Friday, 28 November 2014

The nine labours of Cameron: Analysis of the plans to change EU free movement law


Steve Peers

As a penance for terrible sins committed in a moment of insanity, the mythical hero Hercules was required to perform ten great labours – extended to twelve after he cheated on two of them. Similarly, David Cameron has today pledged to insist upon major changes to the rules on free movement of EU citizens, failing which he will advocate Britain leaving the European Union. He clearly considers this necessary in light of the political sin of failing to meet his earlier promise to hugely reduce the amount of migration to the UK – which was, in hindsight, made in a moment of political madness.

Cameron’s commitments will not be easy to meet, for the reasons I set out below. I have broken them down into a list of specific commitments. I leave it to the classicists to match each of them to Hercules’ tasks.

If no changes to EU free movement law are made, the Prime Minister implicitly suggested that he would campaign for the UK to leave the EU. He did not state that he would insist upon all of his suggested amendments being agreed. So it is important to assess how realistic each of these proposals is.

A crucial factor in determining what is politically realistic is the nature of the current EU legal rules. Where Cameron’s proposals reflect the status quo, they are obviously realistic. Where they would require EU legislative change, they will require only a qualified majority of EU Member States in favour, plus a proposal from the Commission and agreement of the European Parliament. But where would need an EU Treaty amendment, they would have to be agreed by all Member States and ratified by each national parliament. That isn’t impossible, but it won’t be very easy; and it means that Cameron will more likely have to compromise on the proposals made today.

What about the other political parties? UKIP, of course, wants the UK to leave the EU. The Liberal Democrats generally support benefit reforms, and the Labour party’s policy, as announced recently is broadly consistent with Cameron’s proposals. The difference is that the Liberal Democrat and Labour parties do not (or not yet) support a referendum on continued UK membership of the EU.

As can be seen from the detailed comments below, most of Cameron’s specific proposals will require a Treaty amendment. I have elaborated on some of the points made below in my recent blog post on the legal limits on the amendments to EU free movement law.

Specific proposals

1. No access to tax credits, housing benefits and social housing for four years for EU citizens

For EU citizens who are not workers, work-seekers or former workers, this confirms the status quo, as set out in the CJEU’s recent ‘benefits tourism’ judgment in Dano (see discussion of that judgment here).

For EU citizens who are work-seekers, the free movement of workers in the Treaties (as interpreted by the CJEU) requires Member States to give them access to benefits linked to labour-market participation. These benefits would probably not be covered by that rule. So this confirms the status quo.

For EU citizens who are workers (as defined by the Treaties and CJEU interpretation) or former workers (as defined by EU legislation, and the CJEU interpretation of the Treaties), there is a right to equal treatment as discussed in my prior blog post.  So this change would require a Treaty amendment.

2. Removal if job-seekers do not find a job within six months

For EU job-seekers, the EU legislation states that they cannot be expelled as long as they ‘can provide evidence that they are continuing to seek employment and that they have a genuine chance of being engaged’. This reflects the case law of the CJEU, interpreting the Treaties (Antonissen judgment). Therefore this change would require a Treaty amendment.

3. Ending the entry of non-EU family members without restrictions

Currently EU citizens can bring with them their spouse or partner, the children of both (or either) who are under 21 or dependent, and the dependent parents of either. (As an exception, students can’t bring their parents with them – but it’s rather doubtful that many students would want to do that if they could). This applies regardless of whether the family members are EU citizens or not. In 2003, in the judgment in Akrich, the CJEU ruled that Member States could insist that non-EU family members had previously been lawfully resident in the Member State concerned (previously no such rule appeared to exist). But in 2008, in Metock, the CJEU overturned this ruling and said that a prior legal residence requirement was not allowed.

This rule could possibly be overturned by EU legislative amendment, but it is possible that the CJEU would find that for those EU citizens whose rights are based on the Treaties (workers, job-seekers, former workers, self-employed persons, students), such a restriction was a deterrent to free movement, so a Treaty amendment might be needed.  

In fact, Cameron’s speech appeared to suggest that all non-EU family members of all EU citizens would have to be subject to the same restrictive income and language requirements that apply to UK citizens who seek to marry non-EU citizens. This would clearly deter free movement of those EU citizens who have non-EU family members, and would surely require not just a legislative amendment but a Treaty change These changes would therefore probably restrict the possibility of UK citizens to obtain family reunion by going to another Member State to be with their family and then returning, ie the so-called Surinder Singh route.

4. Tougher and longer re-entry bans for foreign rough sleepers, beggars and fraudsters

This assumes that re-entry bans are possible at the moment for such persons. That’s clearly not the case for rough sleepers and beggars: Article 15(3) of the EU citizens’ Directive states unambiguously that a ban on entry cannot be imposed where a person was expelled for grounds other than public policy, public security and public health; and Article 27(1) states clearly that such grounds ‘cannot be invoked to service economic ends’. This rule could possibly be overturned by EU legislative amendment, but it is possible that the CJEU would find that this would constitute a disproportionate restriction on free movement for those who were entering to obtain work later. So a Treaty amendment might be needed.

For those convicted of fraud, they would arguably have been expelled on grounds of public policy, public security and public health. In that case, entry bans are allowed, but EU law does not set absolute upper limits on the duration of those bans. Instead, Article 32 of the citizens’ Directive merely sets out a right to challenge them. So in principle longer and stricter entry bans for fraudsters are possible, as a matter of national law, under the status quo. However, if Cameron wants to make sure that they cannot be challenged successfully in individual cases as a disproportionate limit on free movement rights, a legislative amendment or Treaty amendment might be necessary.

5. Stronger measures to deport EU criminals

Currently, as well as possible expulsion for not meeting the economic criteria for free movement, EU citizens can be deported if they are a ‘sufficiently serious threat’ to public policy or public security. Expulsion decisions must be made on a case-by-case basis and a criminal conviction cannot automatically lead to expulsion. In principle, there is a higher degree of protection against expulsion for those who have got permanent residence status (due to living in a Member State for more than five years) or who have lived in that Member State for more than ten years. However, the CJEU’s case law has somewhat undercut that extra protection in practice.

Any changes to the rules on deportation after a criminal conviction would require at the very least a legislative amendment. It is likely that they would also require a Treaty amendment, since the protection against removal on grounds of public policy, public security or public health is set out in the Treaties for EU migrant workers (Article 45(3) TFEU).

6. Longer waiting periods for free movement of persons from new Member States

The details of each new Member State’s adaptation to EU law are set out in each accession treaty, which has to be approved by each Member State. The UK would be in a position to veto any future accession treaties unless longer waiting periods for free movement are included. So this is entirely consistent with the status quo.

7. EU citizens to have a job offer before entry

EU legislation says that EU citizens can stay for three months as long as they have a valid passport and do not apply for social assistance. After that point, they can stay if they are workers, self-employed persons, students or otherwise have sufficient funds. Also they can stay if they are a job-seeker, subject to the proviso (described above) that they have a genuine chance of getting work. So this proposal would require a legislative amendment.

In fact, it would also require a Treaty amendment, since the CJEU has said (in Antonissen) that the Treaty right to free movement of workers also applies to job-seekers, giving them the right to enter and stay in a Member State to look for work.

If such an amendment were made, it would clearly be unfeasible in practice, since it is not realistic to check at the border or before getting on transport if all EU citizens have job offers, considering also that there are many other legal reasons for EU citizens to come to the UK (short-term visits, self-employment, for study, or self-sufficient stay).  

8. No taxpayer support for job-seekers

The EU legislation already rules out social assistance for job-seekers, so this reflects the status quo. However, the CJEU has said that job-seekers have a right to access benefits linked to labour market participation, if they have a link already with the labour market in question. While David Cameron suggested that the UK’s future Universal Credit would not fall within the scope of the CJEU’s case law, that would likely be challenged in practice. So a Treaty amendment is likely necessary as regards current rules, and possibly necessary as regards universal credit.

9. Payment of child benefit to children abroad

Non-payment of child benefit to children living in other Member States is arguably indirectly discriminatory, since it affects more non-UK nationals than UK nationals. Since the equal treatment of workers is guaranteed by the Treaties, a Treaty amendment would likely be necessary to put this change into effect as regards workers’ family members.


Barnard & Peers: chapter 13

Wednesday, 26 November 2014

So long, and thanks for all the fish: the CJEU clarifies international law and institutional issues applying to fisheries


Steve Peers

The EU’s fisheries policy is controversial both within the EU and outside it, due to its impact on both fish stocks and the livelihoods of fishing communities. Until the Treaty of Lisbon, its application was essentially the sole preserve of the Council. The European Parliament (EP) tried to obtain joint control of some of the external aspects of the policy by claiming that its consent was needed for international fisheries treaties that impacted the EU budget significantly, but the CJEU rejected this challenge.

Following the Treaty of Lisbon, however, the EP has joint decision-making power internally over agriculture and fisheries policies, since Article 43(2) TFEU states that the ordinary legislative procedure now applies to the adoption of legislation in this field. However, certain aspects are still reserved to the Council, with the European Parliament only consulted, namely ‘measures on fixing prices, levies, aid and quantitative limitations and on the fixing and allocation of fishing opportunities’ (Article 43(3)).

As regards fishing, this provision is used each year just before Christmas, to determine the annual allocation of fish between Member States. It’s safe to say that the tone of these usually bitter negotiations never induces fisheries ministers to kiss each other under the mistletoe.  

What exactly is the dividing line between the areas where the EP shares power with the Council, and where fisheries ministers are left to hold acrimonious discussions among themselves? There are pending cases on the dividing line as regards internal EU measures. But today the CJEU ruled on the division of powers externally, in a case which also raised interesting issues of international law.

The judgment

Today’s judgment concerned a Council Decision which awarded Venezuelan fishermen the possibility to fish in the waters of French Guiana (which is a constituent part of French territory). This confirmed what those fishermen had been doing for some time before. The EU had felt it was necessary to put this practice on a more formal footing, but the rather left-wing Venezuelan government did not want to sign a treaty with such dastardly agents of global capitalism.

So the EU adopted an apparently unilateral Decision on this issue. The Council believed that it fell within the scope of Article 43(3), so the EP only had to be consulted, while the EP and the Commission argued that it fell within the scope of Article 43(2), so that the EP had the power of consent. The rules on the EP’s role in the approval of international treaties to which the EU becomes a party are set out in Article 218 TFEU. Basically the EP has the power of consent whenever a treaty concerns an issue regulated by the ordinary legislative procedure (ie fisheries law generally), but is only consulted when a treaty falls within the scope of other decision-making rules internally (ie the non-legislative procedure that applies when the EU fixes and allocates fishing opportunities). (In fact, the rules on the EP’s role in approving international treaties are slightly more complicated, but only this basic distinction is relevant to today’s judgment).

But was the Council Decision an international agreement in the first place? The Advocate-General’s opinion argued that it was not. Rather, it was a unilaterally binding declaration, an interesting form of international law. In fact such legal creatures are so rare that international law had not yet clarified whether international organisations like the EU could adopt them. In the Advocate-General’s view, they could. But that left the awkward question of how exactly the EU could adopt one as a matter of its internal law, since Article 218 clearly only refers to agreements concluded by the EU (or on behalf of the EU by its Member States). She considered various options, but ultimately argued that the relevant provisions of Article 218, including the powers of the EP to give its consent and receive information on negotiations, as well as the special jurisdiction of the CJEU, applied by analogy.

In the Court’s view, however, the Decision did constitute an international agreement. It based itself on the relevant rules of the UN Convention on the law of the sea, to which the EU and its Member States (but not Venezuela) are parties, and confirmed its position in the recent ruling on the Hague Convention on child abduction (discussed here), that treaties could be concluded in two steps. In this case, the EU had extended an offer, which Venezuela was free to accept, reject or suggest changes to. It had chosen to accept.  

As for the internal division of powers, both the Advocate-General and the Court reached the same conclusion: the Parliament’s argument was correct. In the Court’s view, the main powers relating to agriculture and fisheries set out in Article 43(2) concerned ‘policy decision[s] that must be reserved to the EU legislature’. In contrast, Article 43(3) provided for ‘measures of a primarily technical nature’ to implement the legislation adopted in the field. Applying that distinction to this case, the ‘treaty’ with Venezuela set out only a general framework, which had then been implemented by further measures based on Article 43(3). So that ‘treaty’ could not itself be based on Article 43(3); rather it was subject to the consent of the European Parliament, as it was based on Article 43(2).


On the international law issue, the Advocate-General’s analysis is more convincing than the Court’s, given the clear unwillingness of Venezuela to engage in any formal negotiations along its failure to ratify the UN Convention on the law of the sea, which the CJEU relied on so heavily. Nor is the Court’s use of the language of contract law very convincing. True, Venezuela’s application for fisheries authorisations might be described as the acceptance of an offer, but what is the consideration? Why should Venezuela’s actions be characterised in light of a treaty it had not ratified? Possibly the relevant rules reflect customary international law on the law of the sea which apply to Venezuela, but the Court does not make that argument.

Nor is its analysis of the text of the Law of the Sea Convention very convincing. The relevant clause refers to making ‘agreements or other arrangements’ regarding surplus fish. Could not a unilateral binding declaration constitute a form of ‘other arrangement’? Possibly that interpretation has been rejected by the Law of the Sea tribunal or by experts in that field of law (I confess that I’m not one), in accordance with the sources of international law as defined in the UN Charter. But if that is the case, the Court needs to bolster its interpretation by citing such evidence.

The distinction between the forms of international obligation matters mainly as regards the EU’s internal law. If the Decision was a unilateral binding declaration, the CJEU would have the awkward job of deciding whether the EU can adopt such measures, and if so how. Since the Court didn’t have to address these issues today, they must be considered open. But if it is every necessary to consider them in future, there is much to recommend the Advocate-General’s very thorough analysis of both of these points.

As for the internal decision-making rules, the judgment is more convincing, particularly in light of the Advocate-General’s arguments that Article 43(3) cannot apply to everything concerning fishing opportunities, since that would render the main legislative powers set out in Article 43(2) superfluous.

The broader implications of this judgment remain to be seen. But it’s an early indication that the Court is inclined to tilt in favour of a broad interpretation of the scope of the EP’s legislative and treaty approval powers over agriculture and fisheries following the entry into force of the Treaty of Lisbon.

Barnard & Peers: chapter 5

Tuesday, 25 November 2014

The Domino Effect: how many EU treaties violate the rights to privacy and data protection?


Steve Peers

Earlier this year, the Court of Justice of the European Union (CJEU) ruled in the Digital Rights judgment against the validity of the EU’s data retention directive, on the grounds that it provided for mass surveillance without any effective safeguards. Subsequently it ruled against Google, in what has become known as the ‘right to be forgotten’ judgment.

What are the longer-term consequences of the Court’s ‘Privacy Spring’? An Irish court has already referred the ‘Europe v Facebook’ case (discussed here) to the CJEU, asking in effect whether the EU’s ‘Safe Harbour’ arrangement on data protection with the USA is compatible with the rights to privacy and data protection, in light of the Snowden revelations. Now the European Parliament (EP) has decided to refer the proposed EU/Canada agreement on passenger name record (PNR) data to the CJEU, asking if it is compatible with the rights to privacy and data protection in light of the Court’s recent case law. That judgment would implicitly determine whether the separate EU/USA and EU/Australia treaties on PNR data, and the proposed PNR Directive, violate those rights also. And if the PNR treaties breach the rights to privacy and data protection, it would then be more likely that the EU/USA treaty on banking data transfers also breaches those rights in turn.

So, are we at the start of a ‘domino effect’ of a series of EU laws and treaties being ruled in breach of the rights to privacy and data protection by the Court of Justice, all falling in sequence now that the data retention Directive has been overturned? Or are the features of the different measures different enough to avoid this?  


There’s a little bit of déjà vu in today’s decision by the EP to ask the CJEU about the EU/Canada treaty on PNR. Back in 2004, it asked the Court to rule on the original EU/USA treaty on the same subject. The Advocate-General’s opinion in that case ruled against all of the EU’s arguments, including the right to privacy point. However, the Court’s 2006 judgment only ruled on one of the EP’s legal arguments – that the EU/USA treaty had the wrong ‘legal base’, and should have been approved by using a different procedure (relating to police cooperation, instead of the internal market). And that procedure meant that the EP had no role in the approval of the treaty, or any power to ask the Court of Justice about its compatibility with EU law.

Eight years later, the legal environment is quite different. Since the Treaty of Lisbon entered into force in 2009, the EP (or the Commission, Council or a Member State) can ask the CJEU for rulings on the compatibility with EU law of EU treaties with third States on police or criminal law cooperation. Indeed, this will be the first such ruling. And while waiting for the Court’s ruling, the EP can prevent the EU/Canada treaty from being concluded, since it now has the power of consent over such treaties (back in 2004, the Council circumvented a separate request by the EP for the CJEU to rule on the EU/USA PNR treaty by concluding that treaty without waiting for the Court’s opinion). Furthermore, the substantive legal environment has obviously been transformed by the Court’s ruling against mass surveillance earlier this year.

The CJEU had another chance to rule on the right to privacy in the international context when the Commission asked it to rule back in 2012 whether the international Anti-Counterfeiting Agreement (ACTA) violated EU law. However, the Commission left it too late to send its request to the Court, and the EP simply vetoed that proposed agreement before the Court could rule (the Commission then withdrew its case). So we should now get a long-awaited ruling from the Court on the compatibility of international data transfers with the EU rights to privacy and data protection – unless the EP can be talked into withdrawing its request to the Court.

The procedure which the EP has invoked today is a special process which allows the Court to rule on the compatibility with EU law of a draft treaty to be concluded by the EU (or by its Member States on behalf of the EU), before that treaty comes into force. (For Canadian readers: this process is broadly similar to sending a request to the Supreme Court to rule on the constitutionality of a draft law. The EU process only applies to treaties, though.) If the CJEU rules (probably in about 18 months’ time, unless the ruling is expedited) that the draft treaty is incompatible with EU law, either the draft treaty has to be amended to comply with the Court’s ruling, or (improbably) the EU Treaties themselves have to be amended to permit its ratification.

The EU/Canada PNR treaty is distinct from the EU/Canada treaty liberalising air transport (already in force), and the proposed EU/Canada free trade agreement (CETA) – although the latter treaty, along with the EU/USA free trade agreement now being negotiated, will be indirectly impacted by a pending case in which the EU Commission has asked the CJEU to rule on whether the EU/Singapore free trade agreement is compatible with EU law.


So does the EU/Canada PNR treaty violate the right to privacy? There’s a detailed analysis of the broader impact of the data retention judgment on other EU measures in a study by Boehm and Cole, published earlier this year. So this is only a short summary of the issues discussed further in that study. The starting point is how to interpret that judgment: does it rule out all mass surveillance, or just in cases where there are insufficient safeguards? In my view, it does indeed rule out all mass surveillance where it’s linked to EU law, and any draft treaty to which the EU is party would obviously be linked to EU law.

But there’s a prior question: when does a treaty with another State entail mass surveillance? The data retention case concerned collection of data on all phone and Internet use in the EU. This could be compared to the use of social media (in the pending Facebook case), or to international banking transfers, but it’s harder to argue that collection of data on all flights to a particular third country constitutes, by itself, mass surveillance. Having said that, the proposed PNR Directive, which would apply to all flights within the EU, would probably meet the criteria.

If (contrary to my interpretation) the Digital Rights judgment does permit mass surveillance, as long as there are sufficient safeguards, then what must these safeguards be? According to the judgment, there have to be: definitions of the ‘serious crimes’ or other purposes of the data exchange; rules on the subsequent access to the data; limits on the number of people who can access that data; independent control by a court or supervisory authority; strong rules on the data protection period; provisions on protecting data from unlawful access and use; and a requirement to retain the data within the EU only. Obviously, in the context of treaties with non-EU States, the latter requirement must be understood as an obligation to retain the data in the EU or that particular third country.

Do the EU’s treaties with third States meet these criteria? This has to be assessed on a case-by-case basis. At first sight, for example, the EU/Canada PNR treaty contains provisions addressing all of these safeguards issues except one: the transfer of PNR data to other countries, which is permitted (although subject to conditions). But it might be argued that in practice, the right to privacy and data protection is not protected as strongly under such treaties as it might first appear, due to inadequacies in national legislation or practice, such as NSA access to Facebook data or limitations on non-USA citizens claiming privacy rights in the courts.

Finally, there’s an important practical question here. Let’s imagine that the CJEU rules that the proposed EU/Canada treaty violates privacy and data protection rights; or that it approves that treaty, but its reasoning in that judgment casts doubt on the compatibility of other EU treaties with those rights. How can those other treaties be challenged, now that they are already in force?

Time has run out to bring annulment actions against those treaties, or to ask the CJEU for an advance ruling on their compatibility with EU law. But it is still possible for individuals to challenge the application of those treaties via the national courts (as in the Digital Rights and Facebook cases). Or the EP could argue that in order to secure effective protection of rights under the EU Charter of Fundamental Rights, the other EU institutions must take steps to denounce the treaties concerned. If they don’t do so, the EP can sue them for ‘failure to act’ as set out in the EU Treaties.

Barnard & Peers: chapter 9

Monday, 24 November 2014

Amending EU free movement law: What are the legal limits?

Steve Peers

Much public debate in recent weeks has centred upon the possible amendment of EU rules on the free movement of people. In particular, the think-tank Open Europe and the UK Labour party have set out policies on this issue, and Prime Minister David Cameron is expected to make an announcement of his views soon. While most of the discussion focuses on the political and economic aspects of these proposals, there is also an important legal dimension to the debate, because EU law determines how easy (or difficult) it would be to put any potential changes into effect.

In particular, there are three principal ways to change the EU free movement rules: (a) by changing national law, while still remaining consistent with EU law; (b) by changing EU legislation; or (c) by amending the EU Treaties. The first course of action needs only (at most) a national parliamentary majority; the second course of action needs a Commission proposal and support from the European Parliament and a qualified majority in the Council; but the third route needs unanimous support from all 28 Member States’ governments, then ratification in national parliaments.

So it’s important to know which of these categories the proposed reforms fall into. For the reasons set out in this blog post, some of the proposals of Open Europe fall into the second category (EU legislative amendment). But contrary to their arguments, the most significant proposals made by Open Europe fall into the third category (Treaty amendment), making them much more difficult to accomplish than their authors suggest.

As for the Labour party proposals, they broadly reflect the Open Europe suggestion to impose a waiting period on access to in-work benefits for EU migrants. Finally, it’s obviously not possible to comment on David Cameron’s proposals for EU policy reform (which might go beyond EU free movement law) until he makes them. So this blog will return to the issue after his forthcoming speech.

This blog post is in three parts: an analysis of the proposals on cutting back benefits for workers, job-seekers and former workers;  an examination of the other proposals made by Open Europe; and my suggested text for proposed amendments to EU legislation to implement those Open Europe proposals which are (in my view) both desirable and legally feasible.  

Cutting back workers’ benefits

The central proposal of Open Europe today is that newly-arrived EU citizens should be denied equal treatment as regards out-of-work or in-work non-contributory benefits, social housing and apprenticeship schemes for a period of between one and five years, to be negotiated. This suggestion builds upon another recent Open Europe paper which suggests further details of changes to EU legislation on these issues. My comments are mainly directed at the latter paper, which sets out the legal reasoning underpinning today’s proposal. As noted already, the Labour party proposals reflect Open Europe's suggestions, and so the following comments equally apply to those proposals.

I won’t mince words: there’s a fundamental legal problem at the core of these proposals. The Open Europe paper suggests that both in-work and out-of-work benefits could be limited simply by amending EU legislation. This is true for those EU citizens who are not workers, and who have never looked for work or held work in the host Member State, as the CJEU has recently clarified in the ‘benefit tourism’ case of Dano (see discussion here). But look at the Court’s reasoning: it deferred to the wording of EU legislation in that case because the Treaty rules on the rights of EU citizens (Articles 20 and 21 TFEU) defer to the limitations on EU citizens’ rights ‘defined by the Treaties and the measures adopted thereunder’, and ‘laid down in the Treaties and in the measures adopted to give them effect’. Equally the Court referred to Article 18 TFEU, which provides for non-discrimination against EU citizens ‘[w]ithin the scope of application of the Treaties, without prejudice to any special provisions contained therein’.

For people like Ms. Dano, there are no Treaty rules which govern their legal situation, because she was not a worker, and was not a job-seeker or former worker either. Neither was she a self-employed person, service provider or student. Therefore the Court only looked at her position under EU free movement legislation.

But the vast majority of EU citizens in another Member State do fall within one of those other categories – most notably as workers, including job-seekers and former workers. The crucial point here is that all of those other categories of EU citizens have rights not just pursuant to EU legislation, but also pursuant to the Treaties. Indeed, that point is recognised in the very wording of Articles 18, 20 and 21 TFEU: Article 18 defers to the special rules on non-discrimination in the rest of the Treaty, while Articles 20 and 21 defer equally to the Treaties and secondary legislation. The intention of the Treaty drafters is perfectly clear.

So what does this mean for the position of workers? (I’ll concentrate on them, since they are the main focus of the Open Europe proposals). The specific Treaty provision regulating workers’ free movement is Article 45 TFEU, which first specifies that ‘Freedom of movement for workers shall be secured within the Union’. Article 45(2) further specifies that free movement entails the abolition of discrimination between EU workers as regards ‘employment, remuneration and other conditions of work and employment’. Article 46 TFEU then confers powers to adopt legislation ‘setting out the measures required to bring about freedom of movement for workers’. Article 48 TFEU sets out powers to adopt rules to coordinate social security for workers.

The wording of these Treaty articles doesn’t defer to the legislation in order to set out workers’ rights, and this has long been recognised by the Court of Justice. So free movement of workers (and the associated right to equal treatment) has long been recognised in CJEU jurisprudence as a directly effective Treaty right, precisely because it is not dependent on the adoption of further measures. This also means that the CJEU controls the definition of ‘worker’, and extends it to cover also those who are former workers or job-seekers.

Indeed, as recently as this summer, in the Saint-Prix case (discussed here), the CJEU expressly asserted that the definition of former workers (and therefore the access to benefits) as defined in EU legislation didn’t matter, since the Court would determine which former workers still qualified for access to benefits. So the Court ruled that female workers who were former workers at the time when they gave birth still had access to benefits (as long as they got work soon afterward), even though the legislation did not define them as former workers.

So unless the CJEU does a U-turn on this issue, there would be little point in trying to amend the legislation in order to require former workers to wait longer for benefits. At the moment, EU citizens retain worker status, and access to benefits, if they become involuntarily unemployed after working for more than one year in a Member State. If they become unemployed before that point, they retain worker status for a maximum of six months. There are also several other cases where they retain that status.

What about job-seekers who were not formerly employed in the host State? Well, there’s no need to amend EU legislation in order to refuse them some types of benefits, since the EU citizens’ Directive expressly makes clear that they are entitled to no equal treatment as regards ‘social assistance’ at all. However, the CJEU has said (in the Collins case) that, pursuant to the Treaties, job-seekers cannot be refused equal treatment as regards benefits which are intended to facilitate access to employment. And in the Vatsouras case, it expressly distinguished this type of benefit from ‘social assistance’ benefits which job-seekers are not entitled to pursuant to the citizens’ Directive.

The CJEU will have a chance to clarify the position in the pending case of Alimanovic. It might possibly do a U-turn on this issue in its judgment on that case. If it doesn’t, then again, in order to cut back on such labour-market benefits for job-seekers, a Treaty amendment would be necessary.

Next, as regards workers, the key free movement rule, and the equal treatment right, are laid down in the Treaty itself. When ruling on workers’ equal treatment rights, the Court regularly refers to the Treaty rule and the relevant legislation, and both of these confer equal treatment as regards benefits: see the judgment in O’Flynn, for instance. Since the right to workers’ equal treatment is expressly set out in the Treaties, then removing in-work benefits for workers – the core of the Open Europe proposal – would be manifestly contrary to the Treaties, and would require a Treaty amendment.

Of course, a Treaty amendment is not impossible; it’s simply much more difficult than a legislative amendment. It would most likely take the form of a Protocol which sets out a derogation from the Treaty rules.

Other Open Europe proposals

The Open Europe paper suggests an amendment to EU legislation to specify that EU law cannot derogate from national powers as regards matters such as social assistance. It’s not possible for EU secondary legislation to limit EU powers as defined in the Treaties, but an amendment to EU secondary legislation could refer to such limits. This is my suggested amendment 1 below.

In the context of a waiting period for benefits, the Open Europe paper correctly mentions that the time period spent in detention in the home State for committing offences does not count for this purpose. This rule is set out in the case law, but it could also be added to EU legislation. This is suggested amendment 2 below.

Next, the Open Europe paper suggests a number of protections for national workers as regards unfair competition from other Member States. It is suggested here that the EU posted workers’ Directive provides such protection, but this isn’t always the case. So I suggest an amendment to that Directive in order to overturn the controversial CJEU judgment in Laval (amendment 3).

The specific suggested amendments relating to unfair recruitment standards, living wages et al can best be addressed by amendments to the specific Regulation on the free movement of workers. My suggestions to this end are set out in amendment 4.  

Finally, the Open Europe paper gets round to suggesting some protection for migrant EU workers in areas other than employment, on the grounds that the proposed three-year waiting period for equal treatment ‘might contribute to a sense of vulnerability on their part’. The report’s authors should win an award for their contribution to the great British art of understatement here. One could equally say that the Titanic’s collision with an iceberg might have contributed to a sense of dropping body temperature on the part of its passengers.

Open Europe's specific suggestion is to adopt EU legislation protecting EU migrant workers against discrimination on grounds of nationality as regards supply of goods and services. This is possibly superfluous in light of the non-discrimination rules in the Treaty. But to be on the safe side, I have suggested amendment 5, to the Regulation on migrant workers.

Finally, a number of specific statements in the Open Europe paper need correction or clarification. On page 3, there are more than two ‘important protections’ (for States) as regards exceptions from equal treatment for benefits. Nor is sickness insurance a requirement for all EU citizens, only those who are not workers or self-employed. The Brey judgment (text at note 6) has been effectively overturned by Dano, which was delivered after this paper was published. Equal treatment applies only to those EU citizens who are legally resident. Former workers are entitled to benefits not just in the cases referred to in note 9. On page 9, the CJEU has yet to rule whether Article 31 of the EU Charter of Fundamental Rights creates justiciable rights; this issue is the subject of the pending Fenoll case. It’s an overstatement to say that national law giving effect to Charter social rights is always protected by EU law, as the CJEU judgments in Laval, Viking Line and Alemo-Herron make clear.

Proposed amendments

Amendment 1

Add a new Article 1(2) to the citizens’ Directive:

This Directive shall not affect Member States’ competence to define national rules on social assistance and social security (including pensions), public health, public education and employment policies, in accordance with the Treaties.

Amendment 2

Add a new sentence to Article 21 of the citizens’ Directive:

Periods spent in prison as the consequence of a conviction for a criminal offence shall not be taken into account.

Amendment 3

Add a new sentence to Article 2(10) of the posted workers’ directive:

In particular, equality of treatment as regards pay shall be assured as regards all collective agreements, whether or not they are covered by the definition set out in paragraph 8.  

Amendment 4

Regulation 492/2011 on the free movement of workers should be amended to add a new Section 1a, ‘Equal treatment of host State workers’, consisting of a new Article 6a:

Member States shall not permit recruitment of or advertising for employment to nationals of other Member States only.

A new sentence should be added to Article 7(1):

The principle of equal treatment in working conditions applies in particular to any rule relating to wages, including a minimum wage or living wage requirement in the host Member State, as well as any rules relating to health and safety.

A new Article 7(5) should be added:

In order to ensure a dignified standard of living for workers exercising free movement rights, Member States may prohibit bonded agreements or tied housing.

Amendment 5

A new Article 7(6) should be added to Regulation 492/2011:

Workers shall enjoy equal treatment without discrimination on grounds of nationality as regards access to and supply of goods and services which are available to the public.

*Note: This would be enforceable in the way that the Open Europe paper suggests, by means of the Directive on enforcement of migrant workers’ rights, adopted in 2014.


Barnard & Peers: chapter 13

Sunday, 23 November 2014

Building the EU Judicial System: Politicians 1, (Judicial) Architects 0


Steve Peers

The Court of Justice has recently proposed that the EU should double the number of judges on the General Court, and abolish the EU’s Civil Service Tribunal. Why did the Court propose such a dramatic change to the EU judicial system? And is it a good idea?  


The development of the EU judicial system has been dominated by a steady increase in the volume of litigation, leading to disputes over how best to manage this workload. Traditionally, the solution has been to increase the number of EU courts. So the Single European Act, in force 1987, amended the Treaties to give the EU the power to establish a Court of First Instance (later renamed the ‘General Court’ by the Treaty of Lisbon) to assist the Court of Justice with its tasks. This power was used to create that Court in 1989; both Courts have always had one judge per Member State.

As the case load of the Court of Justice continued to increase, more and more of that Court’s jurisdiction was transferred to the Court of First Instance, so ultimately that Court had jurisdiction for almost all actions brought against the EU’s institutions, agencies and other bodies. As an exception, the Court of Justice still has jurisdiction over such cases if they are particularly ‘constitutional’ in nature, namely actions brought by the EU’s institutions against each other, and challenges by Member States to the EU institutions’ essentially legislative acts. The Court of Justice hears appeals from the judgment of the Court of First Instance (now the General Court), and retains full jurisdiction over the other types of EU law proceedings, mainly references from national courts for a preliminary ruling and infringement actions against Member States for breach of EU law.

Yet over time, this transfer of jurisdiction to the Court of First Instance overburdened that Court in turn. So the Treaty of Nice, in force 2003, gave the EU power to create a lower tier of EU courts, called ‘judicial panels’ (renamed ‘specialised courts’ by the Treaty of Lisbon). Only one such specialised court has been created: the EU Civil Service Tribunal, which began its work in 2005.  It has seven judges.

But as the EU built more courts, more litigants came. By 2011, the CJEU was again trying to find a way to cope with the increasing volume of litigation. This time, instead of suggesting the creation of a new Court, it proposed that 12 more judges be appointed to the General Court. At the same time, it also proposed other amendments to the Court’s statute, as well as the creation of a class of temporary judges to assist the Civil Service Tribunal.

Member States in the Council, as well as the European Parliament (the ordinary legislative procedure applies to such changes), adopted the proposal on temporary judges for the Tribunal, as well as most of the other proposed changes to the CJEU Statute. But they did not adopt the increase in the number of judges on the General Court. While they agreed with the increase in principle, they could not agree how to share the extra judges among Member States, or how to rotate the appointments between them.

While these discussions dragged on without any conclusion, the backlog of cases at the General Court has continued to rise. And from this autumn, the situation at the Civil Service Tribunal has begun to worsen as well, with the Council unable to fill two vacant slots due to the same arguments about the principle of rotation.

So the Court of Justice has made a new proposal: to double the number of judges in the General Court to two/Member State, while abolishing the Civil Service Tribunal. This would take place in three stages: 12 new judges in the near future, transfer of the seven Tribunal judges in 2016, and appointment of the remaining nine judges from 2019. It remains to be seen whether the European Parliament and the Council will accept this proposal.  


Undoubtedly, some remedy is necessary to deal with the ever-increasing workload of the General Court. The EU has an obligation under Article 47 of the Charter of Fundamental Rights to deal with cases within a reasonable time. While the new judges will cost money, failing to appoint them will cost money too: as the Court’s proposal points out, there are now cases pending seeking damages because the EU judicial system has failed to uphold the obligations imposed by Article 47.

There are two options available to this end (without a Treaty amendment): the creation of more specialised courts, or the addition of more judges to the General Court. (It should be noted that the planned ‘Unified Patent Court’ is not a specialised court, but a form of shared national court; for more details, see last week’s blog post on the developing EU patent system).

The EU’s judges disagree among themselves which of these two options is best. It was an open secret, at the time of the 2011 proposal, that the General Court judges wanted to create more specialist courts, but the Court of Justice’s preference for appointing more General Court judges prevailed. In the 2014 proposal, there is no longer an attempt to hide this disagreement: it’s openly acknowledged that the Court of Justice and Civil Service Tribunal favour the proposal, while the General Court does not.

Why this difference of opinion? One reason is that EU judges look forward to dealing with civil service cases about as much as a cat looks forward to a bath. The judges at the General Court thought they had got rid of these cases forever (except for appeals from the judgments of the Civil Service Tribunal). Indeed, there’s a rumour that on the day that Tribunal was created, every shop in Luxembourg ran out of champagne. But now, to the horror of its judges, these cases seem set to return to the General Court.

On the other hand, the judges of the Civil Service Tribunal, after years of dockets consisting entirely of feuding functionnaires, can look forward to ruling also on bolshy Belarussians and money-grubbing monopolists in the near future. Of course, by the time that the General Court consists of twice as many judges as the Court of Justice, the former will outnumber the latter - although the Court of Justice, not the General Court (or the CJEU as a whole), which has the power to make proposals to amend the rules. (So does the European Commission).

Viewed objectively, which of these solutions is best? The Court’s proposals make a number of good arguments for increasing the number of General Court judges. First of all, there is a good case that the creation of specialist courts is less flexible, since litigation can increase in areas not within their jurisdiction (such as foreign policy sanctions, in recent years). Also, the creation of such courts does not necessarily solve the workload problem, since the cases most suitable for such courts (like trademark disputes) are often easy to deal with, and about a third of such cases will be appealed to the General Court anyway. Moreover, it makes sense to keep such disputes closer to the Court of Justice (via means of appeals from General Court decisions), since that Court has to deal with similar cases in this context of references from national courts on the EU’s trademark Directive, and on the relationship between intellectual property law and EU free movement and competition law.

Furthermore, it takes longer to set up a new specialist court than to appoint more judges to the General Court. While it is possible in theory to deal with the ‘national representation’ issue by ensuring that there is one more judge per Member State across the General Court and a number of specialised courts, the Court of Justice is right to believe that it will be difficult to operate such a system in practice. The Council’s recent illegal refusal to appoint judges to the Civil Service Tribunal has shown that it values pork-barrel politics more than the rule of law.

Finally, the most recent proposal will have the singular advantage of simplifying the EU judicial system, eliminating an entire tier of courts as well as the complex rules applying to screening whether the Court of Justice ought to review a judgment of the General Court on an appeal from a specialist court. This would make the EU’s judicial system easier to understand for the general public – and in particular, for the beleaguered students of EU law.

Barnard & Peers: chapter 10


Thursday, 20 November 2014

Capping bankers' bonuses: a step too far for the EU?

Steve Peers

Bankers are never going to win a popularity contest. The collapse of international financial markets which started in 2008 and has led to austerity across Europe has been widely blamed on lax regulation of banks and irresponsible behaviour by bankers. It has led to a huge overhaul of EU banking regulation, including the transfer of banking supervision to the European Central Bank, new rules on bank bail-outs, and provision for criminal law sanctions against bankers involved in market abuse (discussed here). EU law has gone further still, and adopted rules which cap the amount of bonuses paid to bankers.

The United Kingdom, home to the biggest financial services industry in the EU, has had reservations about some of these new laws. It has opted out of some of them (the market abuse rules, the banking supervision rules and aspects of the bank bail-out rules), and has challenged others in the CJEU. Earlier this year, its challenge to the ban on ‘short-selling’ failed in the Court (see discussion here), and today’s Advocate-General’s opinion suggests that its challenge to the restrictions on bankers’ bonuses should fail too.

These restrictions are found in the EU’s revised rules on capital requirements and the authorisation to take up banking services, which are set out in a parallel Regulation and Directive adopted in 2013. In effect, they require that bankers’ bonuses cannot usually be more than the amount of their ordinary annual salary. By way of exception, the bonuses can be double the amount of the banker’s ordinary annual salary, if bank shareholders agree pursuant to a special procedure.

Advocate-General’s Opinion

The UK raised six main complaints against the bonuses rules: lack of competence by the EU to regulate pay; infringement of the principles of subsidiarity and proportionality; violation of the principle of legal certainty; illegal delegation of power to an EU agency (the European Banking Authority); breach of EU rules on data protection and privacy, due to the potential disclosure of the pay received by bankers; and a breach of the principles of customary international law, due to the extraterritorial effect of the rules. Advocate-General Jaaskinen argues that all five complaints be rejected.

First of all, the Advocate-General argues that Article 53 TFEU (the legal base for this measure) is correct, because that legal base can extend to banking regulation generally, not just the promotion of the freedom of establishment for banks. The pay cap does not constitute a ‘social policy’ measure, since it does not regulate the basic salary paid to bankers, which is the basis for calculating any additional bonus.

Secondly, data protection rules are not violated, because the disclosure of bankers’ pay is only discretionary, not mandatory. In the event that Member States make a request for such disclosure, they would then be bound by EU data protection law.

Thirdly, conferring powers upon the EU agency is not illegal, because the powers do not concern the essential elements of the legislation, and the EU Banking Authority does not adopt the measures itself, but merely recommends their adoption to the Commission.  Fourthly, the principle of legal certainty is not infringed by applying the new rules to pre-existing employment contracts. Fifthly, the principles of proportionality and subsidiarity are not violated, because the creation of a uniform system of risk management was better achieved at EU level, rather than national level, and the EU institutions have great discretion to assess how these principles apply. Finally, the UK has not made out its argument that customary international law rules out the extraterritorial application of such limits.


This case is not about whether limiting bankers’ bonuses is a good idea. Rather it concerns whether it is legal for the EU to limit them. If the EU lacks such power, there would in principle nothing to prevent Member States from limiting bankers’ bonuses individually, if they wished. The argument about whether to do so would then be held at a national level, rather than the EU level.

Some of the UK’s complaints are clearly unconvincing.  As the Advocate-General suggests, the argument about international law is not fully fleshed out or convincing. The legal certainty argument fails to consider that employment law regulation usually impacts upon existing contracts; this is justifiable in light of the public-interest principles underlying the very nature of employment law. Anyway, bonuses are inherently variable. As for the data protection argument, the Opinion largely follows what the CJEU established already in EP v Council (family reunion): if EU law provides for options for Member States, the compatibility of those options with human rights law should be judged when and if Member States exercise those options. In any event, prior case law on data protection and salary disclosure does not set out an absolute ban on release (see Satamedia, for instance).

The UK’s other arguments are rather stronger. While it is true to say that the EU’s banking agency does not actually take the final decision relating to implementation of the bonus cap, it does more than simply provide expert advice on this issue. The Commission must then either act on this advice or do nothing at all: so it does not have full discretion to adopt the delegated acts (see the complex decision-making system set up by the Regulation establishing the Banking Authority). This process is fundamentally questionable because it blurs the accountability for the decision being taken (and moreover, it is too convoluted to be transparent).   

As for proportionality and subsidiarity, certainly the events of the last six years have demonstrably indicated that a more decentralised system of managing banking risks was ineffective. Hopefully the EU-wide measures will be more successful, but in any event the nature of the subject-matter calls for an EU-wide response, in light of the level of integration between European financial markets and the potential cross-border impact of bank failures. But that isn’t the point: the UK is not challenging the entirety of the capital requirements rules, but only some of the handful of provisions which regulate bankers’ bonuses. In fact, it is not challenging those provisions which prevent bankers from receiving bonuses as a consequence of risky behaviour, but only those provisions which regulate bonuses regardless of bankers’ actions. So the opinion should instead have asked whether these provisions meet the requirements of the subsidiarity principle. It is hard to see how they do.

This brings us to the biggest problem with the Opinion: the argument that the legal base on freedom of establishment can regulate bankers’ bonuses. The legal base point here can only be understood by viewing the Treaty as a whole. It has separate provisions on social policy, which include a ban on EU regulation of pay (Article 153 TFEU). The general internal market power (Article 114 TFEU) specifically states that it ‘shall not apply to’ measures ‘relating to the rights and interests of employed persons’. The Treaty drafters’ intention was clearly to provide for lex specialis rules relating to regulation of pay.

The ban on EU regulation of pay has been clarified in the case-law of the CJEU. In the Impact judgment, for instance, it ruled that the EU could not regulate the level or components of pay, but it could establish non-discrimination rules relating to pay as regards categories of workers. Similarly, the working time directive provides for holiday pay, but does not regulate the level or components of pay which a worker normally receives (which then constitute the basis on which the holiday pay is calculated).

Following the logic of these precedents, it is true to say that the capital requirements legislation does not set the level of bankers’ pay, on the basis of which the bonuses are capped. But it does regulate the components of pay, by determining how much of the total amount of pay can be variable. The Advocate-General’s reasoning would mean that the EU would be free to regulate at least some aspects of workers’ pay in any area of law subject to special rules in the Treaty, rather than the general internal market legal base. So the EU could regulate aspects of the pay of farmers, fishermen, transport workers and anyone in other service industries.

It could reasonably be argued that aspects of pay in these other fields can exceptionally be regulated by EU law where that is an essential component of the regulatory framework. This could be the case in banking, for instance if the overall amount of pay could damage the existence of the bank or bonuses were linked to risky behaviour. The legislation does have rules on these issues, but the UK has not challenged them. So it follows that the opinion is fundamentally unconvincing on the legal base point.

In light of the financial crisis, there are many good reasons to regulate banks more effectively, and it would not be shocking if Member States wanted to react to understandable public anger at the huge cost of bank bail-outs by limiting bankers’ income. But resentment at bankers’ pay, even it is entirely justified, cannot authorise the EU to exercise powers which any reasonable interpretation of the Treaties suggests that it just does not have.

Postscript (November 21st): Like any Advocate-General's opinion, this view is non-binding, although a number of British journalists and politicians forgot this when the opinion was released. In any event, the point is moot since, following publication of the opinion, the UK's Chancellor decided to drop the legal challenge. His official reason was to save taxpayers' money, but this is not convincing since a large majority of the legal fees will surely already have been incurred, and there is still a chance to get them reimbursed if the UK wins the case. A victory for the UK would have not have been improbable, given that the CJEU did not follow this Advocate-General's views in the last major banking law case (concerning the ban on short-selling), and that the analysis of the legal basis point is not very convincing. 

Barnard & Peers: chapter 14, chapter 19