showing that any resulting infraction was in manifest disregard of the UK’s legal obligations. On the contrary, the fact that the Commission looked into the matter carefully and decided not to proceed, points - at the least - to the difficulty of the issues raised and hence militates against a finding that there has been ‘manifest disregard’ in this case.

Ms Carss-Frisk was strongly critical of the letter that the Government wrote in response to the Commission’s query as to the whereabouts of the transposition of the public security grounds from the Amended Article 5 of the Authorisation Directive, given its absence from the new section 8 WTA 2006. Simon Towler, a senior official at DCMS, was cross-examined on this point. The allegation put to him was that, by referring to section 5 of the CA 2003 without also referring to the fact that no direction had actually been made by the Secretary of State under that section in respect of GSM Gateways, the letter of 10 August 2011 was misleading. I do not accept that criticism of the letter. As I have held (see paragraph 89 above), there was no need for a direction to be made, since the 2003 Exemption Regulations were carried forward by the transitional provisions of the CA 2003. There was nothing misleading about the letter of 10 August 2011.

My conclusion is therefore that the breach I have found by DCMS did not amount to a manifest and grave disregard of its obligations under the Authorisation Directive.

Given that I have held that the second criterion is not satisfied, there is no need strictly to consider the third. However, much, if not most of the evidence presented at trial was directed at the issues of causation and quantum. Since the matter may go further and there were facts in dispute between the parties, I will set out my findings on these issues.

The third criterion for Member State liability is that there is a direct causal link between the breach of EU law and the loss sustained by the individuals. The Claimants say that the question of causation is straight-forward here. There is no doubt that the five Claimants who were GGOs were operating businesses using gateways and that these businesses were severely restricted by the decision in July 2003 not to lift the Commercial Use Restriction. They must have lost some profit and that is enough to complete the Francovich cause of action. So far as Easyair is concerned, it lost major customers for its wholesale business selling SIM cards to GGOs. All other questions as to what would have happened in the counterfactual world if gateways had been legal are relevant to the issues of quantum and not to causation.

DCMS say that the causation issue is more complicated than that. They point to various matters which they say break the chain of causation. The first is the fact that the Claimants did not apply for a licence to use GSM Gateways. The second is the fact that for their own commercial reasons, the MNOs would not have supplied the GGOs with the SIM cards they needed to conduct their business even if GSM Gateway business had been entirely lawful.

The Claimants characterise the Commercial Use Restriction as a prohibition on the use of GSM Gateways. DCMS say that the legal effect of the 2003 Exemption Regulations was merely to require the GGOs to apply for an individual licence if they wanted to operate GSM Gateways. Evidence on this point was given by Mr Regan of the Home Office on behalf of DCMS. Mr Regan was not called for cross-examination by the Claimants. He said that if an application for a licence had been made:

“… it is my expectation that the Home Office would have been or would be consulted as to the form and content of any licence conditions which might be imposed. The Home Office would consult the law enforcement agencies… to ensure that their concerns were adequately addressed.

The Home Office would be concerned to ensure that if and to the extent that licences were granted for the operation of GSM gateways, such licences were subject to conditions requiring operators to ensure that public security concerns were satisfactorily addressed”.

Thus the Claimants, DCMS say, are the authors of their own misfortune in failing to apply for licences.

I reject this argument as being entirely inconsistent with the facts. Throughout the period under consideration, I do not see that there was any prospect of a GGO obtaining a licence for a commercial GSM Gateway. When the application of the 2003 Exemption Regulations to GSM Gateways was first debated within Government, the stance taken by OFCOM was that it was impossible to grant an individual licence to a GGO because the spectrum had been exclusively licensed to the MNOs for their commercial use. This seems to have been the view held until about mid 2004. OFCOM’s view was that the only way that GGOs could lawfully operate a commercial GSM Gateway would be pursuant to a sub-licence granted to them by an MNO. During the period that this view held sway in OFCOM, there was no prospect of a successful application for an individual licence to operate a commercial GSM Gateway.

In the course of the Floe litigation, OFCOM’s stance on this score changed. OFCOM argued before the Competition Appeal Tribunal that the MNOs’ spectrum licences did not cover the use of commercial GSM Gateways so that they could not sub-licence the spectrum to anyone else for this use. That view was ultimately upheld by the Court of Appeal in Floe Telecom v OFCOM [2009] EWCA Civ 47. The Court of Appeal noted that “[t]here was clear evidence before the tribunal that if Floe (or any other company) had applied for an individual licence to provide COMUG services the application would have been refused”. Having considered the evidence placed before me, including various statements in the draft and published consultation documents, I agree with the Claimants that there was no realistic possibility of GGOs being granted licences to deploy GSM Gateways commercially at any time relevant to this dispute.

I note that Mr Regan stops short of suggesting that the Home Office’s response, if consulted on a particular application, would have been markedly different from the stance that they took when they were consulted about the continued application of the Commercial Use Restriction in general. Their response was to oppose any liberalisation of the current regime.

Further, I accept the Claimants’ submission that the case law shows that a person who is adversely affected by a breach of EU law is not required to challenge the illegality of the measure in proceedings before the national court before bringing a claim for Francovich damages. Cases C-397/98 & 410/98 Metallgesellschaft v HMRC [2001] ECR I-1760 concerned the obligation imposed on companies resident in the UK to pay advance corporation tax in respect of dividends paid to their parent companies. The Court considered whether the claim could be refuted on the grounds that the Claimants had not challenged the obligation to pay the tax in advance through the domestic courts. The UK Government argued that there was a general principle common to the legal systems of the Member States according to which an injured party must mitigate its loss when seeking to recover damages. The Court confirmed that actions are subject to national rules of procedure which may require claimants to act with reasonable diligence in order to avoid loss or damage or to limit its extent. However, on the facts it was clear that any application by the claimants in that case for relief from the obligation to make payments would have been refused because that relief was not available to them under the national law which the Court found to be an infringement of EU law. The Court said:

“105.

It therefore appears that, in the cases in the main proceedings, the United Kingdom Government is blaming the plaintiffs for lack of diligence and for not availing themselves earlier of legal remedies other than those which they took to challenge the compatibility with Community law of the national provisions denying a tax advantage to subsidiaries of non-resident parent companies. It is thus criticising the plaintiffs for complying with national legislation and for paying ACT without applying for the group income election regime or using the available legal remedies to challenge the refusal with which the tax authorities would inevitably have met their application.

106.

The exercise of rights conferred on private persons by directly applicable provisions of Community law would, however, be rendered impossible or excessively difficult if their claims for restitution or compensation based on Community law were rejected or reduced solely because the persons concerned had not applied for a tax advantage which national law denied them, with a view to challenging the refusal of the tax authorities by means of the legal remedies provided for that purpose, invoking the primacy and direct effect of Community law.”

The same principle applies here. OFCOM’s position throughout the period was that commercial GSM Gateways caused insuperable problems. The GGOs did not want to have to apply for an individual licence, they wanted a general authorisation to be granted. The Claimants were not obliged to go through the process of applying for a licence, being refused and seeking a judicial review of that refusal before they could claim damages for the loss that the Commercial Use Restriction caused them.

DCMS’ second argument on the issue of causation is that the evidence shows that the MNOs were intent, for their own commercial reasons, on stamping out the use on their networks of SIM cards in GSM Gateways. There is no reason to suppose that this intent would have changed merely because the Commercial Use Restriction was revoked. They say that unless the Claimants can establish that they would have been able to force the MNOs to supply them, by relying on competition law rules, the Claimants have failed to establish that the Government’s hypothesised breach of EU law has caused them any loss.

At this point I am considering only the question whether this submission by DCMS goes to the question of causation – and therefore the fulfilment of the third Francovich criterion - or is relevant only to quantum once the three Francovich criteria have been satisfied.

DCMS’ arguments proceed on the basis that this case should be treated as a loss of a chance case. They say that the Claimants’ alleged loss of profit was dependent on them being able to acquire enough SIM cards from the MNOs to operate a commercial GSM Gateways business. Thus the existence of any loss depends how third parties to the proceedings, the MNOs, would have behaved, rather than just on the conduct of the Claimants and DCMS. They referred to the well-known case of Allied Maples Group Ltd v Simmons and Simmons [1995] 1 WLR 1602 which establishes that where loss would only arise from the defendant’s wrongdoing on the supposition that a third party acted in a certain way, the court’s task is to decide whether the claimant has proved that there was a real and substantial chance that the third party would have acted in that way. If there is no such real and substantial chance, then the claimant has not shown that the defendant’s conduct has caused him any loss. If the claimant can show that there was a real and substantial chance, then he is entitled to the profit lost, discounted to reflect how likely the court considers it was that the third party would have so acted.

The distinction between the loss of a chance situations covered by Allied Maples and an ordinary loss of profit case was considered by the Court of Appeal in Vasiliou v Hajigeorgiou [2010] EWCA Civ 1475. There the claimant asserted that his restaurant in rented premises would have traded profitably over the relevant period if the defendant landlord had not been in breach of the covenant to repair. The Court urged caution in identifying the contingency which is said to represent the lost chance. The loss of chance doctrine is primarily directed to issues of causation and needs to be distinguished from the evaluation of factors which go only to quantum. The Court said:

“Where the quantification of loss depends upon an assessment of events which did not happen the judge is left to assess the chances of the alternative scenario he is presented with. This has nothing to do with the loss of a chance as such. It is simply the judge making a realistic and reasoned assessment of a variety of circumstances in order to determine what the level of loss has been”

The Court held that once the judge had reached a view about the prospects of success of the restaurant, there was no reason to apply a discount to that loss to reflect the fact that the restaurant might have been less profitable.

In my judgment exactly the same principle applies here. Every loss of profit claim depends on the conduct of third parties; on customers wanting to buy the claimant’s products and suppliers being willing to supply the claimant with the raw materials he needs to make those products. Those factors affect the likelihood of success of the claimant’s business but do not turn every loss of profit claim into a claim for the loss of a chance. There is no doubt that there was some truncation of the Claimants’ business by reason of the imposition of the Commercial Use Restriction. The third criterion of Francovich is therefore established. Questions about the likely conduct of the MNOs is a matter I will need to consider in relation to quantum.

My conclusions on whether the Francovich criteria are satisfied in this case can be summarised as follows:

Much of the evidence presented at the trial of this action in fact related to the issues about quantum. The quantum of damages sought by the Claimants together amounted to about £415 million. The issues raised about quantum can be divided into two sets. The first set concerns the construction of the counterfactual: how would the market for the commercial supply of GSM Gateway services have developed if their use had not been limited by the Commercial Use Restriction? The second set of issues concerns the individual characteristics of each Claimant’s business: how would this particular Claimant have fared in the counterfactual world and how much damage has it therefore suffered. The two sets of issues are, of course, linked, since the features of the counterfactual world will affect how much money the Claimants might have made.

Each side at trial relied on the expert evidence of a forensic accountant on the quantification of loss. Each Claimant put forward one or more witnesses to describe their business and the effect on that business of the Commercial Use Restriction. All of them were cross-examined on behalf of DCMS. Although some of these witnesses were criticised by DCMS in their closing submissions, I found them all generally straightforward and helpful. In fact the number of purely factual disputes that I have to resolve in this judgment is surprisingly small.

DCMS also relied on the evidence of two witnesses from the MNOs, both of whom were cross-examined by the Claimants. The first was Lawrence Wardle who is Head of Regulatory at Telefónica UK Limited (which trades as O2) and who was responsible for developing the company’s approach to GSM Gateways. The second was Ian Greenstreet who was the Product Manager of Everything Everywhere from 2010 until 2013 and who had previously been a Product Manager at Vodafone between 1998 and 2004 and Head of Voice Services at Orange between 2005 and 2010. Mr Greenstreet also gave evidence about, amongst other matters, whether Vodafone or Orange would have supplied MCT services to GGOs had the Commercial Use Restriction been lifted.

The Claimants provided a report prepared by Mr Geoffrey Senogles who is a chartered accountant with the economics consultancy firm Charles River Associates. He has worked in the field of forensic accounting for many years. In his report describes how he arrived at the sum of £413,856,409 as the total sum of damages claimed as follows:

To work out the loss of profit suffered by each Claimant, Mr Senogles took the following steps.

Mr Senogles is frank about the difficulties of the task facing him. As regards the first of these parameters – how quickly the business would have expanded - Mr Senogles relied on information provided to him by each of the Claimants. He notes (at paragraph 4.8.3 of his Report) the problem in this case of relying on contemporaneous forecasts as a guide for constructing the counterfactual. He says that the data available for the period of actual trading by the Claimants are both limited and potentially not representative of “normal” trading behaviour. The period of trading as a gateway operator was in most cases less than one year making it hard to assess the trend of actual data. He also notes the impediments created by the fact that some of the Claimants are no longer operating and that the period of loss extends back over a decade in some cases. This inevitably causes problems as regard the availability of supporting data.

Mr Senogles makes clear that he has not attempted to model a wider ‘but for’ scenario in which he would try to predict and then simulate how the telecoms industry would have developed if the use of GSM Gateways had not been restricted. Any attempt at such a model would have to consider the possible effects, for example, of the MNOs and BT setting their own prices differently; or BT and/or the MNOs deciding to enter the GSM Gateway market themselves or changing the structure of tariff pricing altogether.

As to this last point, the tariffs that the MNOs would have charged the GSM Gateway operators, Mr Senogles notes that tariffs have changed considerably over the period of loss in response, for example, to the increased popularity of text messaging and the increased use of smart phones. When modelling losses for each claimant, he has chosen the one tariff which is the one most used when a Claimant was in fact in business and assumed that it was the tariff that the Claimant would have used for the entire period of loss. The exception to that approach was where he was instructed specifically that a Claimant would have moved from one tariff to another as it became available. In that case, he has assumed that all the Claimants would have moved to that tariff.

There was a great deal of material accompanying Mr Senogles’ report in the form of Annexes (which contained his own calculations) and Exhibits (which contained raw material from the Claimants). He described in his oral evidence the iterative process that he had gone through with each of the Claimants when preparing his report.

One oddity of his Report was that he says, in respect of each of the Claimants:

“4.7.1

My team and I have been provided with electronic and hard copy data and we have met with, and interviewed, witnesses of fact and other representatives of the claimant companies.

4.7.2

I am instructed that the witnesses of fact will each provide a statement to deal primarily with the data and assumptions that they have provided to me, and on which I have relied in calculating losses.”

However, as Mr Moser pointed out to him, and to each of the witnesses of fact, in cross-examination, the statements of the Claimants’ witnesses simply said that they had read his report and agreed with the conclusions reached. It was not clear, therefore, where, if anywhere, in the many bundles of documents in court there was evidence which supported the assumptions that underlay his report about the rate at which each Claimant’s business would have grown, their likely costs and so forth. I accept that Mr Senogles prepared his report carefully and conscientiously and that the facts that underpin it are taken from material given or told to him by the Claimants’ personnel whom he and his team interviewed. However, it does not appear that the Claimants’ assumptions passed on to Mr Senogles took into account the various factors I discuss below when estimating how their businesses would have developed in the counterfactual world.

DCMS relied on the evidence of Michael Taub. Mr Taub is a chartered accountant and a partner in the firm RSM Tenon. Mr Taub’s report can be summed up in a sentence: it is all just too difficult. He spells out the many elements of uncertainty over how this market would have developed if the Commercial Use Restriction had not been in place. He concludes that one cannot really make any estimate of the loss suffered. DCMS therefore rely on passage in McGregor on Damages to the effect that where it is impossible to quantify loss, a nominal sum should be awarded in damages.

I reject this approach. It is very easy for a defendant to criticise a counterfactual presented by the claimant and much more difficult to present a convincing counterfactual to the Court. The exercise is necessarily uncertain but that is not a reason for a judge to avoid the task. I reject the suggestion that the exercise that Mr Senogles carried out was not worthwhile. He set out clearly what he had done and also what he had not attempted to do. Where I accept the Claimants have erred is by simply adopting the values Mr Senogles arrived at as being the quantum of loss they had suffered and now claim as damages. They should either have ensured that the assumptions on which Mr Senogles report was based did take these points into account or they should have discounted the figures he arrived at to take account of the uncertainties. For example, in relation to a particular element of uncertainty which Mr Senogles accepts he had not allowed for, the Claimants could have acknowledged this and proposed a reduction of, say, 10 per cent to reflect that. That would have been a more realistic approach. DCMS could then have argued that the correct deduction was 30 or 50 per cent. The court would have been in a better position to evaluate the parties’ respective arguments as to the size of the deduction and make a decision. As it is, I am faced with two extreme positions – the Claimants not proposing any deduction despite the acknowledged uncertainties and DCMS inviting the Court to award nothing on the basis that the uncertainties prevent any proper assessment.

Since the issue of quantum does not arise on my findings, I will not consider in detail the proper counterfactual for the assessment of the Claimants’ loss. I offer only the following comments on some of the elements of uncertainty that Mr Taub highlights in his report, in so far as they raise issues of fact. I have not considered the issues that were raised about each individual Claimant’s loss computation.

The history of the relationship between the GGOs and the MNOs as regards the supply of SIM cards is broadly that the MNOs were initially prepared – even keen – to provide SIM cards for use in GSM Gateways but that in early 2003 this changed. From that time onwards the MNOs took steps so far as possible to prevent SIM cards being used and shut off services to cards where the pattern of call traffic from the card indicated that it was being used in a gateway. The Claimants’ witnesses are clear that the MNOs’ motivation in obstructing the use of GSM Gateways was the realisation that the gateways could seriously erode the profits to be earned from MCT charges for F2M calls.

In cross-examination, DCMS put to the Claimants’ witnesses the allegation that the reason the MNOs had initially been prepared to provide SIM cards was that the Claimants had deliberately misled the MNOs into thinking that the SIM cards would be used in COSUGs and small scale gateway devices (incorporating only a couple of SIMs) whereas in fact they intended to use them in COMUGs and large scale devices (incorporating several dozen channels). There was also some criticism of the witnesses in failing to make clear in their evidence whether they bought their SIM cards directly from the MNOs or from third party wholesalers and the extent to which they had tried to circumvent, by fair means or foul, the MNOs’ attempts to stop their use of the SIM cards.

I find that the picture that emerges is a mixed one. I find that some of the MNOs did display an inconsistent approach towards the desirability of selling SIM cards to GGOs. Some MNO sales staff worked on terms which rewarded them with bonuses linked to the number of SIM cards sold and/or on the basis of the intensity of the use of the cards they sold. For these employees, selling dozens of SIM cards to GGOs who would use them intensively was very attractive. It was only when a different department within the company realised that the volume and intensity of use of these cards actually worked against the wider commercial interests of the company by replacing lucrative F2M minutes with less lucrative M2M minutes, that the company sought to rein back its sales teams and prevent them from supplying these cards to GGOs.

I accept also that the MNOs were experiencing congestion problems caused by the use of the SIM cards in COMUGs and that they were concerned that this was causing disruption to the service enjoyed by their ‘normal’ mobile phone subscribers. This was another reason why they wanted to stop SIM cards being sold to GGOs. I also find that some of the GGOs did not make clear to the MNOs the use that they intended to make of the cards and that they took steps to acquire cards from third party suppliers and keep using them after they knew that the MNOs did not want the GSM Gateways operating on their networks.

Following the Secretary of State’s decision in July 2003, the MNOs had a cast iron reason for halting the supply of services to SIM cards incorporated in commercial GSM Gateways. That is what the Court of Appeal found at the end of the Floe litigation. The question is: if at that stage the Commercial Use Restriction had been lifted rather than confirmed, how would the MNOs have responded?

The Claimants argue that the MNOs would have expanded their capacity to accommodate the use of SIM cards in commercial GSM Gateways. They point to what happened with smart phone use, as described in Professor Webb’s report (see paragraph 133, above). There, initial problems with congestion caused by the sudden increase in spectrum use by the new devices were resolved by the MNOs investing in upgraded their facilities. I do not see, however, why the MNOs would have chosen to do so. The difference is that the use of smart phones offered a new and valuable revenue stream for the MNOs and justified the investment they needed to make to improve their network to cope with the demand. The use of GSM Gateways is inimical to the MNOs interests since it simply transfers value from the MNOs’ business to the GGOs’. I do not see how circumstances could have arisen whereby it was in the interests of MNOs to facilitate the expansion of the GSM Gateways.

The Claimants put forward two reasons why the MNOs might have be persuaded to make this investment. First, there was undoubtedly a market demand for GSM Gateways to reduce the cost of F2M calls. The MNOs, the Claimants argue, would have suffered reputational damage if they were seen to take steps to stifle the GGOs’ ability to meet this demand. There is certainly some evidence of adverse media comment criticising the MNOs when they started to cut off supply of services to GGOs in early 2003, prior to the July 2003 Decision (see paragraph 191, above). There is, however, no evidence to support the suggestion that the MNOs would have been under such commercial pressure from the market that they would have agreed to supply SIM cards in the numbers demanded by the GGOs.

There is some evidence as to how the MNOs would have responded if the use of GSM Gateways had been liberalised because there was a period when it appeared that the correct legal position was that the MNOs were able to sub-licence the right to use spectrum to the GGOs under the auspices of their own individual spectrum authorisation. That was the finding of the Competition Appeal Tribunal in the Floe litigation which pertained between the date of that ruling and the date that it was overturned by the Court of Appeal. It does not appear that the MNOs felt obliged by market pressures to grant such licences to the GGOs.

I do not accept therefore that reputational pressures would have caused the MNOs to invest in expanding their networks to accommodate wide scale use of GSM Gateways on their networks. I find rather that they would have relied on the congestion problems arising from the use of GSM Gateways in certain cells to continue to restrict supply of SIM cards to GGOs.

The second reason why the Claimants say that the MNOs would have provided unlimited numbers of SIM cards to GGOs if the Commercial Use Restriction had been lifted was that they would have been obliged to do so under the competition rules. In other words, the Claimants say that the outcome of the Floe litigation would have been entirely different if the MNOs had not been able to rely on the illegality of the use of GSM Gateways as an objective justification for their otherwise abusive cessation of supply. OFCOM or the Competition Appeal Tribunal would have held that it was an abuse of the MNOs’ respective dominant positions to refuse to supply SIM cards or to stop supply of services to the SIM cards already installed.

The Claimants and DCMS differed in the approach they invited me to take to the question whether the MNOs would have been required by the competition rules to supply SIM cards to the GGOs. I consider that the correct question is whether the risk of such a claim succeeding would have acted as a constraint on the behaviour of the MNOs when deciding whether to supply SIMs to Gateway operators. The high point of Claimants’ case was advice provided in 2002 to Vodafone by two specialist counsel. The Opinion said that there were at least arguable points on which Vodafone could rely in stopping services but that ‘it might be open to legal challenge on the basis of competition law’. Counsel concluded that :

“While we fully understand the basis on which Vodafone would wish to justify such action, it seems to us that the arguments are finely balanced and we cannot advise with confidence that Vodafone would succeed”

The view expressed by counsel (both of whom have, in the intervening years, become eminent in the field of competition law) was very preliminary and this point was one of a number of issues covered in a lengthy opinion. They quite properly flagged up that there was a vulnerability to a competition challenge that the MNO should be aware of. But it did not purport to be a full advice on the merits of defeating a particular claim. I do not regard this as enough to show that if the use of GSM Gateways had been liberalised, the MNOs would have regarded themselves as bound to keep supplying whatever the commercial downside.

The questions raised by the Opinion were overtaken by events, namely the Floe litigation which proved that there was a complete answer to the allegation of abuse. If that complete answer had not been available then, given the loss of revenue that the MNOs would suffer as a result of unlimited use of GSM Gateways, there would have been a much more detailed examination of their obligations. The MNOs no doubt would accept that they were dominant in the wholesale supply of mobile call termination on their networks. But the supply of the SIM card is not simply a supply of mobile call termination. The SIM card enables call origination (and termination on that and other networks) at the retail level. There is a complex interrelationship between the upstream and downstream markets that would need to be analysed in this case and the arbitrage opportunity for the GGO arises from the anomalous fact that the retail price for on-net minutes in a SIM card is often cheaper than the wholesale price for terminating those minutes. It is not at all straightforward to say that MNOs are dominant in a market that is the relevant market for the supply of SIM cards and SIM card services.

If the GGOs could establish that the MNOs were dominant, they would then need to show that refusal to supply was an abuse. I can see that there would be various interesting legal issues to be resolved on this point including whether this constituted a cessation of supply to an existing customer or a new customer; whether the obligation of a dominant firm extended to maintaining a price differential on which the downstream customers’ business was based; or whether the problems of congestion on the MNOs’ networks amounted to an objective justification for conduct that would otherwise be abusive or conversely whether the MNOs were under some duty to invest to accommodate this extra traffic. Given these difficulties, I do not accept that the risk of proceedings would have exercised a substantial constraint on the decision making of the MNOs if they were being asked to supply SIM cards on demand or at the rate that is posited in Mr Senogles’ report.

Although MCT charges were for much of the period covered by this dispute controlled by price controls imposed by OFCOM under its CA 2003 powers, the controls generally allow the MNOs to flex the prices for different services in the relevant basket of different services provided that the average price does not exceed the price set for the basket in the price control. DCMS say that just as it was open to the MNOs to maintain this differential in the pricing of MCT it was also open to them to abolish it overnight if they so chose. Whether they would have so chosen if GSM Gateways had become as widespread as predicted in Mr Senogles’ report is a question that the parties addressed.

In March 2005, OFCOM posed a series of questions to interested parties about the legal effect of the 2003 Exemption Regulations and the issues raised by GSM Gateways. Question 16 asked what obstacles, if any, there were to prevent the MNOs adjusting their tariffs to remove the arbitrage opportunity on which the GGOs’ business was based and what, if such an adjustment took place, the likely effects would be on consumers. Mr Greenstreet was asked about Vodafone’s response to that question which was that there would be three options; raising on-net call charges, reducing wholesale MCT and rebalancing peak and off peak MCT. All of these were described as impractical and as having detrimental effects on consumers. I accept that to remove the arbitrage opportunity would have involved a considerable upheaval in the pricing policies that the MNOs had all adopted before then. Whether they would have done so, and thereby pulled the rug from underneath the GGOs’ business would have depended on a number of practical and reputational considerations. Most importantly it would depend on whether the benefits of earning the higher F2M MCT rates from all the people who, through ignorance or inertia, do not use a GSM Gateway outweigh the losses that are incurred on calls by people who do. I have no doubt there are complex computer models devised by the MNOs capable of calculating exactly where the tipping point lies for any particular tariff and any particular split between fixed line calls using a gateway and those not. I do not have evidence before me as to where that tipping point lies but I accept that the risk of being on the wrong side of that point would have acted as a constraint on the expansion of the GGOs’ businesses if the Commercial Use Restriction had been lifted. This needs to be reflected in the counterfactual and hence in the assumptions underlying the predictions.

Mr Wardle from O2 and Mr Greenstreet from Vodafone described how interconnection services are supplied by MNOs to other MNOs and to fixed line operators to transfer the huge volume of individual calls from one operator to another. This is not done over the spectrum but through a cable. Thus, when a Vodafone caller rings an O2 phone, the call initially travels over Vodafone’s spectrum to the Vodafone base station. It is added by Vodafone to all the other Vodafone to O2 calls, transmitted down a cable to O2 and then transmitted by O2 over O2’s spectrum to the recipient. This option of transferring a high volume of calls over a cable rather than using spectrum is something that the MNOs can offer to customers who wish to make a large number of F2M calls. The O2 service, called Mobex, offers a customer much cheaper F2M calls because when someone in the office rings a Mobex number the call is transmitted to the mobile phone of the colleague working out of the office using a dedicated fixed line, referred to as a pipe or leased line. Mr Wardle’s evidence was that if O2 discovered that a customer was using a GSM Gateway they would offer them this Mobex service instead. The tariff was attractive to the customer and the service had the advantages for O2 both that it avoided the congestion problems created by GSM Gateways and that it enabled O2 to offer a package which in effect captured all their F2M calls to the O2 network.

Mr Greenstreet’s evidence was also to the effect that where Vodafone identified that a customer wanted to use a self-use GSM Gateway, the company would instead try to sell them a solution which reduce the costs of their F2M calls. For customers with a smaller volume of calls this solution would be a small gateway supplied by Vodafone itself, or, subsequently, use of a different technology called ‘indirect access’. If the customer’s demand was for 20 channels or more, then Vodafone would offer to install a pipe or leased line to transmit calls. It was put to him in cross-examination that if Vodafone was prepared to bow to consumer pressure and install some mechanism which greatly reduced the income generated by the customer’s F2M calls, why would it not have allowed, in the counterfactual world, a COMUG market to grow in the absence of the Commercial Use Restriction. His answer was that a leased line is provided as part of a wider package for the corporate customer tying all of that customer’s telecoms business into Vodafone. In order to get the leased line, the customer uses Vodafone for all their national and international calls, many hundreds of mobile handsets for their work force, data transmission, text messaging etc. Vodafone may have to forego the high F2M MCT revenues on some calls but it more than makes up for this loss by other profitable aspects of the customer’s business.

This evidence accords with commercial sense, namely that if the Commercial Use Restriction had been lifted, the MNOs and fixed line operators would not have stood by and allowed the GGOs to undermine their revenue stream without responding competitively with their own ways of reducing F2M costs to customers whose business might justify the use of a gateway. This is another factor that would have constrained the expansion of the Claimants’ business in the counterfactual world and which should have been taken into account in computing the loss alleged to have been suffered by the Claimants.

If I had found that the Francovich criteria were satisfied in respect of the imposition of the Commercial Use Restrictions for COSUGs (but not COMUGs) I would have handed down judgment to that effect. I would not have been in a position to arrive at a conclusion on quantum. I would have adjourned the matter, having given the indications of my views on quantum that I have set out in the preceding paragraphs. That would have given the Claimants an opportunity to take stock and reconsider the value of their claim.

My conclusions in summary are as follows:

The claims are therefore dismissed.

Finally I would like to thank all counsel and their teams for the care that went into preparing this very substantial and complex piece of litigation for trial. I was greatly assisted by the thoroughness and clarity of their submissions.