PROFESSIONAL CONTRACTORS’ GROUP AND ORS v COMMISSIONERS OF INLAND REVENUE [2001] EWCA Civ 1945

[2001] EWCA Civ 1945

Case No Case No: C/2001/0895

Between

PROFESSIONAL CONTRACTORS’ GROUP AND ORSAppellants

COMMISSIONERS OF INLAND REVENUERespondents

Before

Lord Justice Auld

Lord Justice Robert Walker

Lord Justice Dyson

Mr Gerald Barling QC and Miss Kelyn Bacon (instructed by Bond Pearce for the appellants)

Dr Richard Plender QC and Mr Stephen Morris (instructed by the Solicitor of Inland Revenue for the respondents)

Date 21 December 2001

Lord Justice Robert Walker:

Introduction

[1]The distinction between a contract of service and a contract for services (in other words, the difference between the position of an employee and that of a self-employed contractor) has important implications in several different fields of law. The law student first meets the distinction in studying vicarious liability in tort. This appeal is concerned with the tax implications of the distinction, and in particular with the lawfulness under Community law of legislation which the Government announced in 1999, and which Parliament has since enacted, to counter tax avoidance by individuals by the use of what are loosely called service companies.

[2]Employees are liable to income tax on their earnings under Schedule E, and they and their employers have to pay National Insurance contributions (“NIC”) on the Class 1 (employed) basis. Taxation under Schedule E has several well-known disadvantages as compared with the taxation (under Schedule D Case I or II) of those who carry on a business or profession. These disadvantages include immediate taxation at source under PAYE, and a much more restricted scope for the deduction of expenses. Moreover if an individual employee of a company became the controlling shareholder of a service company which (as an independent contractor) provided his services to the former employer as a client, he could achieve a double advantage. The service company would pay a low rate of corporation tax on its profits as assessed under Schedule D, and the individual could decide how much of the company’s revenue should be distributed either as remuneration or by way of dividend (free of NIC) to himself and other members of his family who might be employed by or shareholders in the service company.

[3]On 9 March 1999, which was Budget Day, the Inland Revenue published (among numerous other press releases) one designated IR 35, which has achieved unusual notoriety. The press release began with what the trial

judge described as unduly colourful language:

“The Chancellor announced today that changes are to be introduced to counter avoidance in the area of personal service provision. This move underlines the Government’s commitment to achieving a tax system under which everyone pays their fair share. There has for some time been general concern about the hiring of individuals through their own service companies so that they can exploit the fiscal advantages offered by a corporate structure. It is possible for someone to leave work as an employee on a Friday, only to return the following Monday to do exactly the same job as an indirectly engaged ‘consultant’ paying substantially reduced tax and national insurance. The Government is going to bring forward legislation to tackle this sort of avoidance. The Inland Revenue will be discussing the practical application of new legislation with interested parties and will work with representative bodies on the production of guidance. The new rules will take effect from April 2000.”

[4]The expression IR 35 has come to be used, and has been used in this litigation, as a shorthand identification for the measures which have since been enacted (although only after substantial modification of the original proposals). They consist of section 60 of, and Schedule 12 to, the Finance Act 2000 (“the 2000 Act”) and (in relation to NIC) sections 75 and 76 of the Welfare Reform and Pensions Act 1999 (“the 1999 Act”) and the Social Security Contributions (Intermediaries) Regulations 2000 (“the Regulations”).

[5]In June 2000 three claimants (the appellants in this court) sought permission to apply for judicial review of the lawfulness of IR 35. They are(1) Professional Contractors’ Group Ltd (“PCG”) a body formed to represent the interests of service companies,

(2) Mr Ruud Van Zundert, a Dutch national resident in the United Kingdom who has since 1997 operated a service company in the information technology (“IT”) field, and

(3) Square Mile Projects Ltd, an English service company. PCG represents about 11,000 members and was originally formed for the specific purpose of opposing IR 35, although it now has wider purposes. Its members are predominantly in what has been called the ‘knowledge-based’ sector, an imprecise but useful expression which covers IT, specialised engineering skills, telecommunications and management and business consultancy.

[6]On 10 October 2000 Gibbs J gave the applicants permission to apply for judicial review. The principal relief sought by the application as reamended was a declaration that the IR 35 legislation is “ … incompatible with European Community law as being:(a) an unnotified State aid contrary to Articles 87 and 88 EC in respect of the following areas of business activity: (i) Information Technology (ii) Engineering (including oil and gas) (iii) Telecommunications (iv) Management and Business Consulting;

(b) an unlawful hindrance to free movement of workers, freedom of establishment and freedom to provide services, contrary to Articles 39, 43 and 49 respectively; and cannot lawfully be applied.”

[7]Both sides put in a volume of written evidence and full and detailed skeleton arguments. After a hearing which extended over four days Burton J on 2 April 2001 dismissed the application with costs and refused permission to appeal. Permission to appeal was granted on paper by Laws LJ on 16 May 2001.

[8]Before looking at the judge’s reasoning and the grounds of appeal I should say more about the genesis, purpose and legal and economic effects of the IR 35 legislation. The judge made eight findings of fact which neither side has squarely challenged, although there has been a good deal of argument about their implications and the legal inferences to be drawn from them. The IR 35 legislation

[9]As a matter of parliamentary procedure the enactment of the IR 35 proposals had to be split between the 1999 Act (which received the Royal Assent on 11 November 1999) and the Regulations made under the 1999 Act, on the one hand, and the 2000 Act, on the other hand. But all the measures had the same objectives. They came into force or operated from the same day, 6 April 2000, and the conditions for their operation were expressed in the same language.

[10]Between the publication of IR 35 on 9 March 1999 and the operative date of 6 April 2000 there was extensive consultation which led to some changes in the proposals, announced by the Paymaster General in a press release issued on 23 September 1999. There were also two Regulatory Impact Assessment exercises, the results of which were published on 21 May and 8 October 1999. The main changes resulting from the consultation process were the abandonment of a new test for distinguishing between employment and self-employment, and the placing of responsibility for compliance with the new system on the intermediary (rather than the client). These changes are reflected in the summary which follows.

[11]The basic conditions for the application of the new regime are set out in section 4A of the Social Security Contributions and Benefits Act 1992 (“the 1992 Act”) as inserted by section 75 of the 1999 Act, in paragraph 6(1) of the Regulations and in paragraph 1(1) of Schedule 12 to the 2000 Act. These are in almost identical terms and it is sufficient to set out the provision in the 2000 Act:

“This Schedule applies where – (a) an individual (“the worker”) personally performs, or is under an obligation personally to perform, services for the purposes of a business carried on by another person (“the client”), (b) the services are provided not under a contract directly between the client and the worker but under arrangements involving a third party (“the intermediary”), and (c) the circumstances are such that, if the services were provided under a contract directly between the client and the worker, the worker would be regarded for income tax purposes as an employee of the client.”

[12]In the provision set out above sub-paragraph (c) is of great importance, and it needs to be stressed because some of the written evidence of Mr David Gareth Williams, the Chairman of PCG, tends to overlook its importance. The legislation does not strike at every self-employed individual who chooses to offer his services through a corporate vehicle. Indeed it does not apply to such an individual at all, unless his self-employed status is near the borderline and so open to question or debate. The whole of the IR 35 regime is restricted to a situation in which the worker, if directly contracted by and to the client “would be regarded for income tax purposes as an employee of the client”. That question has to be determined on the ordinary principles established by case law (see for instance two cases mentioned in the written evidence, Market Investigations v Minister of Social Security

[1969] 2 QB 173

and Hall (Inspector of Taxes) v Lorimer

[1994] 1 WLR 209

[13]The following summary adopts the terminology of Schedule 12, paragraph 1(1) of the 2000 Act in referring to “the worker”, “the client” and “the intermediary”, and it assumes that the intermediary is a trading company established and resident in the United Kingdom. But it is important to note that the intermediary need not be a company. It could be a partnership or even another individual, and the IR 35 regime is concerned primarily with the taxation of the worker, not the intermediary (although it includes provisions to avoid double taxation of the intermediary). The tax avoidance at which it is aimed is avoidance by the individual worker, not by the intermediary.

[14]Where the intermediary is a company the IR 35 regime applies (Schedule 12, paragraphs 2 and 3) only if the worker has a material interest in the company (in broad terms at least a five per cent interest, aggregating the interests of the worker himself and any associates of his) or receives what the judge called a ‘traceable dividend’. The application of the regime is triggered by the worker receiving, or becoming entitled to receive, directly or indirectly, a payment or benefit not chargeable to tax under Schedule E. The consequence of its application is that the worker is treated as receiving from the intermediary a “deemed Schedule E payment”. This is treated as made at the end of the tax year. Its amount (Schedule 12, paragraph 7) is determined by a fairly complicated code but the general effect is to tax the worker under Schedule E on the full amount, less the deductions mentioned below, of all payments and other benefits received by the intermediary during the tax year in respect of the worker’s “relevant engagements” (Schedule 12, paragraph 2(3)). The only permissible deductions are 5 per cent of the gross amount and any actual expenses which would be deductible under the restrictive test applied for Schedule E purposes. There are special provisions for “multiple intermediaries” and for avoidance of double taxation (Schedule 12, paragraphs 13-17).

[15]The Regulations produce the same effects for the purposes of Class 1 NIC. The key provision is in regulation 6, paragraph (1) of which is a close parallel to Schedule 12, paragraph 1(1) to the 2000 Act. Paragraph (3) of regulation 6 provides, “Where these Regulations apply –(a) the worker is treated, for the purposes of Parts I to V of the [1992] Act, and in relation to the amount deriving from relevant payments and relevant benefits that is calculated in accordance with regulation 7 (“the worker’s attributable earnings”), as employed in employed earner’s employment by the intermediary, and

(b) the intermediary, whether or not he fulfils the conditions prescribed under section 1(6)(a) of the [1992] Act for secondary contributors, is treated for those purposes as the secondary contributor in respect of the worker’s attributable earnings, and Parts I to V of that Act have effect accordingly.” The judge’s findings of fact

[16]The aims and likely effects of the IR 35 legislation were the subject of a good deal of written evidence placed before the judge. Mr Williams put in two witness statements, the first very lengthy (it has over 300 paragraphs) and accompanied by numerous exhibits, including two reports made in January 2001 by Frontier Economics. There were several other witness statements on behalf of the claimants including one from Doctor Leslie Willcocks of Templeton College, Oxford, exhibiting a report dated 3 May 2000 which he had prepared. The evidence on behalf of the Inland Revenue consisted largely of a witness statement of Miss Sarah Walker, who had since June 1999 been Assistant Director of Personal Tax at the Inland Revenue with special responsibility for IR 35, and (exhibited to a witness statement of Mr Paul Lanser) a report prepared in February 2001 by the Inland Revenue’s Analysis and Research Department in response to the reports from Frontier Economics.

[17]It is not necessary to summarise all this evidence. Its general effect is best addressed by reference to the judge’s findings of fact. Mr Gerald Barling QC and Miss Kelyn Bacon (for the claimants) have claimed in their skeleton argument that the judge found in their favour on almost all factual issues, and rejected their case on what they described as narrow points of law. Counsel for the Inland Revenue, Dr Richard Plender QC and Mr Stephen Morris, have put the matter rather differently. They have pointed out that the judge was from the outset sceptical as to whether there were major differences between the parties on factual issues (as opposed to differences of approach and mindset). In the course of argument the judge put forward eight provisional conclusions about which he said this:

“ … I put them to both Counsel, and they were amended to some extent in the course of argument. Neither side of course agreed all of them, and they are my conclusions: however equally neither side was able with any great vigour to contest that they were conclusions open to me upon the evidence.”

The judge then set out his conclusions, with some short comments. I repeat his conclusions verbatim with some reference to his comments and a few further comments of my own.

[18]The first conclusion was that the intention of the IR 35 measures is “ … to eliminate the avoidance of tax and NIC on payments made by clients in respect of services provided by those who are in fact equivalent to employees; and it has that effect on the companies to which it applies.” The judge added some comments about tax evasion, tax avoidance and tax mitigation, implying that the individuals referred to (echoing the original press release) as ‘Friday to Mondays’ might be regarded as tax evaders but that the aim of the proposals was not limited to such blatant cases. I do not think it helpful to explore the obscure boundary between avoidance and mitigation. The judge referred to the Revenue’s suggested figure of £350m a year as to the overall tax loss from the increased use of service companies (estimated by the Revenue as an increase from 30,000 in 1981 to 90,000 in 1999). He said (paragraph 23), “Whatever the precise figures are, the reality is that if the service company is indeed substantially or wholly captured by IR 35, then it will pay more tax. Hence the Revenue’s desire for change and the Claimants’ desire to challenge it.”

[19]That led inevitably to the judge’s second conclusion, that “Many service contractors will be required to pay more monies and earlier to the Inland Revenue under IR 35 than under the previous arrangements.”

[20]The judge rightly emphasised the point, which I have already noted, that the service contractors adversely affected would be those who provided the equivalent of employees’ services (and not the services of self-employed independent contractors). The extra tax paid by the companies would be Schedule E tax and NIC paid on account of the workers who were equivalent to employees. The judge’s third conclusion was that at least two-thirds of service contractors are in the sector referred to in the amended relief (that is, IT and the three other components of the ‘knowledge-based’ sector mentioned in paragraph 6 above). The judge noted that it was not contended that the legislation was targeted against this (or any) particular sector. It has also affected many other workers in fields as diverse as locum doctors, construction workers and pizza deliverymen. It is true that all the examples in official guidance published in February 2000 were in the IT field but that was explained by PCG and others having complained to the Revenue about the absence of case-law as to the boundary between employed and self-employed status in that field.

[21]The fourth conclusion was as follows:

“Instead of certainty as to the impact of tax and NIC, service contractors as a result of IR 35 have uncertainty as to whether IR 35 will or will not apply to a particular engagement.”

The judge commented that this was very much a part of the claimants’ complaint, and that the Revenue could not really deny it. Service companies did until 6 April 2000 shield those who used them from having to face up to the often difficult question of whether they would, on the terms and in the context of a particular engagement, be on the employed or the self-employed side of an elusive dividing-line. The immunity conferred by the service company had now gone, and the service contractor had to decide the question for himself, with such help as the Revenue could provide either in the way of informal advice or (once an engagement had been entered into) a formal ruling in the course of the tax year. Mr Barling made the valid point that a taxpayer’s difficulties may be increased by the question arising at one or even two removes (that is through separate contractual links between client and agency, agency and service company, and service company and individual; the latter difficulty may be self-inflicted but the former is not if the client recruits contract services only through an agency).

[22]The judge’s fifth conclusion was that in respect of engagements or contracts sought, or services to be provided, by service contractors, there is or would be competition with companies which would be unaffected by IR 35. In view of the vast field of economic activity to which IR 35 relates, this conclusion was probably inevitable, despite Miss Walker’s evidence (in relation to the IT sector) that the services offered by large IT companies (such as EDS, Andersen Consulting, Cap Gemini Ernst and Young, Logica and Sema) are fundamentally different in scale and in nature from those likely to be offered by small service contractors. The main report from Frontier Economics indicated that the pattern is quite complex and that some small service companies are in direct competition with the largest companies.

[23]The judge recorded the Revenue’s submission that the relevant comparison, in terms of competition, was not between a large company and a small service contractor but between a genuine employee and a service contractor:

“The Revenue submits that the real competition was between those who were doing the work at the clients’ premises, sometimes in the same team: on the one hand employees, whether of the client or of the service provider, and on the other the service contractor. That is the competition that Dr Plender QC submitted to be the relevant one.”

[24]The sixth conclusion was as follows:

“Companies unaffected by IR 35 would have greater flexibility to arrange their tax affairs, to allocate tax between income tax and corporation tax, to defer tax liabilities, and to pay lesser salaries to those providing the services and higher dividends to shareholders, than service contractors.”

This is an important building-block in the claimants’ case on state aid. The judge said that he was entitled to make this finding on the evidence, but noted three points made by the Revenue. First, the companies alleged to enjoy this greater flexibility would already be deducting income tax and NIC under PAYE in respect of the whole of their employees’ remuneration. Secondly (and following from the first point) IR 35 was restoring a level playing field. Thirdly, service contractors would continue to enjoy the same flexibility so far as they were engaged in genuine ‘independent contractor’ work.

[25]The seventh conclusion was another important building-block in the claimants’ case on Community law, although the judge emphasised that it was a limited finding based on expert evidence which had not been tested by cross-examination:

“Some service contractors may not continue to operate in the United Kingdom as a result of IR 35, and some who have intended to come to the United Kingdom to set up or work as service contractors may not now come to the United Kingdom.”

Finally the judge reached his eighth, ‘very limited’ conclusion, that his fifth, sixth and seventh conclusions might have an effect on trade between Member States. State aid and general measures

[26]As well as issues of state aid and freedom of movement (on which he was not asked to make a reference to the Court of Justice under Article 234 because of the disputed issues of fact), the judge also had to consider arguments based on Article 1 of the First Protocol to the European Convention on Human Rights. He rejected those arguments and they have not been pursued on appeal. I can therefore proceed at once to the issue of state aid, considered at paragraphs 52 to 68 of the judge’s judgment.

[27]Article 87 (formerly 92) begins with paragraph (1) in the following terms:

“Save as otherwise provided in this Treaty, any aid granted by a Member State or through State resources in any form whatsoever which distorts or threatens to distort competition by favouring certain undertakings or the production of certain goods shall, insofar as it affects trade between Member States, be incompatible with the common market.”

Paragraphs (2) and (3) then instance types of aid which are compatible, or may be considered compatible, with the common market. Article 88 (formerly 93) relates to the review of aid by the Commission and the notification of aid proposals to the Commission. Neither side has placed much reliance on these provisions in this court and I need not make any further reference to them.

[28]The judge identified six elements which are characteristic of the type of aid which may contravene Article 87. In view of the judge’s unchallenged findings only the first and third of these elements are directly in issue (and they are, as the judge said, intertwined). But for the sake of completeness I will set out the judge’s list in full:

“(1) An ‘aid’ in the sense of a benefit or advantage which (2) is granted by the state or through state resources, (3) favours certain undertakings over others (the ‘selectivity’ principle) (4) distorts or threatens to distort competition, (5) is capable of affecting trade between Member States and (6) has not been notified to the Commission.”

All these points go to make up what this court (in Queen (Lunn Poly) v Commissioners of Customs & Excise [1999] EuLR 653, 662) called a ‘global question’. It is a question for the national court.

[29]In this case the two contentious elements of aid and selectivity are particularly closely intertwined because at first sight IR 35 is not providing an aid or benefit to anyone. At first sight it is imposing a detriment on service providers in order to prevent tax avoidance and restore fiscal parity. It is very well established that state aid may consist of a tax concession or relief (rather than a direct subsidy): see the early Case 30/59 Steenkolenmijnen v High Authority [1961] ECR 1, 19. But the detriment imposed by IR 35 can constitute aid only if it is to be seen as favouring other competing undertakings, in breach of the selectivity principle.

[30]The principle of selectivity is of fundamental importance in this case, as both sides agree, because it is the primary means of differentiating between objectionable state aid (which favours one or more identifiable undertakings or sectors of the economy) and general measures which do not have that effect. A recent example is the judgment of the Court of First Instance in

Case T-55/99

CETM v Commission (judgment 29 September 2000). In 1994 the Spanish Government had introduced a ‘plan renove industrial’ (“PRI”) under which small and medium-sized enterprises and regional public bodies (but not large commercial enterprises) could obtain loans on favourable terms in order to purchase commercial vehicles, so long as an old vehicle was withdrawn from service on each purchase of a new vehicle.

[31]In rejecting the argument that the PRI should be regarded as a general measure the Court of First Instance said in paragraphs 52-54 (with some references inserted):

“Measures entailing differences in treatment between categories of undertakings or between sectors of activity may be justified by the nature or structure of the system of which they form part (see Case 173/73 Italy v Commission , [1974] ECR 709, paragraph 33 and Case C-75/97 Belgium v Commission [1999] ECR I–3671, paragraphs 33 and 34; see also Case T-67/94 Ladbroke Racing v Commission [1998] ECR II-1, paragraph 76). In the present case, however, the sole circumstance, put forward by the applicant, that the PRI was aimed at modernising the commercial vehicles on the road in Spain in the interest of environmental protection and improving road safety cannot suffice for a finding that the PRI constituted a system or a general measure in itself or formed part of any ‘Spanish system’, which, moreover, the applicant does not even identify. If that argument were followed, it would be sufficient for the public authorities to invoke the legitimacy of the objectives which the adoption of an aid measure sought to attain for that measure to be regarded as a general measure outside the scope of Article 92(1) of the Treaty. That provision does not distinguish between measures of State intervention by reference to their causes or aims but defines them in relation to their effects (Case C-56/93 Belgium v Commission [1996] ECR I-723 , paragraph 79, Case C-241/94 France v Commission , [1996] ECR I-4551 , paragraph 20, and Case C-75/97 Belgium v Commission , [1999] ECR I-3671 , paragraph 25). Furthermore, as the Commission points out in its written submissions, the applicant does not explain how the exception for large undertakings was justified by the nature or structure of the alleged system to which the PRI corresponded or of which it formed part. In any event, the objectives which the applicant claims the Spanish authorities sought to attain by means of the PRI do not justify such an exception, since the age of commercial vehicles used by large enterprises also presents risks in terms of environmental protection and road safety.”

[32]An element of official discretion as to the recipient of special treatment is also inconsistent with the character of a general measure: see

Case C-200/97

Ecotrade v Altiforni e Ferriere di Servola (Court of Justice 1 December 1998), a case on the so-called Prodi law for the extraordinary administration of large insolvent companies. The Court of Justice said in paragraph 40:

“In those circumstances, having regard to the class of undertakings covered by the legislation in issue and the scope of the discretion enjoyed by the minister when authorising, in particular, an insolvent undertaking under special administration to continue trading, that legislation meets the condition that it should relate to a specific undertaking, which is one of the defining features of State aid (see, to that effect, Case C-241/94 France v Commission [1996] ECR I-4551 , paragraphs 23 and 24).”

[33]Another instructive case is

Case C-75/97

Belgium v Commission

[1999] ECR I-3671

. In 1981 the Belgian Government had introduced a social security system (Maribel I) under which lower rates of contributions were paid by manual workers. The Commission did not object to this scheme because it regarded it as ‘general and automatic’. But in 1993 and 1994 the Belgian Government introduced the ‘Maribel bis’ and ‘Maribel ter’ schemes which further reduced contributions for workers in various sectors most exposed to international competition. These new measures had a selective effect in favouring large but identifiable sectors of industry, and so they amounted to state aid. Their social character was not sufficient to exclude them (see paragraphs 24-34 of the judgment).

[34]The judge referred to these and other authorities as examples of what he called ‘positive aid’ and then said that a more difficult question arose in relation to what he called ‘negative aid’. This expression seems to have been coined by the judge. Although it is a vivid expression I respectfully doubt whether it is useful. For the state to confer a benefit on an identifiable group of undertakings is at first sight state aid. For the state to impose a detriment (for instance, a ‘windfall’ tax on privatised utilities) can be state aid only if it can be seen as occasioning a corresponding advantage to identifiable business competitors of those who have to bear the detriment (see Miss Bacon’s article State Aids and General Measures (1997) 17 YEL 269, 318). The expression ‘negative aid’ may be unhelpful by appearing to assume what has to be proved.

[35]The judge then referred to two Commission notices, 0J 1995 C 312/07 (on co-operation between national courts and the Commission on state aid) and 0J 1998 C 384/03 (on the application of the state aid rules to measures relating to direct business taxation). The latter notice contains a section headed ‘Distinction between state aid and general measures’ which merits full quotation:

“13. Tax measures which are open to all economic agents operating within a Member State are in principle general measures. They must be effectively open to all firms on an equal access basis, and they may not de facto be reduced in scope through, for example, the discretionary power of the State to grant them or through other factors that restrict their practical effect. However, this condition does not restrict the power of Member States to decide on the economic policy which they consider most appropriate and, in particular, to spread the tax burden as they see fit across the different factors of production. Provided that they apply without distinction to all firms and to the production of all goods, the following measures do not constitute State aid: tax measures of a purely technical nature (for example, setting the rate of taxation, depreciation rules and rules on loss carry-overs; provisions to prevent double taxation or tax avoidance) measures pursuing general economic policy objectives through a reduction of the tax burden related to certain production costs (research and development (R&D), the environment, training, employment). 14. The fact that some firms or some sectors benefit more than others from some of these tax measures does not necessarily mean that they are caught by the competition rules governing State aid. Thus, measures designed to reduce the taxation of labour for all firms have a relatively greater effect on labour-intensive industries than on capital-intensive industries, without necessarily constituting State aid. Similarly, tax incentives for environmental, R&D or training investment favour only the firms which undertake such investment, but again do not necessarily constitute State aid. 15. In a judgment delivered in 1974 [Case 173/73 Italy v Commission [1974] ECR 709], the Court of Justice held that any measure intended partially or wholly to exempt firms in a particular sector from the charges arising from the normal application of the general system ‘without there being any justification for this exemption on the basis of the nature or general scheme of this system’ constituted State aid. The judgment also states that ‘Article [87] does not distinguish between the measures of State intervention concerned by reference to their causes or aims but defines them in relation to their effects. The judgment also points out that the fact that the measure brings charges in the relevant sector more into line with those of its competitors in other Member States does not alter the fact that it is aid. Such divergences between tax systems, which, as pointed out above, are covered by Articles [94] to [97], cannot be corrected by unilateral measures that target the firms which are most affected by the disparities between tax systems. 16. The main criterion in applying Article [87] (1) to a tax measure is therefore that the measure provides in favour of certain undertakings in the Member State an exception to the application of the tax system. The common system applicable should thus first be determined. It must then be examined whether the exception to the system or differentiations within that system are justified ‘by the nature or general scheme’ of the tax system, that is to say, whether they derive directly from the basic or guiding principles of the tax system in the Member State concerned. If this is not the case, then State aid is involved.”

[36]Mr Barling submitted (although the point was, I think, developed only in the course of his reply) that IR 35 is on its face selective because paragraph 1 of Schedule 12 to the 2000 Act (and the corresponding provisions in the Regulations) do relate to an identifiable sector of the economy, that is small worker-owned companies which provide the personal services of workers to other businesses. A sector may be very broadly defined but still fall within the principle of selectivity: see the Commission’s proposal for appropriate measures in respect of Irish corporation tax OJ 1998C 395/19 (preferential rate of corporation tax for the whole manufacturing sector).

[37]This submission calls for serious consideration but I do not accept it. I think Dr Plender had already answered it in anticipation by referring to the decision of the Commission on the Danish tax relief notified to the Commission in 1999 under Article 88(3), and cleared as not amounting to state aid. The tax relief, intended to attract highly-qualified experts to Denmark, took the form of taxation at a gross, fixed rate on foreign experts employed for between six months and three years in a Danish business or research organisation. The Commission’s decision stated, “The Act itself is not seen as benefiting certain businesses or certain activities, since it applies to experts in all areas and its area of application is not restricted to regions or sectors. The application is assessed on the basis of objectives and non-discriminating criteria without tax authorities being given discriminating powers. The information from the Danish authorities shows that the scheme is used in more than 150 sectors, by small and medium-sized businesses and by large businesses in the private and public sectors. The reason for the concentration in the office equipment, software and fuel sectors is that salaries are especially high in these sectors. In the correspondence from the Commission regarding the control of state aid and the reduction in employer costs, the Commission writes:

“Salaries in the chemical, oil refining, office machinery and IT industries are on average twice as high as salaries in the textile industry and about three times higher than salaries in the footwear and clothing industry and hotel and catering industry. Schemes aimed at the low-paid will therefore have a much greater effect in the latter sectors than in those first mentioned without actually constituting state aid.”

In contrast, schemes aimed at the high-paid will have a considerably greater effect in sectors with high salaries but this still does not constitute state aid.”

[38]Dr Plender cited that decision as showing that a mere propensity for a measure to favour one sector rather than another cannot amount to selectivity. I agree. The lack of uniformity in the practical effects of the Danish measure (as between different sectors) arose not from any selectivity but simply because that is ‘how things are’. The position is the same with IR 35, except that there the claimants face the further difficulty of a measure which has a propensity to disadvantage more workers in some sectors than in others. Lunn Poly and Ferring

[39]The judge rightly devoted some time to considering the decision of this court in Lunn Poly [1999] EuLR 653. It is an important decision which calls for examination in some detail. It was a direct challenge to the validity of domestic primary legislation, that is section 21 of the Finance Act 1997, which amended section 51 of the Finance Act 1994 by introducing new differential rates of insurance premiums tax (IPT). The original uniform rate of 2.5 per cent was replaced by a standard rate of 4 per cent and a higher rate of 17.5 per cent, the higher rate being charged on premiums for travel insurance charged by tour operators or travel agents. The lower rate was payable on travel insurance arranged by anyone else. Lunn Poly (a large travel agent and part of the Thomson Group, which included tour operators) obtained a declaration from the Divisional Court that the new differential rates were unlawful state aid contrary to Article 87 (then Article 92).

[40]The fiscal and economic background to the case was that insurance premiums were exempt from value added tax (under Article 13B(a) of the Sixth Directive, 77/388/EEC) but Member States were free (under Article 33) to impose tax on insurance contracts, so long as the tax was not a turnover tax. The standard rate of value added tax in the United Kingdom at that time was the current rate of 17.5 per cent. The economic background was that there was fierce competition between tour operators, resulting in low margins on package holidays for both operators and agents. However the convenience of being able to arrange and pay for travel insurance at the same time enabled them to obtain a higher margin on related travel insurance. This point was developed in affidavit evidence on behalf of the Commissioners, who attacked it as a form of tax avoidance which justified the new differential rates in order to correct the anomaly.

[41]The Court of Appeal followed the Divisional Court in rejecting this analysis as not being established on the evidence. Lord Woolf said that the approach of the Commissioners’ deponent was ‘logically indefensible’. He continued (at p.664):

“All he has succeeded in demonstrating is that the demand for travel insurance is highly price inelastic. This enables travel agents, in particular, to charge their customers a premium which they should find uncompetitive. They do not do so because they are guided by factors other than price when making their purchasing decision on insurance. Having come to this conclusion, there is no loss of tax which provides an objective justification for the discriminatory rate of tax imposed on tour operators and agents providing insurance. The higher rate contrary to the stand adopted by the commissioners cannot be objectively justified as an anti-tax avoidance measure.”

[42]Once the court had reached that conclusion the differential rates were inevitably a form of state aid, since it was obvious that the imposition of the higher rates of IPT on travel agents and tour operators conferred a corresponding advantage to an identifiable class of undertakings, that is their competitors in the travel insurance market. As Clarke LJ said (at p.668), it would be startling if the differential rates could escape merely because the lower rate (rather than the higher) was designated as the ‘standard’ rate.

[43]However Lord Woolf (with whom Schiemann and Clarke LJJ agreed, although Clarke LJ added further reasons) expressed the clear view that the result would have been different if tax avoidance had been established. After referring to the judgment of the Court of Justice in Case 173/73 Italy v Commission (which had stated that the point of departure must be the competitive position within the common market before the adoption of the relevant measure) Lord Woolf observed (at pp.663-4):

“ ‘The point of departure’ here was the position before the differential tax rates were introduced. In accordance with that decision, it is necessary to focus on the effect the introduction of the differential rate of tax had on the previous position in order to decide whether the change in the rates constituted an aid. In doing this I do not observe from the authorities any suggestion that it is not permissible to look at the reason the member state put forward for imposing the differential rate. To not do so is to approach the issues in a vacuum. It is here that the question of there being an objective justification for the implementation of the measure could be relevant. To take an example relevant to the present case; if a higher rate of tax were imposed to rectify an actual loss of tax due to a tax avoidance scheme initiated by certain members of a group of taxpayers, that would not mean that the remainder of the relevant group of taxpayers were receiving an aid because of the higher discriminatory rate of tax imposed specifically on the tax avoiders. Nor should it make any difference if, instead of the rate of tax being increased for those who are involved in the tax avoidance, it is reduced for those not so involved. In both situations what is being achieved is a level playing field. Where an explanation of this nature is put forward, if the court is satisfied that what had happened is justifiable, the result would be that there would be no discrimination which could constitute an aid.”

[44]In this important passage (read with that set out at paragraph 41 above) Lord Woolf was not saying that the rectification of a fiscal anomaly was a form of state aid, but was objectively justifiable. What he was saying was that because the apparent discrimination was objectively justifiable in order to prevent tax avoidance, it did not selectively favour those who were not tax-avoiders and it was not therefore state aid.

[45]Since the judge’s decision in the present case the effect of differential taxes has been considered by the Court of Justice in Ferring v Agence Centrale des Organismes de Sécurité Sociale (opinion of Advocate General Tizzano 8 May 2001; Court of Justice 22 November 2001). In France the supply of medicinal products is sharply divided between pharmaceutical wholesalers (who are subject to statutory duties as to the maintenance of stocks and the guarantee of deliveries) and retailers (who may obtain direct supplies from manufacturers, and are not subject to the same statutory duties). In 1997 the French social security code was amended to introduce a tax of 2.5 per cent on direct sales by pharmaceutical manufacturers to retail pharmacies. This tax was not imposed on sales by wholesalers and was devoted to funding sickness insurance (paragraph 9 of the judgment), “ … to restore the balance of competition between the various distribution channels for medicines, which had been regarded as distorted by the fact that wholesale distributors are under a duty of public service which is not imposed on pharmaceutical laboratories.”

[46]The Court of Justice, while noting that state aid is a wide concept, observed (paragraph 17) “Nevertheless, the fact that undertakings are treated differently does not automatically imply the existence of an advantage for the purposes of Article [87](1) of the Treaty. There is no such advantage where the difference in treatment is justified by reasons relating to the logic of the system (see, to that effect,

Case C-353/95

P Tiercé Ladbroke v Commission

[1997] ECR I-7007

, and particularly paragraphs 33-35).”

[47]After referring to a case on the ‘polluter pays’ principle the Court observed (paragraphs 27-29), “In like manner, provided that the tax on direct sales imposed on pharmaceutical laboratories corresponds to the additional costs actually incurred by wholesale distributors in discharging their public service obligations, not assessing wholesale distributors to the tax may be regarded as compensation for the services they provide and hence not State aid within the meaning of Article [87] of the Treaty. Moreover, provided there is the necessary equivalence between the exemption and the additional costs incurred, wholesale distributors will not be enjoying any real advantage for the purposes of Article [87] (1) of the Treaty because the only effect of the tax will be to put distributors and laboratories on an equal competitive footing. In this case it is for the national court to decide whether that condition is satisfied. The answer must therefore be that Article [87] of the Treaty is to be interpreted as meaning that, because it is charged only on direct sales of medicines by pharmaceutical laboratories, a measure such as the tax introduced by Article 12 of Law No 97-1164 amounts to State aid to wholesale distributors only to the extent that the advantage in not being assessed to the tax on direct sales of medicines exceeds the additional costs that they bear in discharging the public service obligations imposed on them by national law.” This is a clear recognition by the Court of Justice that a measure which was apparently discriminatory (in that case, as between manufacturers and wholesalers) might on examination prove to be the opposite (that is, to put the two groups of undertakings on an equal competitive footing).

[48]The reference in Ferring (at paragraph 17) to “the logic of the system” echoes similar expressions in C-72/91 and C-73/91 joined cases Sloman Neptun Schiffarts v Seebetriebsrat

[1993] ECR I-887

, paragraph 21 (“consequences … inherent in the system”) and

Case C-353/95

P Tiercé Ladbroke v Commission

[1997] ECR I-7007

, paragraph 35 (“reasons relating to the logic of the totalisator betting system”; this case is an instructive study in the difficulties of fair comparisons in a specialised economic activity which can be organised in different ways). The point is summed up in paragraph 7 of the Commission notice on co-operation already mentioned:

“ …measures which have neither as their object nor as their effect the favouring of certain undertakings or the production of certain goods, or which apply to persons in accordance with objective criteria without regard to the location, sector or undertaking in which the beneficiary may be employed, are not considered to be state aid.”

State aid: conclusions

[49]After considering all these points the judge reached his conclusions on state aid in paragraphs 67 and 68 of his judgment. His conclusions were that IR 35 is a general measure (and not an exception to or a derogation from a general measure). Its aim is to ensure (so far as possible) that all those who supply employee-like services should pay income tax and NIC under the system appropriate to employees, and should not be able to avoid that system by the interposition of an intermediary . The judge thought that the reamendment of the relief claimed (so as to refer to four component parts of the ‘knowledge-based’ sector) only served to emphasise that the measure was not targeted at any particular sector, and did not confer advantages on an identifiable group of competing undertakings.

[50]Despite the very clear and helpful written submissions prepared by Mr Barling and Miss Bacon, and despite Mr Barling’s further oral submissions, I am not persuaded that the judge was wrong in his conclusions or that there was any significant fault in his reasoning. I have already mentioned that I do not find the concept of ‘negative aid’ particularly helpful. Criticism was also directed at the judge’s reliance on a ‘pragmatic’ approach, following Clarke LJ in Lunn Poly (at p.674). I see little force in this. ‘Pragmatic’ is a word which can convey a range of attitudes, from ‘realistic and in touch with the facts’ to ‘unprincipled and driven by expediency’. I do not think either Clarke LJ or the judge was suggesting that the court has any discretion to steer its findings towards what is convenient and Dr Plender disavowed any such suggestion.

[51]The heart of the matter (as Auld LJ pointed out at an early stage in the argument) is what sort of ‘system’ IR 35 is concerned with, and (following from that) what is the relevant comparison for the purposes of “an equal competitive footing” ( Ferring ) (the United Kingdom equivalent, which is becoming rather a well-trodden cliché, being the level playing field). Mr Barling has skilfully and determinedly argued that the relevant system is corporate taxation, and that the relevant comparison is between a service company whose activities are caught by IR 35 and a larger trading company, providing similar services, which is not caught. But for the reasons which I have already explained I cannot accept that approach. Certainly the tax provisions of IR 35 have to address the service company’s liability to corporation tax, in order to avoid double taxation. But the aim of both the tax and the NIC provisions (an aim which they may be expected to achieve) is to ensure that individuals who ought to pay tax and NIC as employees cannot, by the assumption of a corporate structure, reduce and defer the liabilities imposed on employees by the United Kingdom’s system of personal taxation. I would therefore reject this ground of appeal. Freedom of movement: introduction

[52]The free movement of goods, workers, services and capital is of fundamental importance to Community law. Title III of the Treaty deals with free movement of persons, services and capital, and it contains the three articles on which the claimants rely in their other grounds of appeal: Article 39 (formerly 48) which is in Chapter 1 (Workers); Article 43 (formerly 52) which is in Chapter 2 (Right of Establishment); and Article 49 (formerly 59) which is in Chapter 3 (Services).

[53]Article 39 is as follows:

“1. Freedom of movement for workers shall be secured within the Community. 2. Such freedom of movement shall entail the abolition of any discrimination based on nationality between workers of the Member States as regards employment, remuneration and other conditions of work and employment. 3. It shall entail the right, subject to limitations justified on grounds of public policy, public security or public health: (a) to accept offers of employment actually made; (b) to move freely within the territory of Member States for this purpose; (c) to stay in a Member State for the purpose of employment in accordance with the provisions governing the employment of nationals of that State laid down by law, regulation or administrative action; (d) to remain in the territory of a Member State after having been employed in that State, subject to conditions which shall be embodied in implementing regulations to be drawn up by the Commission. 4. The provisions of this Article shall not apply to employment in the public service.”

[54]Article 43 is as follows:

“Within the framework of the provisions set out below, restrictions on the freedom of establishment of nationals of a Member State in the territory of another Member State shall be prohibited. Such prohibition shall also apply to restrictions on the setting-up of agencies, branches or subsidiaries by nationals of any Member State established in the territory of any Member State. Freedom of establishment shall include the right to take up and pursue activities as self-employed persons and to set up and manage undertakings, in particular companies or firms within the meaning of the second paragraph of Article 48, under the conditions laid down for its own nationals by the law of the country where such establishment is effected, subject to the provisions of the Chapter relating to capital.”

[55]Article 49 is as follows:

“Within the framework of the provisions set out below, restrictions on freedom to provide services within the Community shall be prohibited in respect of nationals of Member States who are established in a State of the Community other than that of the person for whom the services are intended. The Council may, acting by a qualified majority on a proposal from the Commission, extend the provisions of the Chapter to nationals of a third country who provide services and who are established within the Community.”

[56]These three articles are mutually exclusive in the sense that a particular factual situation cannot fall to be considered under more than one article. Article 39 deals with those who are seeking employment, Article 43 deals with those who are seeking to do business through a permanent establishment in another member state, and Article 49 deals with those who are seeking to provide services in another member state without being established there. Article 49 can therefore apply in factual situations where the service provider never even sets foot in the other member state (for instance, if the business is television broadcasting). But although the three articles are mutually exclusive it is for the purposes of this case necessary to consider each of them.

[57]Counsel have referred, either at length or in passing, to a large number of authorities on freedom of movement. As to the principles which the authorities establish or illustrate, there has (as might be expected) been a good deal of common ground between counsel. They agreed that in relation to each of the three relevant articles the judge had to ask himself whether there was on the face of it a contravention of the article, and if so whether the contravention could be justified (the burden of demonstrating justification being on the member state).

[58]They agreed that the only possible justification would be, not on any of the ‘statutory’ grounds expressly mentioned in the Treaty (see Articles 39(3), 46 and 55) but on the Cassis de Dijon principle as applied to freedom of movement under Title III (see Case 120/78 Rewe Zentrale v Bundesmonopolverwaltung [1979] ECR 649 ). That very important case was decided on Article 28 (then 30) and in that context it was held that an apparent infringement could be justified if the measures in question were (paragraph 8 ) “ … necessary in order to satisfy mandatory requirements relating in particular to the effectiveness of fiscal supervision, the protection of public health, the fairness of commercial transactions and the defence of the consumer.” In Cassis de Dijon the interests of the consumer justified a national rule as to the minimum alcohol content of liqueurs. Justification includes (but is not limited to) the requirement of proportionality.

[59]Counsel were also largely in agreement that each of the three articles could be contravened in three different ways. The first and most obvious contravention is by measures which are discriminatory. Dr Plender gave as examples of discriminatory contraventions (under Article 39)

Case C-204/90

Bachmann v Belgium

[1992] ECR I-249

; (under Article 43)

Case C-264/96

ICI v Colmer

[1998] ECR I-4695

and (under Article 49) Case C 55/98 Skatteministeriat v Bent Vestergaard

[1999] ECR I-7641

[60]Bachmann was concerned with a provision of Belgian tax law which allowed pension and life assurance contributions to be deducted only if paid in Belgium to a Belgian undertaking or the Belgian establishment of a foreign undertaking. This provision applied regardless of the taxpayer’s nationality, but it amounted to indirect discrimination because it was likely to operate (paragraph 9) “to the particular detriment of those workers who are, as a general rule, nationals of other member states”. ICI v Colmer (a reference to the Court of Justice from the House of Lords) was concerned with consortium relief for the purposes of United Kingdom corporation tax. The provision contravened Article 43 (then 52) because it denied relief unless the business of the holding company owned by the consortium consisted wholly or mainly in holding shares of companies resident in the United Kingdom. This was a deterrent to the exercise of the right of establishment in other member states. Vestergaard was concerned with an administrative practice of Danish tax law under which attendance at professional conferences or courses held in other countries (generally at tourist resorts) was disallowed unless the location could be justified on professional grounds.

[61]The second type of contravention was described by the judge as dislocation. It is really a species of discrimination (indeed the judge classified Bachmann as a case of dislocation). Dr Plender’s examples of dislocation were (under Article 39) Case C 10/90 Masgio v Bundesknappschaft

[1991] ECR I-1119

, (under Article 43)

Case C-250/95

Futura Participations v Administration

[1997] ECR I-2471

; and (under Article 49)

Case C-6/98

ARD v PRO Sieben Media

[1999] ECR I-7599

[62]The facts of these cases illustrate the notion of dislocation (which can perhaps be described as a discriminatory effect, often unintended, arising from the interaction of different systems in different member states). Mrs Masgio’s husband had worked as a migrant worker in mines in Belgium and Germany and had contracted silicosis, for which he received a disability pension from a Belgian institution. His Belgian old-age pension was reduced as a result, but his pension under the German miners’ scheme was reduced to take account of the gross amount of his Belgian old-age pension. Futura was concerned with the taxation under Luxembourg law of the permanent establishment in Luxembourg of a French company based in France. The Luxembourg tax authorities would not allow trading losses to be carried forward unless they were shown (by accounts maintained in Luxembourg) to be linked to economic activity in Luxembourg. The Court of Justice held expressly (paragraph 22) that this did not entail any discrimination, (overt or covert). Nevertheless the requirement for accounts to be kept in Luxembourg was held to contravene Article 43 (then 52). PRO Seiben was concerned with the interpretation of the ‘Television without frontiers’ Directive, which contained a provision regulating the permitted number of interruptions of feature films by advertisements. There were two possible interpretations, referred to as the gross principle and the net principle, and the gross principle was held to be correct. That led to the further question whether national rules could lawfully impose the stricter, net principle. It was held that this would be contrary to Article 49 (then 59).

[63]The first important difference between counsel arose on the third type of contravention. This type involves neither discrimination nor dislocation and for that reason the judge referred to it as ‘neutral’. The judge seems to have thought that it was relevant only to Article 49, but Dr Plender has not sought to defend that view in this court. He has put forward as examples of neutral contraventions (under Article 39)

Case C-19/92

Kraus v Land Baden-Württemberg

[1993] ECR I-1663

Case C-415/93

URBSFA v Bosman

[1995] ECR I-4921

, and

Case C-190/98

Graf v Filzmoser Maschinenbau (Advocate General Fennelly 16 September 1999, Court of Justice 27 January 2000); (under Article 43) Kraus ,

Case C-55/94

Gebhard v Consigli dell’Ordine degli Avvocati di Milano

[1995] ECR I-4165

and

Case C-108/96

MacQuen (Advocate General Mischo 16 March 2000, Court of Justice 1 February 2001); and (under Article 49)

Case C-275/92

Customs & Excise v Schindler

[1994] ECR I-1039

Case C-384/93

Alpine Investments v Minister van Financien

[1995] ECR I-1141

and

Case C-398/95

Syndesmos ETTG v Ypourgos Ergasias

[1997] ECR I-3091

. Dr Plender’s classification of Schindler and Alpine Investments as neutral supports Mr Barling’s criticism of the judge’s observation that Syndesmos was the only ‘neutral’ case of contravention under Article 49; but as is by now apparent, these classifications are not clear-cut and cannot be regarded as definitive.

[64]Mr Barling has contended that any substantial impediment to any of the relevant freedoms, even if not involving discrimination or dislocation, is nevertheless a contravention in the third, residual (or neutral) category, and is unlawful unless it can be justified. Dr Plender has challenged that, stressing the importance of examining the precise content of whichever right is in issue, and relying particularly on Graf (and the analysis, in that case, of Bosman ).

[65]The second important difference between counsel is as to the grounds on which a measure which is on it face a contravention can be justified, and in particular whether countering tax avoidance can be a justification. This also raises an issue as to economic and non-economic aims. The other important differences between counsel are on the issue of proportionality and as to the need for a reference to the Court of Justice under Article 234. I shall address these points in that order. ‘Neutral’ contraventions

[66]Bosman was the first case in which the Court of Justice had to consider the obstacles which the system of transfer fees places in the way of professional footballers in moving from one job to another. Bosman was concerned with the transfer system in Belgium but there was no element of discrimination. The system derived from UEFA and FIFA regulations and applied regardless of nationality. Nevertheless it contravened Article 39 (paragraph 95) “In that context, nationals of Member States have in particular the right, which they derive directly from the Treaty, to leave their country of origin to enter the territory of another Member State and reside there in order there to pursue an economic activity (see, inter alia ,

Case C-363/89

Roux v Belgium

[1991] ECR I-273

, paragraph 9, and Singh

[1992] ECR I-4265

, (paragraph 17). Provisions which preclude or deter a national of a Member State from leaving his country of origin in order to exercise his right to freedom of movement therefore constitute an obstacle to that freedom even if they apply without regard to the nationality of the workers concerned (see also

Case C-10/90

Masgio v Bundesknappschaft

[1991] ECR I-1119

, paragraphs 18 and 19).”

[67]Dr Plender has placed particular weight on the analysis of Bosman in Graf , both in the opinion of Advocate General Fennelly and in the judgment of the Court of Justice. Mr Graf, a German national, had been employed in Austria for four years. He gave two months’ notice to terminate his employment in order to take up a new job in Germany. Under Austrian employment law he would have had the right to a compensation payment of two months’ salary, but for the fact that he had himself given notice. His claim under Article 39 (then 48) failed.

[68]In his opinion Advocate General Fennelly examined a number of decisions of the Court of Justice, including Kraus , Bosman , Gebhard , Alpine Investments , ICI v Colmer and Masgio . His conclusions (in a section of his opinion headed ‘The limits of a general test’) merit full quotation (paragraphs 31 to 33):

“It would be possible to construe the broadly worded tests quoted above from Kraus , Gebhard and Bosman as relating solely to the sorts of formal conditions of access to the employment market which were at issue in those and the other cases discussed … above. On the other hand, the Court did not avert expressly to any such limitation of the scope of application of the criteria it laid down in those cases. If, however, it were proposed to treat as restrictions on the exercise of freedom of movement neutral national rules which allegedly preclude, deter, impede, hinder or render less attractive such exercise simply by raising material barriers, for example, by establishing commercial and regulatory conditions in the market in question which are less enticing than in other Member States, or by offering benefits which would be lost in the event that a worker changed employment, those criteria could not be applied in the same way as in the case of a formal condition. Prejudice to the exercise of the freedom of movement of workers or self-employed persons cannot be automatically presumed in all cases where an apparently burdensome national regulation of economic activity, or the loss of a benefit in the case of a change in economic activity, is at issue. Such an approach would be equivalent to applying the Dassonville test [see Procureur du Roi v Dassonville [1974] ECR 837 , a case on movement of goods], in its most far-reaching construction, to freedom of movement of persons. Where an alleged obstacle to freedom of movement does not result from a formal condition of market participation but is instead alleged to arise from some neutral material barrier or disincentive deriving from national regulations, the prejudice to the exercise of Community-law rights must be established. In my view, if the possibility of treating such national rules as restrictions on freedom of movement were admitted, the appropriate criterion would be that which has already been employed by the Court in Bosman and in Alpine Investments in order to reject the application by analogy to certain national rules in the field of the free movement of persons of the approach adopted in Keck [joined cases C-267/91 and C-268/91 [1993] ECR I-6097 ] to national provisions governing selling arrangements for goods: that, proposed by the Commission in this case, of a direct effect on access to the market in question of the worker or self-employed person concerned. Although the Court did not have occasion in either case to state whether fulfilment of this criterion was essential in all cases to establish the existence of a prohibited neutral obstacle to free movement, this appears to me to be necessary if the Treaty is not to be exploited as a means of challenging any national rules whose effect is simply to limit commercial freedom. Thus, neutral national rules could only be deemed to constitute material barriers to market access, if it were established that they had actual effects on market actors akin to exclusion from the market. As in the case of rules regarding selling arrangements in the case of goods, there can be no presumption that neutral national commercial regulations, or those governing pay scales, social protection and other matters of concern to workers , have this effect. In the normal case, the migrant worker must take the national employment market as he finds it. The same holds true for neutral national rules which are alleged to affect the worker’s decision as to whether or not to leave a Member State in order to take up an economic activity in another. This is especially important as regards such possible exit restrictions because the number of formal restrictions on leaving a post is likely to be extremely limited relative to those applicable to taking up employment. If the Court established, in principle, that such material disincentives could, in certain cases, constitute restrictions on freedom of movement, aggrieved persons should be required to reverse that presumption by demonstrating that a particular rule has, in all the circumstances, such a burdensome and deterrent effect on market access as to constitute a direct denial of such access. It is, of course, implicit in such an approach that the existence of the alleged material denial of market access must be ascertained by reference to the circumstances of the particular complainant. My analysis is, I think, similar to that of Advocate General Lenz in Bosman , where he sought to establish a distinction between national rules regarding access to the market and those merely governing the exercise of an economic activity. Advocate General Alber has expressed a different view to Advocate General Lenz in Lehtonen , [Case C-176/96, opinion 22 June 1999] arguing, by reference to Keck , that rules regarding the exercise of a profession are closer to product rules than to those regarding selling arrangements, in that they directly affect citizens, who may thus have to take into account different rules and to acquire new skills every time they migrate from one Member State to another. However, I think that the apparent disagreement arises in part from a different understanding of what is meant by rules governing the exercise of an economic activity. According to the scheme I have outlined above on the basis of the case-law, in particular that governing qualifications, national provisions which require certain skills of economic actors and thus tend to subject migrant workers to a dual regulatory regime are more readily classifiable as formally affecting access or, at the very least, as in Kraus and Choquet [1978] ECR 2293, as being sufficiently closely bound up with market access as to be subjected to a similar regime.”

[69]This important passage seems to have been followed by the Court of

Justice in its judgment (paragraphs 23 and 24):

“Provisions which, even if they are applicable without distinction, preclude or deter a national of a Member State from leaving his country of origin in order to exercise his right to freedom of movement therefore constitute an obstacle to that freedom. However, in order to be capable of constituting such an obstacle, they must affect access of workers to the labour market. Legislation of the kind at issue in the main proceedings is not such as to preclude or deter a worker from ending his contract of employment in order to take a job with another employer, because the entitlement to compensation on termination of employment is not dependent on the worker’s choosing whether or not to stay with his current employer but on a future and hypothetical event, namely the subsequent termination of his contract without such termination being at his own initiative or attributable to him.”

Any non-discriminatory obstacle must have a direct and indisputable effect on the exercise of the particular right in question (in Bosman and Graf , the right to accept an offer of employment in another member state). An indirect or debateable influence or tendency is not enough. The facts of Graf are a clear illustration of this principle.

[70]Mr Barling has argued that IR 35 does provide a particularly burdensome obstacle to anyone wishing to offer employee-like services through a small service company in the United Kingdom. But IR 35 is not an obstacle to anyone who is seeking employment in the United Kingdom. It is arguably an obstacle to someone who wishes to offer the sort of services which an employee would undertake, without having the status of an employee. Such a person cannot fall within Article 39. The judge was right, although not quite for the reasons which he gave, in concluding that Article 39 cannot apply in this case.

[71]However Bosman and Graf have a wider relevance because they help to clarify the principles which were being applied in Kraus , Gebhard and Macquen . Mr Dieter Kraus was a German national who had obtained a further degree in Community law from the University of Edinburgh. Under a law of the Reich still in force in the Länder he could not use his academic title from Edinburgh without authorisation from the Ministry of Sciences and Arts of Baden-Württemberg. He challenged this law as infringing both Article 39 and Article 43 (then 48 and 52). His challenge met with only limited success. Much of the judgment of the Court of Justice is concerned with the rather specialised issue of whether recognition of higher academic degrees is relevant at all to freedom of movement.

[72]The situation in Gebhard was that Mr Reinhard Gebhard had qualified as a German Rechtsanwalt in 1977 and had since 1978 lived with his family in Italy. Until 1989 he worked in association with a firm of Italian lawyers in Milan. Then he started his own firm and claimed to be entitled to refer to himself as avvocato . There were disciplinary proceedings against him and in 1992 he was suspended. The Court of Justice upheld the right of the host state to impose conditions (such as – paragraph 35 – “rules relating to organisation, qualifications, professional ethics, supervision and liability”) but that (paragraph 37) “ … national measures liable to hinder or make less attractive the exercise of fundamental freedoms guaranteed by the Treaty must fulfil four conditions: they must be applied in a non-discriminatory manner; they must be justified by imperative requirements in the general interest; they must be suitable for securing the attainment of the objective which they pursue; and they must not go beyond what is necessary in order to attain it (see

Case C-19/92

Kraus v Land Baden-Württemberg

[1993] ECR I-1663

, paragraph 32).”

[73]Macquen was a similar case concerned with restrictions imposed by Belgian law on practice as an optician. There was an issue as to the correct interpretation of the relevant provisions of Belgian law. Mr Barling has submitted that Macquen goes beyond Bosman and Graf , by requiring a ‘neutral’ restriction to be justified even if it does not impede access to the market. That submission was based on some observations in the opinion of Advocate General Mischo (paragraph 43). But the judgment of the Court of Justice (paragraph 26) appears to follow Gebhard and does not appear to establish any new principle.

[74]What I derive from these authorities (and especially from Graf , which is particularly instructive) is that a neutral, non-discriminatory national measure will not contravene the articles relating to freedom of movement unless it has a direct and demonstrable inhibiting effect on the particular right which is asserted. Otherwise (as the judge almost said) the United Kingdom government would be in breach because of the poor state of this country’s public transport infrastructure. Bosman is a very clear example of a direct and demonstrable (although non-discriminatory) inhibition.

[75]Another clear example (in the context of freedom to provide services without an establishment) is Syndesmos . A Greek law passed in 1985 required licensed tourist guides working with certain tourist agencies to be treated as employees and so subject to Greek employment legislation as regards their working relationship. This provision was not discriminatory but it was held to infringe Article 49 (then 59) because it was a direct and insuperable barrier to tourist guides from other member states providing services as self-employed persons. Moreover it could not be justified by the need to secure industrial peace, which was an economic aim and so inadmissible as a justification.

[76]Schindler was decided on a reference by the Queen’s Bench Division relating to section 2 of the Lotteries and Amusements Act 1976. The Court of Justice held that the section’s general prohibition on trans-border lottery business was a direct inhibition which contravened Article 49 (then 59) unless justified on grounds of social policy. The Court held that lotteries are not inherently harmful but recognised that different member states vary in their attitudes to their acceptability.

[77]In Alpine Investments the Court of Justice reached a similar conclusion in relation to Dutch legislation prohibiting cold-calling by telephone by brokers in commodities futures. The member state from which a trans-border cold call is made is best placed to regulate the practice.

[78]I have thought it right to summarise the facts of these cases because Dr Plender rightly emphasised the need to identify the particular Community right which is said to be infringed, and he rightly warned (as Advocate General Fennelly did in Graf ) against too ready a transposition of general principles from one right to another. But the guiding principle is by now clear. IR 35 cannot be regarded as discriminatory, either in the ordinary sense or under what the judge called dislocation. It does not provide a direct and demonstrable inhibition on the establishment of a business within the United Kingdom, or on the provision of services without establishment. Genuine self-employed activities will not be affected and a business of providing employee-like services will be taxed as if there was a real employment situation.

[79]The judge rejected Mr Barling’s arguments on Articles 39 and 43 but concluded that it was “just arguable” (because of Syndesmos ) that IR 35 was a relevant restriction within Article 49. He therefore went on to consider the issue of justification in relation to Article 49. I agree in relation to Articles 39 and 43 but I respectfully disagree, for the reasons set out above, in relation to Article 49. I would accept the submission embodied in the Revenue’s respondent’s notice. Justification: tax avoidance

[80]On the view which I take this issue does not arise, but I will state my views on it briefly, since the court has heard full argument. The main focus of the argument has been as to whether the prevention of tax avoidance (as opposed to fiscal supervision and fiscal cohesion, two other expressions which recur in Community jurisprudence) can be a proper ground of justification.

[81]The issue about tax avoidance resolved itself, in the course of argument, into whether the Court of Justice was in Case 270/83 Commission v France [1986] ECR 273 and in ICI v Colmer holding that all measures against tax avoidance, or only discriminatory measures against tax avoidance, were inadmissible by way of justification. The crucial texts are paragraphs 24 and 25 of the former judgment and paragraphs 24 to 29 of the latter. In my view it is clear that the Court was referring only to discriminatory measures. I note also that in paragraph 26 of ICI v Colmer the suggestion that the more artificial the tax avoidance which has to be countered the easier it may be to find justification. IR 35 is aimed at relatively artificial avoidance since it counters only activities which are the provision, in the guise of self-employment (normally through a corporate vehicle), of employee-like services.

[82]There is a general principle (mentioned in Syndesmos ) that economic aims cannot be a justification for measures which offend fundamental principles of Community law. That principle is in a sense obvious, since a member state cannot rely on its own economic interests to flout the principles on which the common market is founded. But the principle must be understood in its context. It cannot be used to undermine the principle that direct taxation is at present, and unless and until harmonisation takes place, a matter for the legislatures of member states. Diminution of tax revenue cannot be a justification for unequal treatment ( ICI v Colmer , paragraph 28) but the general prohibition on economic aims does not seem to take the argument any further. Proportionality

[83]If the IR 35 legislation had to be justified an essential part of the process of justification would be for the Revenue to show that its terms were appropriate and necessary to achieve its objectives; that whenever there was a choice between different measures which might be appropriate the least onerous had been chosen; and that the burdens imposed by the measures were not disproportionate to the aims pursued.

[84]The judge found that the Revenue had discharged that burden. In particular he dismissed what he called the ‘sledgehammer’ argument to the effect that the only real problem was the extreme ‘Friday to Mondays’ and that that abuse could have been dealt with by simpler and less far-reaching measures. The judge saw the objectives of IR 35 as preventing not only tax evasion but also (and more importantly) widespread tax avoidance by persons who saw themselves as legitimately exploiting a gap in the law. The judge also rejected various suggestions (made in the course of argument before him) as to how the Revenue’s objectives might have been achieved in other ways. The suggestions which he referred to were changing the method of taxing dividends, increasing the rate of corporation tax, increasing capital allowances and abolishing NIC. He said that in any event he was satisfied “ … that for me as a judge to canvass other possible methods of fiscal reform is wholly inappropriate.” On that point he referred to the observations of Lord Slynn in Queen v Chief Constable of Sussex ex parte International Trader’s Ferry Ltd

[1999] 2 AC 418

, 439.

[85]Mr Barling has criticised the judge’s approach as failing to discharge his function when an issue of proportionality is raised. He cited the judgment of this court in Queen v Secretary of State for Health ex parte Eastside Cheese

[1999] 3 CMLR 123

at pp.143 and 146:

“In principle the decision on proportionality has to be taken by the national court which is seised of an issue on Article [30] EC, subject of course to any possible reference to the Court of Justice. But in the case of a legislative measure the national court must not simply accept the view of the national legislature or confine itself to deciding whether what the legislature has enacted is reasonable … The judge’s task was … to see whether the exercise of the Secretary of State’s power … had been objectively justified and had been shown not to be disproportionate. The test is more demanding than that of ‘manifest error’ and is also more demanding than that of Wednesbury unreasonableness.”

[86]That approach has been confirmed, he submitted, by the House of Lords in Queen (Daly) v Secretary of State for the Home Department

[2001] 2 WLR 1622

, 1635. Contrasting the principle of proportionality with the traditional grounds of judicial review, Lord Steyn said:

“Most cases would be decided in the same way whichever approach is adopted. But the intensity of review is somewhat greater under the proportionality approach. Making due allowance for important structural differences between various convention rights, which I do not propose to discuss, a few generalisations are perhaps permissible. I would mention three concrete differences without suggesting that my statement is exhaustive. First, the doctrine of proportionality may require the reviewing court to assess the balance which the decision maker has struck, not merely whether it is within the range of rational or reasonable decisions. Secondly, the proportionality test may go further than the traditional grounds of review inasmuch as it may require attention to be directed to the relative weight accorded to interests and considerations. Thirdly, even the heightened scrutiny test developed in R v Ministry of Defence, Ex p Smith [1996] QB 517 , 554 is not necessarily appropriate to the protection of human rights.”

[87]I have to say that I see some validity in these criticisms. If proportionality were an issue in this appeal I would be inclined to think that the judge did not go far enough in asking himself whether the IR 35 measures were the least onerous which could be adopted in order to achieve their objective. In expressing that tentative view I am well aware of the difficulty of the task, and I sympathise with the judge’s view that it was not for him to enter the political arena.

[88]But tax legislation raises technical as well as political issues. The technical difficulties of finding fair and practical tests for distinguishing between employment and self-employment for tax purposes go back to the nineteenth century. But recent developments have intensified the problem. In the knowledge-based sectors from which PCG draws most of its members, the dividing line is often debatable. Moreover the ever-increasing responsibilities of employers for their employees have led to downsizing, outsourcing, ‘bodyshopping’ and similar practices. Many redundant employees, having found that employed status does not guarantee long-term job security, wish to explore the possibilities of self-employment.

[89]All these considerations have led me to wonder whether it might not have been possible to bring forward measures which accorded some recognition to the existence of a sort of no-man’s land between Schedule D and Schedule E, rather than insisting on the gulf which exists in theory (but not, always, in practice) between them. Such measures might still have contained robust sanctions against unacceptable tax avoidance. However it is not necessary or appropriate to express any final view, since in my judgment the appeal has fallen at an earlier hurdle. PCG and its members may continue to work through democratic means for amendment of IR 35 so as to meet their complaints, but they have in my judgment failed to strike it down under Community law. Article 234

[90]At the end of his submissions Mr Barling asked the court to conclude that it was clear that his appeal should be allowed; or if the court was not of that view, that there should be a reference to the Court of Justice under Article 234. For the reasons already mentioned, I consider that the appeal should be dismissed; and I do not consider it necessary or appropriate for this court to make a reference under Article 234.

[91]The principles stated by Sir Thomas Bingham MR in Queen v International Stock Exchange ex parte Else

[1993] QB 534

still hold good. But in applying them the court must also take account of the guidance given by the court (following European authority) in Trinity Mirror plc v Commissioners of Customs & Excise

[2000] 2 CMLR 759

, 783-5. The latter case has drawn attention both to the very heavy case-load of the Court of Justice and also to the greater familiarity with Community law which domestic courts now have. In this case as in many others, the real difficulty is not in ascertaining the relevant principles of Community law but in applying them to the facts; and that is a task for the national court. Lord Justice Dyson:

[92]I agree.

Lord Justice Auld:

[93]I also agree.

Order

Appeal Dismissed; with costs; application for permission to appeal to the House of Lords refused. (Order not part of approved judgment)

Cited in 36 later judgments

and 27 more of the most senior