I. Introduction
BMW challenges the 70 mph speed limit on British motorways as an obstacle to the free movement of goods. It also says that the requirement to comply with British labour laws when it employs staff at its Oxford plant is a restriction on its freedom of establishment, as is the fact that it must pay British taxes. Far-fetched? Well there is nothing in the ECJ's recent case law on free movement to prevent BMW from running such arguments. If BMW chooses to do so, the Court of Justice may find that the UK rules are “restrictions” or “obstacles” to BMW's freedom of movement, contrary to Articles 28, 39, 43 and 49 of the EC Treaty, and are thus prima facie unlawful. The burden would then shift to the UK both to justify its national rules, by pointing to good policy reasons taking precedence over free movement, and to show that its laws are proportionate.
Speed limits can of course be justified on the grounds of road safety.Footnote 1 But the problem lies with proportionality. While setting a speed limit might be a suitable means of achieving the public safety objective, it is not obvious that a 70 mph limit is no more restrictive than necessary to achieve that objective (why not 69, 68 or 71 or 72 mph?).Footnote 2 Employment laws can also be justified on the grounds of worker protection,Footnote 3 but again they raise the question of whether the challenged rules can be considered proportionate. The taxation case is more difficult still: pure economic arguments, such as the loss of tax revenueFootnote 4 or the cost to the national exchequer,Footnote 5 are not even accepted as justifications by the ECJ and so the UK's case is likely to fail at this stage, even before the question of proportionality comes into play.
These examples highlight a more general problem facing parties when defending non-discriminatory national rules designed to protect “social” interests found to restrict free movement. As the trade unions found in Laval Footnote 6 and its sister case, Viking,Footnote 7 despite the Court's talk of balance between economic and social policy,Footnote 8 in fact the ECJ's “restrictions” analysis gives primacy to the economic freedoms (Articles 43 and 49) and creates a presumption that the national rule is unlawful. This puts the defendant, usually the state, on the back foot, defending national social policy choices and showing that the legislation is proportionate. Courts are also placed in the difficult position of having to make decisions about the legitimacy and proportionality of national policy choices when they lack the institutional capacity to gather and process the relevant economic/social/cultural information. The Sunday trading saga in the late 1980sFootnote 9 illustrated this point all too clearly: national courts reached different conclusions on the proportionality of a ban on shops opening on Sundays.Footnote 10 This litigation eventually led the ECJ in the (in)famous Keck Footnote 11 decision to carve out an area of national regulation – certain non-discriminatory selling arrangements – to which Article 28 would not apply. However, the 2009 decision of Commission v. Italy (trailers) Footnote 12 suggests that even the Keck settlement is being re-thought in light of the “restrictions” or “market access” analysis which is now the vogue in respect of free movement of persons, services and capital.
This article therefore seeks to address the question of whether there should be limits on the scope of intervention of EC law into national regulation. If so, what are they and what form might they take? This is not just a technical matter for black letter lawyers; it goes to the heart of the federal question – what sort of EU is the Court of Justice trying to shape? If, as generally accepted, Member States should retain some degree of regulatory autonomy how much space can there, and should there, be for Member State action unhindered by Community law? I do not wish to examine this question through the lens of positive integration (the adoption of Community measures), where the safeguards of powerFootnote 13 (competence, subsidiarity and proportionalityFootnote 14) and process (e.g. Member State voting in the Council of MinistersFootnote 15) go a long way to protecting (some) state autonomy. Instead, I will consider the question through the lens of negative integration (the use of Treaty provisions to strike down national law),Footnote 16 which is the area where I perceive there to be the greatest problem. Is there a way that the Court of Justice can combine its push to remove legislative dead wood at national level, and thus continue the dynamic process of creating a single market, with preventing a total free-for-all where all national legislation is potentially up for review as being incompatible with EC law?
To help answer this question I will compare the ECJ's favoured “restrictions” approach with the largely discrimination-based approach to the Dormant Commerce Clause adopted by the US Supreme Court. Superficially at least, the US is a good comparator for the EU.Footnote 17 Broadly speaking, both have similar sized economies. Both are based on a federal or quasi-federal structureFootnote 18 and both recognise that, in certain circumstances, “federal” law needs to control state activity where that activity interferes with interstate commerce and thus the unity of the Union or the single market.Footnote 19
Yet, the parallels between the EU and the US are more apparent than real.Footnote 20 The US has been explicitly concerned with nation-building since its formation over two hundred years ago. Its experience – and thus its maturity – suggests that it is now at the “market maintenance” stage:Footnote 21 the federal regime has already reached the status quo which it wishes to maintain and this permits a lighter federal hand on the tiller. The US Supreme Court therefore takes the view that federal law will intervene only where the state law is discriminatory or unduly burdensome of interstate commerce (as we shall see, this latter category is contestedFootnote 22). The approach of the US Supreme Court has two consequences: first, it allows a much greater degree of deference to state actors and to state regulation; and second, it creates a stable, more certain situation which is not intended to alter market structures.Footnote 23
If the US has reached a “comfortable middle age”,Footnote 24 the EU is a “work in progress”.Footnote 25 It finds itself at what Maduro terms the “market building” stage, focusing on “promoting the new set of rights brought by the larger integrated area and to break the path-dependence of actors from national systems”.Footnote 26 Thus, in contrast to the US, the EU's situation implies a more activist intervention by “federal” law, one based on removing obstacles or restrictions to free movement i.e. eliminating national rules which have accumulated, often over the centuries.Footnote 27 This approach pays little heed as to whether the rule is discriminatory. It focuses instead on the (potentially) restrictive effect of the rule on the out-of-state trader wishing to gain a foothold on the market in another Member State.
While active judicial intervention, through the restrictions model, may generally be necessary in the EU to eliminate obstacles to free movement such as packaging requirements, rules on authorisation criteria and provisions on the recognition of qualifications, it creates problems when applied to non-discriminatory measures which structure the national market on which trade occurs, such as employment and taxation laws and other welfare measures, as the BMW example shows. This article therefore argues that the type of federalism that the Court of Justice is responsible for shaping must leave space for the sub-units (the states) to regulate and develop at least the matters which form the core of the welfare state, as well as social policy more generally, largely unhindered by the application of EC law. Failure to do so may well lead EU citizens blaming the EU for the failure of the European social model which is so dependent on national welfare policies for its substance.Footnote 28
It is here that the ECJ could draw inspiration from the case law of the US Supreme Court under the Dormant Commerce Clause. Non-discriminatory state rules should enjoy a presumption of legality. Only if the trader can show that the rule particularly affects his or her market access, an effect which goes beyond simply incurring additional costs, will the rule's compatibility with the four freedoms be considered. In order to make this argument I shall start by examining the different degrees of intensity with which federal law can control state decision-making in the field of free movement.
II. Approaches by Federal law to Controlling State Law
I will suggest that there are four different levels by which federal law can control state law in this field. I shall describe these as the zero impact, the low impact, the high impact and the total impact approaches.
The zero impact approach would, as the name suggests, mean that federal law does not attempt to control state action in the commerce field in any way: that state and federal government action remain distinct, autonomous and supreme in their two separate spheres. This approach, which comes close to the system of dual federalism recognised in the US from 1789–1937, proved to be unworkable, in part due to the artificial categories that the Supreme Court was forced to draw,Footnote 29 and in part due to the practical difficulties of delimiting the two spheres of competence. It has now been discredited as a system of constitutional interpretation.
The total impact approach is the opposite: all state laws are subject to challenge as being potentially incompatible with federal law unless they can be justified. This would envisage a type of economic due process review,Footnote 30 as employed by the US Supreme Court during the Lochner era (c.1897–1937) under the Fourteenth Amendment.Footnote 31 However, by 1937 this approach had also been discredited.Footnote 32 This was subsequently confirmed in Williamson v. Lee Optical of Oklahoma Footnote 33 where Justice Douglas declared the Court no longer struck down “state laws, regulatory of business and industrial conditions, because they may be unwise, improvident, or out of harmony with a particular school of thought. … ‘For protection against abuses by legislatures the people must resort to the polls, not to the courts.’”Footnote 34
In the EU, the Court of Justice also does not have general jurisdiction to review all state legislation: it has jurisdiction only where the rule under review has an effect, actual or potential, on inter-state trade/movement. Wholly internal situations, however disputed the category, remain outside the scope of EC law.Footnote 35 Therefore the ECJ has no competence to adjudicate on every state rule;Footnote 36 it can do so only where the state rule has some effect on inter-state trade.Footnote 37 For these reasons, I will not discuss the total impact theory further.
More nuanced are the high impact and low impact approaches. The former, which envisages careful scrutiny of most state rules by the federal tribunal, is more closely associated with the EU; the latter, which envisages less scrutiny of state rules, with the US. Since these two approaches are linked to two identifiable legal systems we shall consider them in detail in Sections III and IV below.
III. The Low Impact Approach
A. The Essence of the Approach
The low impact approach envisages a low level of intervention by federal law on the state rules. Only those rules which are discriminatory will be subject to challenge under federal law; all other state rules are lawful. So, under this model, a state can determine whether it should have a minimum wage and if so at what level. This decision will not be subject to challenge under federal law provided that, say, the state does not set a higher minimum wage rate for nationals than for migrants. Were it to do so, the discriminatory differential could be challenged (i.e. the higher wage rate) but not the substance of the underlying rule (i.e. the state's decision to impose a minimum wage).
Various, often closely related, rationales are offered for the prohibition of discrimination in the trade law context. First, discriminatory rules are considered unlawful because they allow irrelevant factors to be taken into account (e.g. the need to protect the insider from the outsider). This weakens the “concept of the union”, whether a political union, as in the US, or a limited federal union, as in the EU.Footnote 38 More specifically, discrimination undermines the solidarity between the constituent elements of the polity. This argument is particularly important in the EU where the principle of solidarity has long been recognised in both the EC and EU Treaties.Footnote 39 In the US such arguments more usually take the form that, absent the Dormant Commerce Clause's prohibition on discrimination, the federal system would face collapse due to growing resentment between the states which would ultimately lead to retaliation,Footnote 40 a very real concern for the drafters of the US Constitution in 1787.
A second rationale for the prohibition of discrimination is more economic than political: discrimination – and protectionism – are inefficient because they divert “business away from presumptively low-cost [foreign] producers [to high-cost (local) producers] without any colorable justification in terms of benefit that deserves approval from the point of view of the nation as a whole”.Footnote 41
A third rationale, and one strongly pushed by Justice Stone in the 1930s in the US,Footnote 42 concerns the operation of the political process: federal law should be used to invalidate discriminatory state rules because they are the product of a successfully functioning democratic process operating at state level in which the outsider most affected by the rules is not represented in that process.Footnote 43 Federal supervision is therefore necessary to inject a national perspectiveFootnote 44 or, putting it another way, to give out-of-state interests “virtual representation”.Footnote 45
Some criticise this argument on the grounds that in-state actors, particularly consumers, often share the interests of the out-of-state actors and so act as a proxy for them in the state legislative process.Footnote 46 Nevertheless this third rationale has particular force in the US because, as Halberstam explains,Footnote 47 the federal Constitution grants “each level of or unit of government a set of regulatory tools that may be used without regards to whether the exercise of these powers serves the system of democratic governance as a whole”. He argues that it is less relevant in the EU because the dominant approach is one of fidelity which holds public institutions to a duty to act responsibly and to promote the well-being of the entire political system.Footnote 48 According to this view, it is less likely that EU states will adopt overtly discriminatory legislation because of their duties of loyalty and cooperation under Article 10 EC.Footnote 49 If this is correct it therefore makes more sense in the EU context to adopt the restrictions, rather than the discrimination, approach considered in section IV below.
Whichever rationale is favoured, it is clear that the discrimination model has a good and well respected pedigree. Its great advantage is that it leaves room for state action, largely unhindered by concerns that the federal level will intervene. State rules can therefore respond to the demands of the local electorate, so long as that electorate does not insist on discriminatory rules being put into place. This results in a diversity of rules between the statesFootnote 50 and the creation of a space for states to experiment so that other states – and even the federal government – can learn from the experience without incurring the risks that might result from a nationwide experiment.
The weakness of the discrimination approach lies in the fact that non-discriminatory national rules are not subject to “federal” challenge even when they interfere with a trader's ability to provide services or goods. So the (non-discriminatory) transfer fees in Bosman Footnote 51 would remain on the rule book as would the (non-discriminatory) rule in Commission v. Greece (opticians) Footnote 52 that qualified opticians could operate not a chain but only one optician's shop in Greece. The desire to attain a fully unified market will therefore always be frustrated so long as the discrimination model is in place.
The discrimination approach also suffers from both technical and conceptual disadvantages. The technical drawbacks concern identifying the comparator. The model is premised on the fact that migrants and nationals are similarly situated when, by definition, they are not.Footnote 53 Even if this intellectual hurdle can be surmounted, the choice of comparator is not always obvious and may be determinative of the outcome.Footnote 54 Even identifying the type of discrimination at issue is not straightforward.Footnote 55 For example, is a residence requirement imposed on a service provider a directly or indirectly discriminatory rule?Footnote 56
Whatever its problems, it is the discrimination approach which has dominated the US Supreme Court's jurisprudence under the Dormant Commerce Clause and it is to this that we shall now turn.
B. The US Supreme Court and the Discrimination Approach
Federal intervention into state laws starts from a weaker base in the US than in the EU. While the EC Treaty contains provisions which expressly provide the basis for ECJ intervening in national systems (e.g. Articles 28, 29, 90, 39, 43, 49 and 56), no express equivalent exists in the US. Article 1, section 8, clause 3 of the US Constitution merely provides:
The Congress shall have Power … to regulate Commerce with foreign Nations, and among the several States …
This “Commerce Clause” gives the federal government the power to legislate in respect of commerce (very broadly defined); it does not prohibit state-created restrictions on inter-state commerce.Footnote 57 However, “by interpreting ‘these great silences of the Constitution’ … the Supreme Court has limited the scope of what the state might do”.Footnote 58 In other words, the Supreme Court has longFootnote 59 read into the commerce clause the negative power (the so-called “dormant”Footnote 60 or “negative”Footnote 61 commerce clause) to restrict “permissible state regulation of interstate commerce”,Footnote 62 albeit that some Supreme Court justices, notably Justice ScaliaFootnote 63 and increasingly Justice Thomas,Footnote 64 remain deeply suspicious about the implication of such a broad, and, in their view, illegitimate prohibition into the Constitution. Nevertheless, they remain in the minority on this point and most commentators accept that the US Supreme Court has developed a two-tiered framework of analysis of state laws under the dormant commerce clause: (1) the discrimination tier and (2) the non-discrimination tier. We shall examine these in turn.
1. The Discriminatory Tier
The discriminatory tier is, in fact, divided in two:Footnote 65 facially discriminatory measures and measures discriminatory in (practical) effect. Facially discriminatory (i.e. directly discriminatory) statutes are subject to a “strict scrutiny” reviewFootnote 66 which means the court checks whether the statute serves “a legitimate local purpose” and Footnote 67 whether there is an “absence of non-discriminatory alternatives”.Footnote 68 In practice, it is extremely rareFootnote 69 for any state legislation to satisfy this hurdleFootnote 70 so that the Supreme Court sometimes says that facially discriminatory rules are subject to a virtually per se rule of invalidity,Footnote 71 especially when economic protectionism is involved.Footnote 72
Statutes which are discriminatory in effect (or “discriminatory impact”) are also subject to strict scrutiny review, albeit of a somewhat more relaxed standard.Footnote 73
2. The Non-discrimination Tier
In respect of so-called non-discriminatory rules the Supreme Court generally says that it adopts a balancing approach. The balancing test is set out in the now classic statement by Justice Stewart for a unanimous court in Pike v. Bruce Church:Footnote 74
Where the statute regulates even-handedly to effectuate a legitimate local public interest, and its effects on interstate commerce are only incidental, it will be upheld unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefits. … If a legitimate local purpose is found, then the question becomes one of degree. And the extent of the burden that will be tolerated will of course depend on the nature of the local interest involved, and on whether it could be promoted as well with a lesser impact on interstate activities. Occasionally the Court has candidly undertaken a balancing approach to resolve these issues.Footnote 75
There is much controversy over the meaning of this test. Blasi suggests that it embodies three threshold requirements: the measure must be (1) non-discriminatory (i.e. it regulates “even-handedly”); (2) it effectuates a legitimate local purpose; and (3) its effects on inter-state trade are only incidental.Footnote 76 If these conditions are satisfied the state measure is lawful and so no balancing needs to be undertaken. Therefore, he says that “the real bite in the Pike standard may lie in the threshold considerations, not the balancing formula”. Yet for Day the emphasis is on the later components listed in Pike and the balancing contained therein, namely: (1) the burden imposed on commerce; (2) putative local benefits; and (3) whether the local interest could be promoted as well with a lesser impact on interstate activities.Footnote 77
The uncertainty generated by the Pike formula is exacerbated by the fact that, as Regan argues in his seminal article on the Dormant Commerce Clause, most of the cases cited as examples of balancing, including Pike itself,Footnote 78 in fact concern “purposeful economic protectionism”Footnote 79 to which the rules on discrimination outlined above apply (presumption of illegality, justified only exceptionally).
Assuming Pike does genuinely embody a balancing test, the Supreme Court's approach appears to resonate with the methodology used by the ECJ in free movement cases like Cassis Footnote 80 and Säger Footnote 81: (1) putative local benefits=mandatory/public interest requirement; (2) lesser impact on interstate activities=proportionality. However, there is a key difference between the Supreme Court and the ECJ's approach to balancing.Footnote 82 In Cassis, as (re)interpreted,Footnote 83 the ECJ starts with the presumption that the national regulation is unlawful. Only if the defendant can show the existence of a mandatory requirement and that the measure is proportionate will the measure be considered lawful. By contrast, the starting point of the US approach is the legality of the state regulation. As Pike puts it, an even-handed national rule will be upheld unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefit. Where the claimant does manage to show that the burden imposed on inter-state commerce is clearly excessive then balancing must occur. However, we can see a further difference between the US and the EU approaches: Pike suggests that the balance occurs before the finding of illegality. This means that in the US the courts undertake a more genuine balance between the competing interests (e.g. economic and social), thereby avoiding the possibility of one policy interest (economic) immediately taking priority over the other, as happened in Viking and Laval. I will return to this point in Section V.
In this section I want to focus on the presumption of legality of the state rule. This point was emphasised in Breard v. Alexandria Footnote 84 where the Supreme Court upheld a ban on door-to-door selling without prior agreement of the householder. Having noted that the rule was non-discriminatory (“regulation that leaves out-of-state sellers on the same basis as local sellers cannot be invalid for that reason”Footnote 85), the Court concluded that “When there is a reasonable basis for legislation to protect the social, as distinguished from the economic, welfare of a community, it is not for this Court, because of the Commerce Clause to deny the exercise locally of the sovereign power of Louisiana”.Footnote 86 In Huron Portland Cement Co v. Detroit Footnote 87 the Supreme Court dispensed altogether with the need for a balancing test. In upholding Detroit's Smoke Abatement Code, which set maximum limits on the amount of smoke that could be emitted into the atmosphere. The Supreme Court said that the “burden on interstate commerce needs no extended discussion”.Footnote 88
The general insulation of non-discriminatory national laws from attack under the Commerce Clause is important to maintaining the integrity of a federal system. As the US Supreme Court put it in Huron Portland Cement Footnote 89 “it must be borne in mind that the Constitution when conferring upon Congress the regulation of commerce, …. never intended to cut the States off from legislating on all subjects relating to the health, life, and safety of their citizens, though the legislation might indirectly affect the commerce of the country”.Footnote 90
The willingness to uphold the constitutionality of non-discriminatory state rules without engaging in any form of balancing can be explained in part by the fact that some justices “abhor” the balancing test because, as Justice Scalia put it, the test allows judges limitless discretion.Footnote 91 In his dissent in Kassel v. Consolidated Freightways,Footnote 92 Justice Rehnquist, with whom Chief Justice Burger and Justice Stewart joined, complained that the balancing approach “arrogate[s] to this Court functions of forming public policy, functions which, in the absence of Congressional action, were left by the Framers of the Constitution to state legislatures”.Footnote 93 They preferred a test which required that state interests be “demonstrably trivial while the burden on commerce is great”Footnote 94 before declaring a state measure unconstitutional.
Kassel concerned an Iowa law prohibiting the use of double-trailer trucks within the state, with the result that larger trucks had to drive greater distances to bypass Iowa. The majority (Powell, White, Blackmun and Stevens) found, under the balancing test,Footnote 95 that the state's arguments based on safety were illusory and that the Iowa law “substantially burdens interstate commerce”.Footnote 96 However, Justices Brennan and Marshall, while concurring with the majority on the result and paying lip-service to the balancing test, in fact advocated a “rational purpose” test, a lighter form of review. They said “the only relevant evidence concerns whether the lawmakers could rationally have believed that the challenged regulation would foster those purposes [the lawmakers had for maintaining the regulation]”.Footnote 97 This approach was also adopted by the US Court of Appeals 7th Circuit's decision in Proctor & Gamble v. Chicago,Footnote 98 which concerned a Chicago ordinance banning the use of detergents containing phosphates.Footnote 99 The Court said:Footnote 100
It is our view, however, that if the burden on interstate commerce is slight, and the area of legislation is one that is properly of local concern, the means chosen to accomplish this end should be deemed reasonably effective unless the party attacking the legislation demonstrates the contrary by clear and convincing proof. If it is determined that this presumption should be applied, no further balancing need by undertaken. The end has already been deemed legitimate and the burden on interstate commerce slight. If the legislation is a reasonable means to that end it is constitutional.
Despite the multiplicity of approaches which fall under the umbrella of “balancing”, it is two aspects of the Supreme Court's approach to non-discriminatory rules that I want to rely on in Section V to argue (1) that non-discriminatory EC rules which have only an incidental effect on free movement should also benefit from a presumption of legality; and (2) those that do have a substantial effect on free movement should be subject to a more genuine balancing. Before reaching this point, I want to consider the ECJ's current approach to its free movement case law which, as I will suggest, is currently based on a “high impact” approach to federal intervention.
IV. The High Impact Approach
A. Introduction
The high impact approach requires intense intervention by the federal authorities. All state rules which affect inter-state trade constitute a barrier or obstacle to “market access” or a “restriction” on free movement and so are presumptively unlawful; the burden rests on the state to justify the continued existence of the rule and to show it is proportionate. So, returning to the example of the minimum wage, the very existence of a law on the minimum wage could be challenged by employers or workers as a “restriction” on their ability to access the host state's market since, although non-discriminatory (and so any challenge would fail under the non-discrimination test), minimum wage legislation nevertheless interferes with the freedom of migrants to sell their services for less than the statutory minimum.Footnote 101
The advantage of the high impact approach is that almost all state rules affecting inter-state trade can be challenged and thus need justifying by the state. This technique is therefore an effective way of removing decades of accumulated regulation which impedes the creation of a unified market. The disadvantage of this approach is that it is strongly deregulatory: the working presumption is that such national rules are unlawful unless justified and often it is hard for states to justify their legislation. And, if the national rules are struck down, it is only rarely that they are re-enacted at the federal level. This risks creating what Advocate General Tizzano describes as “a market without rules”,Footnote 102 thereby threatening the legitimacy of both the federal system (what right has the federal government to intervene in decisions taken by democratically elected state governments?) and the state government (why should the state bother to regulate and thus to experiment when its rules are likely to be found contrary to federal law?).
The legal tool most closely associated with achieving the high impact is the controversial “market access” or “restrictions/obstacles” approach. This is the approach which has taken firm hold in the EU.
B. The ECJ and the Market Access Approach
1. Introduction
Although the original Treaty provisions, particularly those on persons, services and capital, suggested that the Treaty was underpinned by a prohibition against discrimination,Footnote 103 in fact, since Säger,Footnote 104 the “market access” or “restrictions” approach has been in the ascendancy. In Säger the Court said that Article 49 required “not only the elimination of all discrimination against a person providing services on the ground of his nationality but also the abolition of any restriction, even if it applies without distinction to national providers of services and to those of other Member States, when it is liable to prohibit or otherwise impede the activities of a provider of services established in another Member State where he lawfully provides similar services”. The Court continued that any such “restriction”,Footnote 105 could be justified by imperative reasons relating to the public interest.Footnote 106 Despite controversy as to whether it was appropriate to apply this test beyond the field of services,Footnote 107 the Court has nevertheless done so: to the free movement of workers,Footnote 108 freedom of establishment,Footnote 109 free movement of capital,Footnote 110 and free movement of citizens.Footnote 111
This market access approach also appears to be the driving force behind the earlier cases on free movement of goods such as Dassonville (“All trading rules enacted by Member States which are capable of hindering, directly or indirectly, actually or potentially, intra-Community trade”)Footnote 112 and the principle of mutual recognition in Cassis de Dijon:Footnote 113 goods lawfully produced in one Member State should presumptively have free and unrestricted access to the market in another Member State. Although Keck appeared to send a strong message that the market access approach was being curtailed, at least in respect of national rules regulating certain selling arrangements, the test has recently received strong support from the ECJ in Commission v. Italy (trailers). Footnote 114 There the Court identified three situations where a rule could be considered a measure having equivalent effect to a quantitative restriction and so breach Article 28: (1) measures adopted by a Member State the objectFootnote 115 or effectFootnote 116 of which is to treat products coming from other Member States less favourably; (2) indistinctly applicable product requirements;Footnote 117 and (3) “Any other measure which hinders access of products originating in other Member States to the market of a Member State”. While the first two limbs reflect a discrimination approach, the third limb contains a renewed commitment to the market access approach. It was this third limb that the Court relied on in Trailers where the Court said that “a prohibition on the use of a product in the territory of a Member State has a considerable influence on the behaviour of consumers, which, in its turn, affects the access of that product to the market of that Member State”.Footnote 118
2. The Market Access Approach and the Problems it has Generated
(a) What is meant by market access?
While the “market access” test gives EC law broad supervisory power over national law, the problem is that it is based on an “inherently nebulous”Footnote 119 idea. Spaventa offers three possible interpretations.Footnote 120 The first, and narrowest, is that barriers to market access are those created by circumstances or legislation that make it “more costly for new firms to enter an industry”.Footnote 121 There is some overlap between this approach and the discrimination model if it is construed to mean that the out-of-state trader has to incur greater costs than the in-state trader to gain access to the host state's market.
The second, and much broader, approach to market access is that “any regulation can be seen as a potential barrier to access, since any regulation imposes and implies compliance costs”. This version of the market access approach comes close to allowing the complainant to object to the very existence of the national rule. This seems to be the view of the Court in the golden share case, Commission v. UK (British Airports Authority (BAA)). Footnote 122 BAA's Articles of Association prevented any person (other than a permitted person) from acquiring BAA shares carrying the right to more than 15% of the equity in BAA. The Court said that “Rules which limit the acquisition of shareholdings … constitute a restriction on the free movement of capital”.Footnote 123 The Court added that although the restrictions were non-discriminatory “they affect the position of a person acquiring a shareholding as such and are thus liable to deter investors from other Member States from making such investments and, consequently, affect access to the market”.Footnote 124 The answer to the question “What is the deterrence?” is presumably the very existence of the rule: had the rule not existed there would have been no breach. If this analysis is correct then it could be argued that rules laying down, say, a minimum school leaving age or a minimum age at which a person can drive a car or drink alcohol are liable to deter young people from exercising their rights of free movement. This very broad approach to a “restriction”, as the UK warned in BAA itself, would lead to a repetition of the problems generated by Dassonville which the Court was forced to correct in Keck.Footnote 125 Nevertheless, the Court persisted with its “market access” jurisprudence.
The third approach is what Spaventa describes as “intuitive”: rules which the Court intuitively feels interfere with intra-Community trade should be subject to judicial scrutiny while rules considered neutral as regards intra-Community trade should not.Footnote 126 Snell thinks this is the approach which predominates in the Court's case law. He argues that the notion of market access conceals rather than clarifies.Footnote 127 He says that the term covers a range of situations including discrimination against free movers or free movement,Footnote 128 substantial or direct obstacles to trade or free movement,Footnote 129 and substantial reductions in profitability.Footnote 130 He concludes “As the term lacks a clear content, the Court may use it freely either to approve or to condemn measures that it happens to like or dislike. Market access may simply provide a sophisticated-sounding garb that conceals decisions based on intuition.”Footnote 131
Laval Footnote 132 provides a good example of the Court's intuitive approach. The case concerned a Latvian company, which won a contract to refurbish a school in Sweden using its own Latvian workers. It faced ruinous industrial action by the Swedish trade unions when it refused to apply Swedish collectively agreed terms and conditions. Although the industrial action was permissible under Swedish law, Laval brought proceedings in the Swedish labour court claiming that the industrial action was contrary to Article 49 EC. The Court said “the right of trade unions of a Member State to take collective action by which undertakings established in other Member States may be forced to sign the collective agreement for the building sector … is liable to make it less attractive, or more difficult, for such undertakings to carry out construction work in Sweden, and therefore constitutes a restriction on the freedom to provide services within the meaning of Article 49 EC”.Footnote 133
But how, precisely, did the collective action impose costs which made it “more difficult” or “less attractive” for Laval to operate in Sweden? Deakin suggests that one answer might be “more difficult” in relation to the situation which would have prevailed had Latvian law and/or Latvian collective agreements applied”.Footnote 134 If this is correct, it means that the effect of the restrictions analysis is to enable service providers to export their home country laws and have them applied extra-territorially. Another answer may be intuition: Laval was a Latvian company and it was having difficulty fulfilling a contract which it had legitimately won, largely due to its cheaper labour costs. The Swedish strike action therefore prevented it from doing the very thing that EU accession had promised to the new Member States, namely free access to the markets in services in other Member States; and that was enough to trigger Article 49.
But if the effect of Laval is to allow challenges to host state rules which differ from those which apply at home this suggests that a mere difference between the rules of two states is sufficient to constitute a restriction on free movement, an understanding that the Court has rejected in the past.Footnote 135 However, following Laval there is no reason why German tourists cannot now argue that the requirement of having to drive on the left hinders their free movement since British rules are different to those in Germany. Here, the Court's “intuition” might lead it to find that such rules are in fact neutral as regards intra-Community movement and so should not breach Community law.
This might explain why in Burbaud Footnote 136 the Court said that the requirement of passing an exam in order to take up a post in the public service could not “in itself be regarded as an obstacle” to free movement.Footnote 137 Similarly, in Deliège Footnote 138 the Court said that, although selection rules for an international sports competition inevitably had the effect of limiting the number of participants in a tournament, such a limitation was inherent in the conduct of an international high-level sports event. Such rules therefore could not in themselves be regarded as constituting a restriction on the freedom to provide services prohibited by Article 49.Footnote 139 In Graf Footnote 140 the Court went further and equipped itself with a more substantive tool for rejecting cases whose effect on free movement it intuitively felt was insufficient. The case concerned an argument run by a former employee that, by resigning from a job, he lost his chance of claiming compensation when the employer dismissed him, and this breached Article 39. However, the Court said that the loss of a chance of being dismissed was “too uncertain and indirect a possibility for legislation to be capable of being regarded as liable to hinder free movement for workers”Footnote 141 and so there was no breach of Article 39.
These cases are, however, the exceptions, not the rule, and generally the Court, almost unthinkingly, applies the market access approach to most rules which come before it.Footnote 142 In more recent cases the Court tends to favour the language of “restrictions” or “obstacles”Footnote 143 rather than “market access”,Footnote 144 especially in cases which are less obviously commercial,Footnote 145 but the terms are synonymous, a point the Court emphasised in Commission v. Italy (motor insurance):Footnote 146 “the concept of restriction covers measures taken by a Member State which, although applicable without distinction, affect access to the market for undertakings from other Member States and thereby hinder intra-Community trade”. In reaching this conclusion, the Court cited both Alpine Investments Footnote 147 and Caixa-Bank. Footnote 148 However, it was in Caixa-Bank that Advocate General Tizzano emphasised the risks associated with the Court's current pre-occupation with the “restrictions” approach. He said that “rather broad concept of restriction” in the case law covering all “national measures liable to hinder or make less attractive the exercise of fundamental freedoms guaranteed by the Treaty”Footnote 149 allows economic operators, both national and foreign, to abuse the Treaty to oppose any national measure that, solely because it regulated the conditions for pursuing an economic activity, could in the final analysis narrow profit margins and hence reduce the attractiveness of pursuing that particular economic activity.Footnote 150
The risk of abuse lies in the fact that, in contrast to competition cases, there is a striking absence of analysis in the free movement cases of what that market is, the extent to which it has been screened off and whether there are other ways of gaining access to that market.Footnote 151 Rather, in most cases the Court contents itself with assertions, such as the one in Bosman, that the transfer fee rules, although non-discriminatory, still “directly affect players'” access to the employment market in other Member States and are thus capable of impeding freedom of movement for workers”.Footnote 152 Moreover, unlike competition cases, the market access test is not apparently subject to any de minimis or remoteness threshold,Footnote 153 despite the best efforts of certain Advocates Generals to introduce such a requirement.Footnote 154 This means that almost every national rule can be challenged because almost every rule has some effect on inter-state trade, even if that was never the intention of the rule and even though the effect on inter-state trade is very slight.Footnote 155
(b) Justifications and Proportionality
An expansive reading of the Treaty provisions also means that the Court must then evaluate policy choices made by Member States in an ever-wider range of situations.Footnote 156 In the early goods cases, especially those concerning food, the ECJ confidently assessed the justifications put forward by the Member States – normally consumer protection and public health – and their proportionality. In these cases, Maduro argues, the Court was guided by the principle of majoritarianism: if a majority of Member States in the EU did not protect consumers in this way then the national law would not be upheld.Footnote 157 By relying on this majoritarianism approach, the Court avoided making significant policy judgments about national laws.
By contrast, more recent cases involving establishment and services have forced the Court to make complex policy assessments, often without being fully aware of the implications of the assessment it is making. Take the case of Viking as an example. There the Court found that the trade union's proposed strike action was a restriction on Viking's freedom of establishment. In their defence, the trade unions argued that the strike was justified in the name of worker protection, a position that the Court in principle accepted.Footnote 158 However, it then went on to curtail the circumstances in which that argument could be raised:
[E]ven if that [industrial] action—aimed at protecting the jobs and conditions of employment of the members of that union liable to be adversely affected by the reflagging of the Rosella—could reasonably be considered to fall, at first sight, within the objective of protecting workers, such a view would no longer be tenable if it were established that the jobs or conditions of employment at issue were not jeopardised or under serious threat.Footnote 159
It also gave a narrow reading to the proportionality principle. When considering the necessity limb of the proportionality test, the Court said that industrial action should be the last resort; and that national courts would have to verify whether the union had exhausted all other avenues under Finnish law before the industrial action was found proportionate.
In reaching this conclusion the Court, apparently unwittingly, trampled over a range of social values. For example, it ignored the fact that the strike action was protected by the Finnish constitution; it seemed unaware that the criteria it prescribed for taking strike action (jobs and conditions of employment are under serious threat) limits the possibility of taking industrial action even in the UK where national laws are already strict; it certainly rules out political strikes which are protected by various national constitutions. And in its identification of the criteria for assessing the proportionality of a strike, criteria which national courts will have to apply in this politically sensitive domain, the ECJ has created the paradox that the more successful industrial action the less likely it is to be proportionate.Footnote 160 Thus, in its prioritisation of freedom of establishment, the Court failed to see the impact of its decisions on a wide variety of industrial relations systems.
V. Non-Discriminatory Restrictions: Lessons From the US?
A. Introduction
The argument so far is that, by using the restrictions rather than the discrimination analysis, the ECJ has chosen to go down the high impact route, catching a range of discriminatory and non-discriminatory measures in the Treaty prohibition and then allowing defendants (often, but not always, Member States) to justify their rules and show that the national measures are proportionate. Often the defendants fail to defend their laws because the burden of proof is so high. While this approach may be justifiable in respect of discriminatory measures, this is not necessarily the case with some (but not all) non-discriminatory rules which merely structure the market on which trade is taking place, notably employment, tax and environmental laws. As we saw in section III, in analogous situations in the US such rules have benefited from a strong presumption of legality, an approach the Court of Justice experimented with in Keck. Footnote 161
The ECJ is also showing signs of experimenting with a return to a discrimination approach in the field of tax, an area where state treasuries had expressed grave concern about the implications of the Court's “restrictions” analysis.Footnote 162 This shift was signalled in Weigel,Footnote 163 a case concerning the levy of a hefty tax when registering a car in Austria. The Court rejected the arguments raised by a German husband and wife, who transferred their residence to Austria where Mr Weigel had a job, that because this tax discouraged them from moving it breached Article 39(2). The ECJ noted that the Austrian rule applied “without regard to the nationality of the worker concerned to all those who registered a car in Austria and, accordingly, it is applicable without distinction”.Footnote 164 The Court then added that “It is true that it is likely to have a negative bearing on the decision of migrant workers to exercise their right to freedom of movement”.Footnote 165 Under the Säger formula, this observation would have been enough for a finding of a breach of Article 39 which the Member State would then have to justify.Footnote 166 However, in Weigel the Court said that, provided that nationals and migrants were treated in the same way, the fact that the tax regimes of the different Member States were different was not enough to trigger the application of the Treaty: there needed to be something more for the Treaty to apply.Footnote 167
Mobistar Footnote 168 confirmed that Weigel was not a one-off. It concerned a request for the annulment of taxes adopted by the Community of Fleron in Belgium on transmission pylons, masts and antennae for mobile phones on the grounds, inter alia, that the taxes contravened Article 49. Having cited the Säger market access test, the Court said:
By contrast, measures, the only effect of which is to create additional costs in respect of the service in question and which affect in the same way the provision of services between Member States and that within one Member State, do not fall within the scope of Article [49] of the Treaty.
These cases do suggest a move towards a more discrimination based approach. However, as Banks shows,Footnote 169 the Court is far from consistent: a year after Mobistar, the Court made no mention of the discrimination analysis in other tax cases such as Commission v. Belgium Footnote 170 and Commission v. Finland. Footnote 171 Yet, she argues subsequent Grand Chamber decisions, notably ACT Group Litigation Footnote 172 and FII Litigation,Footnote 173 “indicate a definite resurgence in discrimination reasoning in tax cases”. If this is the general trend, then the Court appears to be taking an important step towards preserving perfectly neutral taxation systems from scrutiny under community law and thereby respecting Member States competence.
If the Court can change its mind about tax, I would like to suggest that the time has come for it to reconsider its approach in other areas, like social policy. One possibility would be for it to reintroduce a pure non-discrimination approach, as it has done in respect of tax. However, this would expose the ECJ to the criticism outlined in section III above of the weaknesses of a pure discrimination approach in the evolving context of a single market. Another possibility, and my favoured approach, is to follow the lead of the US Supreme Court to say that other areas of non-discriminatory national regulation should benefit from the presumption of legality (discriminatory legislation would still remain subject to the full force of the restrictions approach). It is however a presumption only and one that can be rebutted by claimants showing that there is discrimination in fact or that they are substantially hindered from gaining access to the market in other Member States. I begin by making the case as to why the time is ripe for change and then to outline in more detail how the Court might apply the US approach to the EU context.
B. The Maturing of the EU's Single Market
It was emphasised earlier in this article that the US system is considerably more mature than the EU's. However, I would suggest that the due to the use of some innovative legislative techniques, the EU is maturing more rapidly than its relative youth might indicate. Two legislative tools have helped in this process. The first is the Technical Standards Directive 98/34,Footnote 174 given renewed vigour as a result of the decisions in CIA,Footnote 175Unilever,Footnote 176 and Sapod-Audic. Footnote 177 It requires states to notify draft technical standards to the Commission which must then screen their compatibility with Community law.Footnote 178 The idea behind this approach is to ensure that fewer restrictive measures get on to the rule books of the Member States which then have to be challenged subsequently in ad hoc litigation. Since CIA, the number of notifications has risen significantly,Footnote 179 giving the Commission and the Member States far greater opportunity to scrutinise proposed national rules.
The second legal tool is the Services Directive 2006/123.Footnote 180 The establishment Chapter of the Directive contains a watered down version of the notification obligations under the Technical Standards Directive 98/34 in respect of new measuresFootnote 181 and an even more dilute version in respect of services.Footnote 182 However, in respect of pre-existing measures, the Services Directive requires Member States to undertake an ambitious programme of screening of their legislation.Footnote 183 This legislative spring-clean, if done conscientiously by the Member States, should ensure the removal of a plethora of unjustified national rules by 28 December 2009.Footnote 184 Furthermore, the mutual evaluation process under Article 39 of the Directive should give Member States the opportunity to scrutinise any remaining national rules for their compatibility with Community law. If this legislative spring-clean has been done in good faith, all that should remain on the national rule books are (1) measures which can be justified and (2) non-discriminatory national rules which are too remote from freedom of establishment and free movement of services to be challenged. This second category reflects the rather obscure second sentence of the 9th Recital of the Services Directive which says that the Directive does not apply “to requirements, such as road traffic rules, rules concerning the development or use of land, town and country planning, building standards as well as administrative penalties imposed for non-compliance with such rules which do not specifically regulate or specifically affect the service activity but have to be respected by providers in the course of carrying out their economic activity in the same way as by individuals acting in their private capacity.” If only justified national rules and national rules too remote from free movement are left on national statute books, there is an argument that the Court should adopt a more hands-off approach to scrutinising national legislation under Articles 43 and 49, at least in those areas covered by the Services Directive.
C. Why Can't the EU be More Like the US?
It may be naïve to think that these legislative techniques will be as successful as their proponents advocate and so there is still a role for the Court of Justice to play through its application of the Treaty provisions on the four freedoms. However, I would argue that these changes provide the pretext for the Court to back away from rigorously applying its restrictions approach in cases where the national rule is non-discriminatory. I would argue that, as in the US Supreme Court's decision in Pike, such rules are presumptively lawful but leave open the possibility for market participants (and, in the case of Article 226 proceedings, the Commission) to rebut that presumption by showing that the rule nevertheless substantially impedes their market access. The burden of proof for claimants would be a stiff one: their case would have to be more than simply that the host state rules are different (and they would prefer to have their home state rules applied instead) and make life more expensive for the claimant than if they had not migrated.Footnote 185 Claimants should at least have to show that the host state rules “partition the market”. The test for this might be that the national rules directly and substantially hinder market access: i.e. the national rules directly affects the claimants' ability to engage in their specific economic activity in the host state. So a challenge to a national speed limit would not satisfy this test because it does not just apply to, say, speeding opticians but to everyone in the state acting in a private as well as an economic capacity. (I think this is what the phrase “providers in the course of carrying out their economic activity in the same way as by individuals acting in their private capacity” in the 9th Recital of the Service Directive was trying to achieve). By contrast, a limit on the number of opticians' shops an entrepreneur could open, using the facts in Commission v. Greece, or even possibly a restriction on who can own and operate a pharmacy, as in Commission v. Italy Footnote 186 does go to the heart of the entrepreneur's business model and so might satisfy the “directly and substantially hinder market access” test.Footnote 187
Only if the claimants satisfy this test will the burden shift to the defendant state to justify the national rule; and even here there should be a more nuanced approach to justification based on a true balance of competing interests. Once again Pike might provide the necessary model: the starting point of the balance should not be the illegality of the national rule. Rather, the Court should recognise that the rule restricting opticians from running more than one shop in order to ensure the quality of care for each customer is on an equal footing with the need for entrepreneurs to have access to the Greek market. Only if the justification is a veneer or the restrictive effects of the measure are excessive would the measure contravene the Treaty.
I wonder whether it was this type of balancing that the Court, somewhat clumsily, tried to engage in, in the much criticised decision in the Sunday trading case Torfaen. Footnote 188 The Court said that “Article [28] of the Treaty must be interpreted as meaning that the prohibition which it lays down does not apply to national rules prohibiting retailers from opening their premises on Sunday where the restrictive effects on Community trade which may result therefrom do not exceed the effects intrinsic to rules of that kind.”Footnote 189 While, in recent years, there has been no appetite for this genuine balancing approach (at least in the context of Sunday trading where it has now been replaced by Keck Footnote 190) it may still have a role to play in the more difficult cases – like Viking and Laval – where intuitively we might sense that there is an impact on interstate trade but there is a need to engage in a genuine balance of interests of equal value.Footnote 191 This exercise is by no means easy, which is why it should be undertaken only rarely, but at least it would give real meaning to the idea, expressed in Article 3 TEU-L, that the Union should work to establish a “social market” economy,Footnote 192 a term which underlines the link between economic and social policy to ensure greater coherence between the two.
VI. Conclusions
In this article I have argued that the time has come for a reconsideration of the blanket application of the “market access” or “restrictions” approach in respect of non-discriminatory measures. This approach threatens to undermine legitimate national regulation and, as Viking and Laval show, recourse to the justification/proportionality formula is insufficiently sensitive to local needs. I have therefore argued that, following the approach of the US Supreme Court in Pike under the Dormant Commerce Clause, such measures should benefit from a strong presumption of legality. Only if claimants can demonstrate that their access to the market has, nevertheless, been directly and substantially infringed (as they might do if the facts of Viking and Laval were repeated today) will the Treaty be engaged and thereafter a genuine balancing of the competing interests should occur. The consequence of the proposed approach is that it recognises that there remain benefits, even in a single market, to maintaining a market with some national rules, particularly those which serve a social purpose. As Article 2 EC makes clear, the attainment of the four freedoms is not an end in itself but a means of achieving “the raising of the standard of living and quality of life, and economic and social cohesion and solidarity among Member States.” If the interpretation of the four freedoms leads to a reduction in the standard of living, due to the deregulatory effect of a blanket application of the restrictions approach, this is clearly not a result that the Treaty drafters envisaged. It is also an approach which also risks alienating EU citizens.Footnote 193
Any change in position by the ECJ inevitably alters the balance of power between the centre and the states. The proposal put forward here is not intended to replace the EU's “high impact” approach with the “low impact” approach found in the US. Instead, it identifies problems with the wholesale application of the restrictions approach and proposes a way in which the model could be finessed by taking one aspect of the US approach and applying it, in the EU, to non-discriminatory national rules. At first sight, it might seem odd to draw on the US model to save the European social model from challenges under the restrictions approach, given that the legitimacy of the states in the US is much less closely associated with the achievements of the welfare state than in the EU.Footnote 194 Nevertheless, the balancing test in Pike has prevented the US courts from, in Justice Rehnquist's words in Kassel, arrogating to themselves public policy formation, functions which were left to the states by the Constitution. Much the same can be said of social and tax matters in the EU: these are largely areas for the Member States and only exceptionally should the ECJ use the Treaty provisions on the four freedoms to strike down any such non-discriminatory rules.
