Hostname: page-component-745bb68f8f-grxwn Total loading time: 0 Render date: 2025-02-11T14:50:01.453Z Has data issue: false hasContentIssue false
  • English
  • Français

The Necessary Complexity of Consent: Rules and Norms in EU Treaty Making

Published online by Cambridge University Press:  14 January 2020

Dermot HODSON  and
Imelda MAHER
Dermot HODSON
Affiliation:
Birkbeck College
Imelda MAHER
Affiliation:
University College Dublin*
Rights & Permissions [Opens in a new window]

Abstract

The idea that EU treaties have become too difficult to amend is a recurring one. This Article explores changing national constitutional rules and norms in the consent stage of EU treaty making in twenty-eight Member States between 1950 and 2016, asking how parliaments, people, and courts came to be much more significant for consent, what the consequences of this shift are, and offering some tentative proposals as to how the challenges this raises could be addressed. EU treaty making has become more complex, but we argue that treaties should be more rather than less difficult to amend where concerns over two-level legitimacy rather than two-level games predominate.

Keywords

EU treaty makingreferendumparliamentcomparative constitutionalismratificationlegitimacy
Type
Articles
Information
Cambridge Yearbook of European Legal Studies , Volume 21 , December 2019 , pp. 297 - 317
Copyright
Copyright © Centre for European Legal Studies, Faculty of Law, University of Cambridge

I. INTRODUCTION – THE CHALLENGE OF CONSENT

The European Union (EU) is founded on treaties to which Member States give their consent to be bound. Three actors have assumed a more prominent role in the process through which such consent is offered: parliaments, people, and courts. Parliaments were pivotal actors in the approval of the Treaty of Paris, which in 1951 founded the first of the three European Communities, but they now play a much more visible role. The involvement of five regional parliaments and the federal Parliament in Belgium is an extreme case.Footnote 1 The people are now routinely offered a say, with ten Member States promising referendums on the failed European Constitution.Footnote 2 National higher courts, especially Germany's Federal Constitutional Court, have also become familiar in EU treaty making.Footnote 3 The Treaty on the Functioning of the European Union (‘TFEU’) Article 136 amendment to establish a stability mechanism led to constitutional challenges in six Member States.Footnote 4

The idea that EU treaties have become too difficult to amend is a recurring one from European leaders. The fraught politics of approving the Lisbon Treaty, which emerged after the rejection of the Constitutional Treaty in referendums in France and The Netherlands and two referendums in Ireland, produced a plea of ‘no more treaties’ from the then UK Prime Minister Gordon Brown.Footnote 5 Treaty making was ‘taboo’ during the euro crisis, French President Emmanuel Macron later admitted, while remaining reluctant to be drawn on specific proposals for treaty amendment.Footnote 6 Given the low likelihood that EU treaties will remain unchanged—there have been more than 20 revisions since 1951—there is strong scholarly support for treaty making becoming more flexible. ‘In the light of comparative constitutional law and the practice of international organisations, the general procedure for amending the treaties is particularly rigid’, argued a high-level report on treaty amendment submitted by leading legal scholars to the European Commission in 2000.Footnote 7 Vivien Schmidt makes a similar point when she criticises current treaty making rules and norms for allowing some Member States ‘to hold the others hostage, delaying the entry into vigour of treaties approved by the others and often watering down measures desired by large majorities in futile attempts to engineer compromise’.Footnote 8 Critical too is Carlos Closa, who argues that treaty revision procedures ‘are too rigid and, hence, that national governments are increasingly tempted to channel reform via treaties outside the EU’.Footnote 9

This Article explores changing national constitutional rules and norms in the consent stage of EU treaty making in the twenty-eight Member States over the period 1950 and 2016.Footnote 10 It examines how parliaments, the people, and courts came to play a new role in this stage and what the consequences of changing consent practices were. Our findings confirm that EU treaty making became more complex but challenge the idea that treaty making was too rigid as a consequence. The increased role of parliaments, the people, and courts are associated with falling treaty amendment rates overall but treaty making by no means ground to a halt as a result of these changes, not least as governments were willing and able to circumvent such constraints. There is a neglected case for making treaties more rather than less difficult to amend where concerns over two-level legitimacy rather than two-level games predominate.

II. NATIONAL CONSTITUTIONAL RULES AND NORMS AND EU TREATY REVISION

Treaty making begins with the negotiation and conclusion stages, in which agreement on a final text is sought and secured.Footnote 11 Before a treaty can enter into force, it must pass through, what we call, the consent stage. Consent is sometimes equated with ratification, but ratification is just one of several means through which states can give their consent to be bound by a treaty.Footnote 12 Other means include signature, the exchange of instruments constituting a treaty, acceptance, approval or accession, or any other agreed means.Footnote 13 Although the primary focus of our discussion is how national constitutional rules and norms have shifted in the consent stage of EU treaty making, comparable changes have taken place at the negotiation stage too. From the Single European Act onwards, the European Parliament gradually gained entry to intergovernmental conferences. The convention process was codified in the Treaty of Lisbon so representatives of the European Parliament and national parliaments are offered a seat in the initial stages of treaty negotiation alongside the heads of state and government and the Commission.Footnote 14 The Treaty of Lisbon introduced ordinary and simplified revision procedures for amending EU treaties.Footnote 15 The ordinary revision procedure is aimed at amendments that increase or reduce the competences conferred on the EU by treaties, although not exclusively so.Footnote 16 The simplified revision procedure allows the European Council to draft treaty amendments without recourse to an intergovernmental conference or convention, but it cannot be used to increase the competences of the EU and it is limited to revisions concerning certain aspects of the Treaty on the Functioning of the European Union (TFEU).Footnote 17Pringle addressed the question of whether EU leaders, when they changed Article 136 TFEU to allow for the creation of a euro area stability mechanism, complied with the conditions laid down in the simplified revision procedure. The Court of Justice of the European Union was consequently drawn into the negotiation stage of treaty making for the first time.Footnote 18

National constitutions typically set out rules governing the consent stage of treaty making.Footnote 19 In the context of EU treaties, three rules are of particular relevance. First, there are those concerning international treaties in general. Second, many Member State constitutions contain provisions specific to the EU treaties and third, where the proposed EU treaty is such as to require an amendment of the constitution then regard also has to be had to the provisions governing constitutional amendment. A specific issue that shapes the consent stage is whether or not the state is monist or dualist, as the former does not require legislation for treaties to enter into force in domestic law with treaties viewed as directly applicable.Footnote 20

When it comes to approving EU treaties, Member States have regard for norms as well as rules. For example in Ireland, referendums are required to amend the constitution but the norm emerged that referendums would be sought for new EU treaties.Footnote 21 Constitutional norms, understandings, habits, and practices that drive constitutional lawFootnote 22 can be seen as precedents that are respected as deemed appropriate.Footnote 23 These norms, practised by the powerful in society, are more than just accepted, moral norms.Footnote 24 They may be articulated in constitutions—or not. They may be enforceable by the courts—or not. Their meaning may change over time reflecting their evolutionary nature—or they may be very stable, being seen as foundational to the constitution and the state. This section examines relevant constitutional rules and norms in all twenty-eight Member States over the period 1950–2016. It is necessary for the sake of comparability to make some simplifying assumptions about the role of parliaments, people and courts and these are explained in each sub-section below.

III. PARLIAMENTS

A. Constitutional amendment

One of the more difficult legal and political questions around consent for EU treaties by Member States is whether or not that treaty will necessitate a constitutional amendment.Footnote 25 Usually constitutional amendment rules are more stringent than treaty consent per se, although this is not always the case. For example, under the Czech Constitution, approval of a treaty that transfers certain powers to an international organisation or institution requires the same three-fifths majorities in the Houses of Parliament as for a constitutional amendment.Footnote 26

Four observations can be made around the question of whether or not constitutional amendment arises as a result of EU treaty revision. First, there are those Member States whose constitutional norms do not require constitutional amendments for EU treaty making, often because the constitution has been changed to introduce larger parliamentary majority requirements where powers are delegated under a treaty (Bulgaria,Footnote 27 Denmark,Footnote 28 Cyprus,Footnote 29 Italy,Footnote 30 Malta,Footnote 31 The Netherlands,Footnote 32 Portugal,Footnote 33 Romania,Footnote 34 Slovakia,Footnote 35 Slovenia,Footnote 36 Luxembourg,Footnote 37 and SwedenFootnote 38). Second, there are those for whom constitutional amendment in connection with the EU treaties is rare (Belgium,Footnote 39 Finland,Footnote 40 Greece,Footnote 41 and HungaryFootnote 42). Third, there are those where constitutional amendment is more likely (Austria,Footnote 43 Czech Republic,Footnote 44 France,Footnote 45 Ireland,Footnote 46 and SpainFootnote 47). There are those Member States where it is not clear when a constitutional amendment will be required as they have yet to be faced with or not yet found treaty changes to be of constitutional significance (Croatia,Footnote 48 Estonia,Footnote 49 Latvia,Footnote 50 and LithuaniaFootnote 51) or where the matter of approval remains contentious politically and legally (PolandFootnote 52). The UK, because of its partially written, wholly uncodified Constitution is a special case, which goes some way towards understanding why EU treaty making proved so contentious in this Member State.Footnote 53

Thus, for the majority of Member States (sixteen) constitutional amendment is not required or is rare. Yet for several of these states the majorities required in parliament are the same as those required for constitutional amendment (of which more below). For six Member States, constitutional amendment is more likely for Treaty changes. There are then several Member States who joined the EU more recently where it is not yet settled when constitutional amendment might be required. Hence it is necessary to look more closely at what is required even where constitutional amendment is not triggered by a proposed EU treaty in order to assess the challenge of consent.

B. Consent beyond constitutional amendment

Parliaments have increasingly had a role in consenting to international treaties and EU treaty making is part of this trend.Footnote 54 The prominence of the parliament can be determined by whether the legislature must: (1) be consulted; (2) approve by a simple or absolute majority; (3) approve by a three-fifths or two-thirds majority; or (4) whether a supermajority is required.Footnote 55 Our focus is on the rules and norms rather than the particular composition of any legislature at the time of treaty making—it may well be the case that for a minority government, even a simple majority may be a difficult if not insurmountable obstacle.Footnote 56

All constitutions in the Member States have rules and norms pertaining to treaty approval and all of them have a role for their national parliaments. There are three main themes in determining the challenge for the government of securing consent. These are first, whether the parliament is unicameral or not, with the challenge of securing consent increasing with the number of chambers involved. Second, whether unicameral or not, the size and nature of the vote required to secure consent is important. Finally, there may also be subnational parliaments and whether their consent is also required can pose a challenge.

There are fifteen unicameral parliaments in the EU-28, seven of which have constitutional norms and rules that require approval for an EU treaty by a simple majority.Footnote 57 Two Member States, Greece and Slovakia, require a three-fifths majority. Six Member States require a two-thirds majority,Footnote 58 and Sweden requires a three-quarters majority. The remaining thirteen Member States have bicameral legislatures. Of these, six Member States require only a simple majority in each House.Footnote 59 Austria and Germany require a two-thirds majority of both Houses of Parliament. Romania requires a similar majority but with both Houses sitting jointly, while Slovenia limits the two-thirds majority to the lower House. The Czech Republic requires a three-fifths majority in both Houses. Finally, two Member States have subnational parliaments that have a role to play. In Finland, the Åland Islands Parliament must consent to any EU treaty that falls within its competence in order for that treaty to take effect within the islands. However, once the Finnish Parliament approves the treaty, it does not have to wait for the Åland Islands Parliament.Footnote 60 Belgium has moved from being a unitary to a federal state and as a result of various constitutional reforms, treaty making powers have extended to the subnational parliaments where the treaty involves issues falling within their competence (which is the case for most treaties).

National parliaments have grown more prominent since the 1950s with a significant increase recorded in the late 1980s and early 1990s. Ten parliamentary chambers gave their consent to the Treaty of Paris setting up the European Coal and Steel Community among six Member States. By 2016, 44 parliamentary chambers are involved in approving EU treaties, reflecting the increased number of Member States but also the growing involvement of parliaments in agreeing to EU treaties. Hence, even among the six original Member States, only two have retained the requirement of a simple majority: Italy and the Netherlands. The practice of acceding states differ. For Greece (1981), Austria, Finland, and Sweden (1995), a three-fifths majority, a two-thirds majority in both Houses, a two-thirds majority, and a three-quarters majority were required, respectively. Post-2005, a range of majorities were required showing no particular pattern but parliaments all have a say.

IV. PEOPLE

Prospective or existing EU Member States account for more than one-third of foreign policy referendums in the world held since 1972.Footnote 61 In studying the evolution of the constitutional rules and norms that determine whether or not referendums can be called, it is important to note the possibility of a referendum, as this can act as a constraint on negotiation even if a referendum never actually takes place.Footnote 62 Portugal is the only Member State that prohibited a referendum until it changed its Constitution in 1997. Following this change, its Constitutional Court twice rejected calls for referendums on the Amsterdam Treaty and the European Constitution on the basis the question posed did not lend itself to the clarity required for a yes/no answer. A further constitutional amendment means such that a referendum is now possible but has never been called.Footnote 63

Referendums are improbable in six Member States, ie there are rules allowing for them but they are unlikely to be called.Footnote 64 Belgium seems to be the only Member State that is not moving towards greater use of referendums.Footnote 65 Under statute, a referendum can be called in Cyprus on any matter of public interest. There is no such provision in this Member State's Constitution and, given the unusual circumstances pertaining there, a referendum on an EU treaty would be improbable especially as Cyprus joined the EU without holding one.Footnote 66 Referendums on treaties are expressly prohibited in the Estonian Constitution but changes to the Constitution require a referendum. To date, no referendum has been sought for an EU treaty so it is improbable but not prohibited, given an EU treaty might be viewed as changing the constitution.Footnote 67 In Germany, a referendum is required for a constitution freely adopted by the people and a recurring issue of debate is when/whether an EU treaty will be proposed that will require a free decision of the people, given its impact on the existing Basic Law. This makes a referendum improbable but not impossible.Footnote 68 In Hungary, a referendum can be called by the National Assembly on any matter falling within its functions and powers. This constitutional provision has not yet been invoked in relation to an EU treaty rendering a referendum improbable.Footnote 69 Finally, the Italian Constitution prohibits a referendum being called on an international treaty. The President is also curtailed from calling one if the treaty has been approved by Parliament. A referendum on amending the Constitution could be held but this has not happened to date rendering such a referendum improbable.Footnote 70

By far the largest category is those Member States where a referendum is possible mainly because the EU treaty may be deemed to amend the constitution—usually by delegating authority. Twenty states fall within this category.Footnote 71 The trend has generally been towards allowing the possibility of a referendum.Footnote 72 A referendum can be binding and non-binding.Footnote 73 They can be called on the strength of a public petition, which can be constrained as one element of the process allowing for a referendum to be called or can be a recent constitutional development. There can be a threshold for the majority required, eg under the Croatian Constitution, the requirement went from 50 percent of the electorate to the less demanding 50 percent of those voting must support the question.Footnote 74 Some constitutions are vague as to the circumstances as to when a referendum might be invoked with no statute underpinning the empowering constitutional provision.Footnote 75 Different actors can trigger a referendumFootnote 76 and calls for referendums can be rejected.Footnote 77 Denmark has had the largest number of referendums (eight) since accession in 1972. Referendums can be called for international treaties or for any bill (including one requiring a referendum on a treaty) with the rules long pre-dating the EU.Footnote 78 However, it does not always have one and a referendum is not inevitable.

The smallest category is that where referendums are probable, and hence most likely, with only two Member States: Ireland and the UK. A referendum has never been held on an EU treaty per se in the UK, but only on whether or not to remain a member.Footnote 79 However, the 2011 European Union Act did allow for referendums on new EU treaties but was never invoked. For Ireland, the constitutional norm developed of having a referendum for EU treaties following a Supreme Court ruling that any treaty that altered the essential scope or objectives of the Communities would require a referendum.Footnote 80

For seventeen Member States, the question of referendum arises because it is required for constitutional amendment and an EU treaty may be deemed to constitute such an amendment.Footnote 81 This may be in addition to other provisions allowing for referendums.Footnote 82 Even if a referendum is prohibited for treaties, if the treaty constitutes an amendment of the constitution then a referendum may become necessary to agree to an EU treaty.Footnote 83 Sometimes the question of a referendum only arises where there is a total revision of the constitution but there is the possibility that an EU treaty may be deemed to necessitate such a revision.Footnote 84 On average, referendums have become possible in the EU, having previously been improbable, leading Qvortrup to note that referendums have become the bargaining chip of choice for EU treaty making.Footnote 85

V. COURTS

Constitutional review has grown with 83 percent of the world's constitutions permitting it in 2011, compared to only 38 percent doing so in 1950.Footnote 86 Similarly in the context of EU treaty making, constitutional review has also become more evident with seven challenges to the Lisbon Treaty.Footnote 87 We are concerned with the extent to which there is ex ante constitutional review of EU treaties as a means through which political actors can realise their political objectives via securing delay or derailment of the EU treaty under consideration.Footnote 88 Our focus is not on how the courts dealt with any challenges or which court, ie we are not concerned with the level of court where the case is initiated provided it can have an impact. Where review is to an advisory body but the constitutional norm is to follow that advice, we regard it as having impact.Footnote 89 Hence we explore what the potential is for court involvement. It is worth noting that this issue is of greater legal complexity than that of the role of parliaments and referendums because some of the law is judge made. Three broad categories can be identified.

First, some Member States do not have ex ante review at all.Footnote 90 Second, there are seven Member States that do not allow for review of the proposed EU treaty with the constitution.Footnote 91 Finally, in others, the Constitutional CourtFootnote 92 or a Constitutional Review Body, that may or may not be quasi-judicial in nature,Footnote 93 may be called on to determine the compatibility of the EU treaty with the constitution. Undoubtedly, the most prominent court in this regard is the German Constitutional Court, which has extended standing rules and the scope of review to allow for constitutional review of EU treaties.Footnote 94 One of the most significant issues is the scope of standing, as the range of potential litigants is one factor in rendering constitutional review more likely. Where review is allowed, typically heads of state, members of parliament, or governments can refer cases to the constitutional court. As of 2016, five states allow citizens to bring actions.Footnote 95 When the Treaty of Paris was being approved, ex ante constitutional review was only possible in one of the Member States. Now there is some scope for constitutional review in eighteen Member States, with a major change to be seen in Cyprus—the only Member State where constitutional review became impossible. For others—including France, Germany, Ireland, Latvia, and Slovakia—the prospect of constitutional review of EU treaties increased underlying the greater role for courts in treaty making in Europe.

VI. THE CONSEQUENCES OF COMPLEX CONSENT

The increased complexity of securing consent for EU treaties poses two questions. First, has it had an impact on the rate of treaty amendment?Footnote 96 Second, how have the Member States responded to instances where a treaty is not approved at the national level or was at risk of non-approval? To answer the first question, it is necessary to look at treaty amendments on an article-by-article basis. This does not pick up on the significance of a change in a particular article (eg the introduction of Article 50 of the Treaty on European Union (‘TEU’) allowing states to leave the EU), but it does give a sense of the increasing quantity of amendments. This is possible to do as each element of EU treaty amendments has its own CELEX number. Hence, the Maastricht Treaty has more than 1,000 articles. This shows comprehensive treaty amendment followed by more piecemeal amendments, with the overall rate of amendment declining after the Treaties of Rome and again after Maastricht. As the treaties grow longer, they become more difficult to amend. Further, regression analysis also shows that the rate of amendment slows as the people assume a more prominent role at consent stage, showing the long shadow of referendums on treaty making.Footnote 97 The greater involvement of parliaments is more subtle with greater selectivity as to what is amended rather than the number of treaties. It is also worth noting that the involvement of courts has not led to a reduction in the rate of treaty amendment. This may reflect the more dialogic nature of the courts as institutions and their tendency not to stop treaties but to delay.

Source: Hodson and Maher (2018), p 224.

Treaty Amendment Rate (Article-by-Article, 1952–2016)

Turning to the second question,Footnote 98 the European Defence Community Treaty is the best example of the Community abandoning a treaty after it had been rejected by domestic constituents, and, even so, it is not a clear-cut case. Signed in May 1952, this treaty had been approved by Germany and the Benelux countries by the time the French National Assembly rejected it in August 1954. In fact, the National Assembly did not vote down the treaty per se but a motion to debate it, although it was accepted by deputies on the basis that the treaty had been defeated. The foreign ministers of the Six made no serious attempt to save the treaty after this vote for two reasons. The first is that the question of how to re-arm Germany, a key motivation for the European Defence Community Treaty, was urgent. Rather than wait for the Six to find a solution, the five signatories of the Brussels Treaty (1948)—Belgium, France, Luxembourg, the Netherlands, and the United Kingdom—agreed to modify this agreement to allow for the creation of a new Western European Union, to which Germany and Italy would accede. The second reason is that Pierre Mendes France, President of the Council of Ministers, had sought unsuccessfully to reopen negotiations on the European Defence Community Treaty prior to its consideration by the French National Assembly. Neither the Assembly nor the foreign ministers of the Six endorsed Mendes France's proposed protocol. This left the government in an ambivalent position on the European Defence Community Treaty—which it put to Parliament without formally requesting its approval—and without a mandate to renegotiate it. The rejection of the treaty was a defeat for the French government in one sense but a victory in another.Footnote 99

National parliaments, in spite of their increased role in treaty making, tend to support treaty amendments. The troubled passage of the Maastricht Treaty before the British Parliament was the closest the Community came to the parliamentary defeat of a treaty since 1954. The European Communities (Amendment) Act 1993, through which former Prime Minister John Major's government sought to give legal effect to the Maastricht Treaty, won a large majority in the House of Commons at third reading, albeit after opponents of the treaty had tabled more than 600 amendments in an effort to derail the treaty. Member of Parliament continued to table amendments even after the Act received Royal Assent, with the government losing a vote concerning the treaty's Social Chapter. The day after its defeat, Major called a confidence vote on a revised resolution and won the support of Members of Parliament. This parliamentary siege of Maastricht, as David Baker, Andrew Gamble, and Steve Ludlam call it, was spectacular but short lived.Footnote 100 Although there was no shortage of parliamentary dissent over subsequent treaty amendments in the United Kingdom and elsewhere, no legislature has since come as close to vetoing a treaty amendment.

Higher courts have shown a willingness to scrutinise EU treaties but with little appetite for precipitating involuntary defection. A possible exception in this regard is Ireland's Supreme Court, which, in the Crotty case, ruled that Title III of the Single European Act required an amendment to Ireland's constitution.Footnote 101 In so doing, the Court upheld an appeal against the approval of the Single European Act via legislation alone, thus necessitating a referendum before the treaty could be ratified. In all probability, however, the Court neither sought nor expected to jeopardise the ratification of the treaty through this ruling. The Single European Act was predictably backed by a sizable majority of Irish voters, and it was only after the Irish government chose to run referendums on major treaty amendments rather than face similar challenges that ‘no’ votes in treaty-related referendums occurred.Footnote 102 France's Constitutional Court also left its mark on EU treaty making, albeit it with less dramatic effects. By ruling that the Maastricht Treaty required a constitutional amendment and concluding likewise in relation to the Amsterdam and Lisbon treaties, the Court raised the threshold for treaty amendments from a simple majority to a three-fifths majority requirement among the combined membership of the National Assembly and Senate.Footnote 103 The higher threshold did not endanger the approval of these treaties, however, with the more consequential ‘petit oui’ for Maastricht and the ‘grand non’ against the European Constitution arising from presidential decisions to hold referendums. The German constitutional court has arguably taken the lead in this regard in the Maastricht and Lisbon Treaty judgments setting down a marker that revisions might at some stage require total revision of the Basic Law.Footnote 104

The risk of treaty rejections is greatest in relation to referendums. The issue is what happens where there is a no vote. Denmark's initial ‘no’ to Maastricht was superseded by a ‘yes’ vote following an agreement to recognise certain exceptions for Denmark. All political parties had agreed to the proposal and, even though the legal status of the exceptions was uncertain, they were sufficient to secure a positive response from the electorate.Footnote 105 Similarly, in the Irish no votes on Nice and Lisbon, an analogous approach was adopted with the EU not reopening the treaty but exceptions being negotiated and ultimately accepted by the people.Footnote 106 The collapse of the European Constitution followed a no vote by France and the Netherlands, with several states cancelling proposed referendums after these votes.Footnote 107

Member States’ increased recourse to international treaties provides another example of how the increased complexity of EU treaty making does not go hand-in-hand with flexibility. There is some evidence to suggest that EU leaders see international treaties as easier to win consent for in their own domestic arenas. In January 2012, Herman Van Rompuy floated the idea that the Fiscal Compact could be drafted in a way to obviate the need for either a referendum or parliamentary approval.Footnote 108 In point of fact, this proposal sought to transform the Fiscal Compact from an exercise in treaty making to one of legislation by activating a provision in the TFEU that allows the Council to amend Protocol No 12, which concerns the excessive deficit procedure, by means of a special legislative procedure.Footnote 109 Although Van Rompuy's trial balloon quickly burst, the idea that the Fiscal Compact would not necessitate approval via the usual channels persisted. Cyprus broke from EU treaty making norms by approving this agreement by means of government decree rather than an Act of the House of Representatives.Footnote 110 Ireland's break from EU treaty making norms proved less successful. Having routinely put significant EU treaty amendments to a referendum since Maastricht, the Irish government initially refused to be drawn on whether the people would be offered a vote on the Fiscal Compact.Footnote 111 In the end, the government followed the Attorney General's advice that a referendum was required,Footnote 112 which was approved by a margin of 60.3 percent to 39.7 percent.

VII. TWO THEORETICAL PERSPECTIVES ON TREATY REFORM

Robert Putnam's seminal two-level game approach explores how domestic and international politics are entangled.Footnote 113 This approach allows for the possibility that governments in treaty negotiation may choose to tie their hands domestically in order to restrict the range of possible outcomes at the negotiations.Footnote 114 A government that ties its hands too tightly, Putnam warns, could produce deadlock in negotiations.Footnote 115 A corollary to this point is that governments that find their ability to make treaties encumbered by the consent stage have a strong incentive to find slack either by changing or circumventing domestic constraints.

From a two-level game perspective, concerns that Member States have tied their hands too tightly through more complex constitutional rules and norms is understandable. Calls to end the unanimity requirement for approving EU treaties and to limit the use of national referendums are consistent with this view. Among academic commentators, Fernando Mendez, Mario Mendez, and Vasiliki Triga offer the most thoughtful proposals. One, inspired by the 1984 Draft Treaty on European Union, would allow a supermajority at both the negotiation and consent stages of treaty making, which the authors suggest could overcome ‘the paralysis and stagnation’ of treaty making in an enlarged Union.Footnote 116 As regards referendums, Mendez, Mendez, and Triga see such votes as ‘extra territorial’, because of the significant consequences generated by voters in one Member State on the rest of the Union. Carlos Closa makes a similar point when he argues that the rejection of a treaty generates negative externalities for other Member States, as occurred in the Exchange Rate Mechanism crisis triggered by Denmark's ‘no’ vote against the Maastricht Treaty in 1992.Footnote 117

The two-level legitimacy approach sees the privileged position of national governments in international diplomacy as being contested.Footnote 118 The rise of parliaments, the people, and courts in the negotiation and consent stages of EU treaty making are a response to, and reflection of, this contestation. As such, attempts to reassert national governments’ privileged position by circumventing the role of other actors in the consent stage could trigger a backlash. A degree of inertia in constitutional amendment procedures can be necessary and desirable if it bolsters the legitimacy of constitutional law as a form of lex superior.Footnote 119 There is a similar case for building in further inertia to EU treaty making, for making it harder for Member States to circumvent parliaments, the people, and courts, for accepting that treaty making can and should occasionally fail to produce agreements in spite of the diplomatic sunk costs involved.

‘Is it right’, Hervé Bribosia asks, that the refusal of a few hundred thousand inhabitants should be allowed to block a reform desired by the representatives of five hundred million people?’Footnote 120 Where legitimacy is a primary consideration, the answer, is ‘yes’. The consent to be bound is a defining principle of treaty making between sovereign states. In practice, new treaties do not typically enter into force unless the parties to it have expressed their consent to be bound.Footnote 121 The question of consent for treaty amendments is more intricate. Under the Vienna Convention on the Law of Treaties, an amendment to a multilateral treaty cannot bind a state that is party to the original treaty without that state having become a party to the amending agreement.Footnote 122 In practice, multilateral treaties often include standard amendment procedures that allow a qualified majority of parties to amend the existing treaties, but states that oppose such amendments typically cannot be bound by such amendments against their will.Footnote 123 Multilateral treaties sometimes include simplified amendment procedures, which in some cases allow for a majority of states to approve amendments that apply to all parties to the original agreement, whether they have given their consent to be bound by such changes or not or by allowing or compelling non-consenting states to withdraw from the treaty.Footnote 124

Insofar as the regulation of referendums can be justified from a two-level legitimacy perspective, it would be to discourage the EU's tendency to overlook referendum results, as occurred in the re-run referendums over Maastricht, Nice, and Lisbon and the willingness to salvage much of the European Constitution in spite of referendum votes against it. There are arguments in favour of second referendums. Sinnott, for example, finds a marked increase in both communication and understanding in Ireland's re-run referendum on the Nice Treaty.Footnote 125 But this practice is all too easily criticised, as Gráinne de Búrca notes, for failing ‘to respect the outcome of legitimate constitutional processes and [undermining] the democratically expressed will of the people’.Footnote 126 Stephen Tierney goes further by seeing second referendums as sending a clear message ‘that national electorates will not be allowed to frustrate closer integration’.Footnote 127

Other ideas for making EU treaties harder to amend, include time-locks on treaty reform, citizen-led treaty making, and greater oversight of treaty making. The simplest form of time lock would be to limit the number of treaty amendments that can occur within a particular time period. The Treaty of Paris included such a provision by prohibiting amendments to itself until a five-year transition period had expired.Footnote 128 One way of giving citizens a say would be to hold periodic pan-European referendums on whether to amend EU treaties, for example, every five years. As regards greater oversight, there is a case for routinely inviting the Court of Justice to give an opinion on whether the use of the simplified revision procedure over the ordinary revision procedure is justified rather than waiting for a reference from a national higher court. There is a similar case for making the use of the simplified revision procedure subject to the consent of the European Parliament, as occurs when the European Council seeks to employ the ordinary revision procedure without convening a convention.

VIII. CONCLUSION

Who makes treaties and how such actors are held to account are recurring concerns in the study of international law.Footnote 129 Few instances of treaty making can match the EU for intensity and controversy. During the period 2010–2011 alone, EU Member States launched a combined 105 national consent procedures connected to treaty amendments.Footnote 130 Once thought of as epoch-making events, treaty amendments are now part of the ‘everyday politics’ of the EU, argues Thomas Christiansen.Footnote 131 And yet treaties are no less controversial for this. The approval of the Maastricht Treaty, which was rejected by Danish voters in a referendum and only narrowly endorsed by their French counterparts, intensified popular concerns over EU treaty change.Footnote 132 Thirteen years later, referendums on the European Constitution in France and the Netherlands produced a popular backlash against a treaty that was designed to bring the EU closer to its people.Footnote 133 The United Kingdom had planned to hold a referendum on this treaty, but its failure to do so on earlier or later agreements goes some way towards explaining why former Prime Minister David Cameron called and lost a referendum in 2016 on the United Kingdom's continued membership of the EU.

The EU is an important case not only because its treaties are in flux but because the process through which Member States give their consent to be bound to such agreements has changed. In 1951, ten parliamentary chambers participated in the approval of the Treaty of Paris. All Member States approved this agreement by means of a simple majority vote; there were no referendums or ex ante constitutional reviews. By 2016, the comparative constitutional analysis presented in this Article has shown, forty-four parliamentary chambers were involved in approving major EU treaties and most Member States required approval by a parliamentary majority of two-fifths or more. Referendums were either possible or probable in twenty-two Member States and there was scope for the ex ante constitutional review of EU treaties in eighteen Member States.

EU treaty making has become more complex but such complexity has not produced gridlock and does not necessarily argue for greater flexibility in the ways that Member States give their consent to be bound. Parliaments and national higher courts, though influential, tend to approve treaties. Member States, meanwhile, generally work around referendum votes against EU treaties rather than abandoning such agreements. As EU treaty making has become more complex, Member States have also turned increasingly to international law treaties, in part because they are perceived as being easier to approve.

Arguments in favour of EU treaty making becoming more flexible rest uneasily with these findings but chime with Robert Putnam's two-level game approach. Member states that tie their hands too tightly in the consent stage of treaty making, it follows, should seek more room for manoeuvre in the domestic arena or risk rejecting treaties. The two-level legitimacy approach, in contrast, warns against untying hands without understanding the reasons why they have been tied to begin with. It argues for less rather than more flexibility in the consent stage, with time locks on treaty reforms, citizen-led treaty making and greater oversight of treaty having the potential to incorporate further inertia into EU treaty making.

Footnotes

*

Earlier versions of this Article were presented at the European Union Studies Association conference, Denver 2019, the Society of Legal Scholars conference Queen Mary London 2018, the UCD Sutherland School of Law research seminar, and the DCU School of Law and Government research seminar 2019. Thanks to participants for comments and to Ronan Riordan for able research assistance. The usual disclaimer applies.

References

1 On the role of parliaments in EU treaty making, see Closa, C, The Politics of Ratification of EU Treaties (Routledge, 2013)CrossRefGoogle Scholar.

2 Mendez, F, Mendez, M, and Triga, V, Referendums and the European Union: A Comparative Inquiry (Cambridge University Press, 2014)CrossRefGoogle Scholar.

3 Mendez, M, ‘Constitutional Review of Treaties: Lessons for Comparative Constitutional Design and Practice’ (2017) 15(1) International Journal of Constitutional Law 84CrossRefGoogle Scholar.

4 These states are Austria, Estonia, Germany, Hungary, Ireland, and Poland.

5 ‘Brown Rules out EU Referendum as Opponents Begin 3-Month Battle to Block Treaty’ (Evening Standard, 14 December 2007).

6 Michel Rose, ‘France's Macron Says EU Treaty Change “Not Taboo”’ (Reuters Business News, 15 May 2017).

7 ‘Reforming the Treaties’ Amendment Procedures’, Second Report on the Reorganisation of the European Union Treaties Submitted to the European Commission on 31 July 2000, Robert Schuman Centre for Advanced Studies, p 9.

8 Schmidt, V A, ‘Re-envisioning Envisioning the European Union: Identity, Democracy, Economy’ (2009) 47(s1) Journal of Common Market Studies 17, pp 26–27CrossRefGoogle Scholar.

9 Closa, C, ‘Between a Rock and a Hard Place: The Future of EU Treaty Revisions’ (2014) 2 SIEPS European Policy Analysis, p 2Google Scholar.

10 This Article draws extensively on Hodson, D and Maher, I, The Transformation of EU Treaty Making: The Rise of Parliaments, Referendums and Courts Since 1950 (Cambridge University Press, 2018)CrossRefGoogle Scholar. At the time of writing the UK is still a member of the EU and hence is included in references to Member States save where otherwise stated.

11 Korontzis, G, ‘Making the Treaty’ in Hollis, DB (ed), The Oxford Guide to Treaties (Oxford University Press, 2012), p 185Google Scholar. For this section see note 10 above, pp 1–34

12 Ibid, pp 195–201.

13 Vienna Convention on the Law of Treaties, Art 11, 23 May 1969, 1115 UNTS 331.

14 Even though the failed European Constitution had been the first time the convention was used. See Christiansen, T and Reh, C, Constitutionalizing the European Union (Palgrave Macmillan, 2009), p 242CrossRefGoogle Scholar; Art 48(3) TEU. The European Central Bank also is consulted in relation to institutional changes in the monetary area.

15 Article 48(1) TEU.

16 Article 48(2–5) TEU.

17 Article 48(6) TEU.

18 Pringle v Government of Ireland, C-370/12, EU:C:2012:756. Constitutional challenges were brought in Austria, Estonia, Germany, Hungary, Ireland, and Poland. See Reestman, J-H, ‘Legitimacy through Adjudication: The ESM Treaty and the Fiscal Compact before the National Courts’ in Beukers, T, de Witte, B, and Kilpatrick, C (eds), Constitutional Change through Euro-Crisis Law (Cambridge University Press 2017), pp 243–78CrossRefGoogle Scholar.

19 In the UK, the rules were contained in legislation. See the European Union Act 2011. There is no single constitutional document in the UK.

20 Although legislation is sometimes required to bind the state internationally. See Sloss, D, ‘Domestic Application of Treaties’ in Hollis, D B (ed), Oxford Guide to Treaties (Oxford University Press, 2012), pp 367–95Google Scholar.

21 Finke, D, ‘Domestic Politics and European Treaty Reform: Understanding the Dynamics of Governmental Position-Taking’ (2009) 10(4) European Union Politics 501CrossRefGoogle Scholar. Hodson, D and Maher, I, ‘British Brinkmanship and Gaelic Games: EU Treaty Ratification in the UK and Ireland from a Two-Level Game Perspective’ (2014) 16(4) The British Journal of Politics and International Relations 651CrossRefGoogle Scholar.

22 Dicey, AV, The Law of the Constitution (Oxford University Press 2013), p 20Google Scholar.

23 Jennings, 1967, quoted in Albert, R, ‘How Unwritten Constitutional Norms Change Written Constitutions’ (2015) 38(2) Dublin University Law Journal 387Google Scholar. Wiener also notes their appropriateness. See Wiener, A, ‘Contested Meanings of Norms: A Research Framework’ (2007) 5(1) Comparative European Politics 1CrossRefGoogle Scholar.

24 Waldron, J, ‘Are Constitutional Norms Legal Norms?’ (2006) 75(3) Fordham Law Review 1697Google Scholar.

25 See generally Besselink, L F M, ‘The Kingdom of the Netherlands’ in Besselink, L F M, et al. (eds), Constitutional Law of the EU Member States (Kluwer, 2014), pp 1187–241Google Scholar.

26 Art 10a Constitution of the Czech Republic. See also Art 23(1) Basic Law of the Federal Republic of Germany; Closa, C, The Politics of Ratification of EU Treaties (Routledge, 2013), pp 50–59CrossRefGoogle Scholar.

27 Arts 155, 161 Constitution of Bulgaria.

28 See Sec 20 Constitutional Act of Denmark. Constitutional amendments are rare. See T Knudsen and U Jakobsen, The Danish Path to Democracy (Paper for the 2nd ECPR General Conference, Marburg, 18–21 September 2003), p 9.

29 Due to the differences between the written constitution and the functional constitution as a result of the separation of Cyprus in 1974, it exceptionally did not amend its Constitution to accede to the EU and constitutional amendment is not envisaged for future treaties. See Art 179(2) Constitution of the Republic of Cyprus; A Emilianides ‘Cyprus: Everything Changes and Nothing Remains the Same’ in S Farran, E Örücü, and Donlan, S P (eds), A Study of Mixed Legal Systems: Endangered, Entrenched or Blended (Ashgate 2014), p 236Google Scholar; Mendez, Mendez, and Triga, note 2 above, pp 227–28.

30 Constitutional Court of the Italian Republic, No 183/1973, Judgment. F Fontanelli and G Martinico, ‘Cooperative Antagonists: The Italian Constitutional Court and the Preliminary Reference: Are We Dealing with a Turning Point?’ (Eric Stein, 2008) Working Paper 5/2008, p 3.

31 Art 66 Treaty of Malta. The norm of not looking for constitutional amendment has been criticised. See A S Trigona, ‘A Sham Ratification’ (Times of Malta, 10 July 2012)

32 See Arts 91(3), 137 Constitution of the Kingdom of the Netherlands. The exception of course is the referendum around the European Constitution. If this is seen as introducing a new constitutional norm then The Netherlands would be categorised as constitutional amendments being rare.

33 Arts 7–8 Constitution of the Republic of Portugal.

34 Art 148(1) Constitution of Romania.

35 A three-fifths majority is required for the delegation of powers by an international agreement which is the same majority for constitutional amendments. See Arts 7, 84 Constitution of the Republic of Slovakia.

36 The Constitution was amended prior to accession removing the possibility of a referendum on EU Treaties. A two-thirds majority has become the norm. See Art 3a Constitution of the Republic of Slovenia.

37 They can take place after ratification. See Art 114 Constitution of Luxembourg. See generally Gerkrath, J, ‘Constitutional Amendment in Luxembourg’ in Contiades, X (ed), Engineering Constitutional Change (Routledge, 2012), p 247Google Scholar.

38 Instrument of Government, chs 4, 5, 8, 10 (1974, amended to 2015).

39 Art 195 Constitution of Belgium; see also Claes, M ‘Constitutionalizing Europe at Its Source: The ‘European Clauses’ in the National Constitutions: Evolutions and Typology’ (2005) 24(1) Yearbook of European Law 86, 97CrossRefGoogle Scholar. C Behrendt ‘The Process of Constitutional Amendment in Belgium’ in Contiades, note 37 above, pp 35–50.

40 Art 73 Constitution of Finland.

41 Art 110 Constitution of Greece;Spyropoulos, P K and Fortsakis, T, Constitutional Law in Greece (Kluwer Law International 2009), pp 79–80Google Scholar.

42 Arts E(4), S(2) Fundamental Law of Hungary.

43 See eg Art 44 Austrian Constitution (amended in 2013). An explicit constitutional amendment is now required where a treaty is at odds with the constitution.

44 Art 10a Constitution of the Czech Republic. See T Dumbrovsky ‘Constitutional Change through Crisis Law: Czech Republic’ (2014) European University Institute: Fiesole. The President has to consent to international treaties and the President sought controversially, and ultimately unsuccessfully, to delay the ratification of the Lisbon Treaty. See Marek, D and Baum, M, The Czech Republic and the European Union (Routledge, 2010), p 49CrossRefGoogle Scholar.

45 Art 54 Constitution of the French Republic. See Closa, note 1 above, p 102.

46 Art 29.5.1 Constitution of Ireland.

47 Arts 95(1), 167–168 Spanish Constitution. The Constitution was revised for the Maastricht Treaty to allow European citizens votes in municipal elections. See Kumm, M and Comella, V F ‘The Primacy Clause of the Constitutional Treaty and the Future of Constitutional Conflict in the European Union’ (2010) 3 International Journal of Constitutional Law 475Google Scholar.

48 Art 143 Constitution of Croatia. Article 140 allows for delegation of powers to international organisations, reducing the need for a constitutional amendment for EU treaty revision.

49 Art 65 Constitution of the Estonian Republic; Sec 126 Riigikogu Rules of Procedure and Internal Rules Act (2007); Mendez, Mendez, and Triga note 2 above, p 56.

50 See Arts 68, 76 Constitution of Latvia. See also Rasnača, Z, Constitutional Change Through Euro Crisis Law: Latvia (European University Institute, 2013)Google Scholar.

51 Arts 148, 150 Constitution of the Republic of Lithuania. In fact, the constitutional amendment for the introduction was deemed to be unnecessary. See Case 22/2013, 24 January 2014.

52 Arts 89–90 Constitution of Poland; Zwolski, K ‘Euthanasia, Gay Marriages and Sovereignty: Polish Ratification of the Treaty of Lisbon’ (2009) 5(3) Journal of Contemporary European Research 493Google Scholar.

53 The European Union Act 2011 introduced referendums for many, but not all, Treaty revisions.

54 See Closa, note 1 above, pp 44–46.

55 On the difficulty of constitutional amendment relative to majorities required for approval, see the influential study of Lutz, D S, ‘Toward a Theory of Constitutional Amendment’ (1994) 88(2) American Political Science Review 360CrossRefGoogle Scholar.

56 For a study of this nature, see Hug, S and König, T, ‘In View of Ratification: Governmental Preferences and Domestic Constraints at the Amsterdam Intergovernmental Conference’ (2002) 56(02) International Organization 454CrossRefGoogle Scholar.

57 Cyprus, Denmark, Estonia, Latvia, Lithuania, Malta, and Portugal. In Denmark, treaties that do not transfer sovereignty—or more accurately that do not delegate powers ‘vested in the authorities of the Realm’—are passed by a simple majority. Treaties entailing such a transfer can be approved by Parliamentary channels alone by five-sixths of the members of the Folketing.

58 Bulgaria (although if it were to be viewed as a constitutionally amending Treaty then a three quarters majority would be required), Croatia, Denmark, Finland, Hungary, and Luxembourg.

59 Belgium, Italy, the Netherlands, Poland, Slovenia, Spain, and the United Kingdom.

60 Sec 58 Act on the Autonomy of Åland 1991/1144.

61 Calculation based on the Centre for Research on Direct Democracy database available at www.c2d.ch.

62 Gerber, E R and Hug, S, ‘Legislative Response to Direct Legislation’ in Mendelsohn, M and Parkin, A (eds), Referendum Democracy (Springer, 2001), pp 88–108CrossRefGoogle Scholar; G Majone ‘The “Referendum Threat”, the Rationally Ignorant Voter, and the Political Culture of the EU’ (University College Dublin Law, 2009) Research Paper 4, pp 17–18.

63 Portuguese Constitutional Court Decision 531/98 and 704/2004. The new provision is Article 295 of the Constitution of Portugal.

64 Belgium, Cyprus, Estonia, Germany, Hungary, and Italy.

65 Qvortrup, M, A Comparative Study of Referendums: Government by the People (Manchester University Press, 2005), p 1Google Scholar.

66 Cyprus, Law on Referendums 208/1989.

67 See Arts 106, 163 Constitution of the Estonian Republic.

68 Art 146 Basic Law of the Federal Republic of Germany. See also German Federal Constitutional Court, Judgment of the Second Senate, 2 BvE 2/08 (30 June 2009).

69 Art 8(3) Fundamental Law of Hungary. There were two unsuccessful challenges to not having a referendum on the Lisbon treaty and the European Constitution. See Constitutional Court of Hungary, Decision No 61/2208 and Decision No 6/2005 (I.13). See also Mendez, Mendez, and Triga, note 2 above, p 82.

70 Arts 11, 75, 78(6) Constitution of the Italian Republic.

71 Austria, Bulgaria, Croatia, the Czech Republic, Denmark, Finland, France, Greece, Latvia, Lithuania, Luxembourg, Malta, The Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, and Sweden.

72 For example, in France referendums became possible under the 5th Republic. See Art 11 Constitution of the Republic of France. In Greece, Article 44(2) of the Constitution of the Republic of Greece, which allows for referendums, has gained traction following the fiscal crisis. See Ungerer, J and Ziaka, L, ‘Reflections on the Greek Capital Controls: How the Rescue of the National Economy Justifies Restricting Private Business’ (2017) 44(2) Legal Issues of Economic Integration 135Google Scholar. In 2003, Article 114 of the Constitution of Luxembourg introduced a referendum for constitutional amendment—which may be required in relation to an EU treaty. The Dutch introduced a popular petition for a referendum in 2015. See Advisory Referendum Act (Wet raadgevend referendum) 2015.

73 See eg Arts 43, Article 49b Constitution of Austria; Pelinka, A and Greiderer, S ‘Austria: The Referendum as an Instrument of Internationalisation’ in Uleri, P V and Gallagher, M (eds), The Referendum Experience in Europe (Springer, 1996), pp 20–32CrossRefGoogle Scholar. Sweden allows for both binding and non-binding referendums favouring the latter. See Bernitz, U, ‘Sweden and the European Union: On Sweden's Implementation and Application of European Law’ (2001) 38(4) Common Market Law Review 903CrossRefGoogle Scholar.

74 Art 87 Constitution of Croatia.

75 See eg Art 10A Constitution of the Czech Republic.

76 See eg Article 68 Constitution of Latvia, which allows the parliament to call a referendum provided half of its members vote for it and there are to be substantial changes to Latvia's membership of the EU. Article 90 of the Constitution of Romania allows the President to call a referendum on matters of national interest (which could of course include an EU treaty).

77 See Art 3a Constitution of Slovenia. A petition by voters to call a referendum on the Lisbon treaty failed as there were not enough signatures.

78 Secs 20, 42 Constitutional Act of Denmark.

79 See R(Miller) v Secretary of State [2017] UKSC 5.

80 Crotty v An Taoiseach [1987] IR 713. The Constitution was amended to ensure specifically that any Treaty adopting a common defence that would include Ireland has to be put to a referendum. See Art 29.4.9 Irish Constitution.

81 Art 42 Constitution of Austria; Art 87 Constitution of Croatia; Art 89 Constitution of the Republic of France; Art 8(3)(a) Foundational Law of Hungary; Art 46 Constitution of Ireland; Art 114 Constitution of Luxembourg; Art 66(3) Constitution of Malta; Arts 11(3), 147(3), 151(3) Constitution of Romania; Art 95(1) Constitution of Spain; Art 15 Instrument of Government (1974, amended to 2015) of Sweden (the referendum must be triggered by members of parliament). See also Besselink, note 25 above.

82 See eg Hungary; Arts 75, 138 Constitution of Italy; Art 148 Constitution of Lithuania; Art 235(6) Constitution of Poland.

83 Art 163 Constitution of Estonia; Art 77 Constitution of Latvia.

84 Austria; see also Art 146 Constitution of Germany.

85 Qvortrup, M, ‘The Three Referendums on the European Constitution Treaty in 2005’ (2006) 77(1) The Political Quarterly 89CrossRefGoogle Scholar.

86 Robertson defines constitutional reviews as ‘a process by which one institution, commonly called a constitutional court, has the constitutional authority to decide whether states or other decrees created by the rule-making institutions identified by the constitution are valid given the terms of the constitution’. See Robertson, D, The Judge as Political Theorist: Contemporary Constitutional Review (Princeton University Press, 2010), p 5CrossRefGoogle Scholar; Ginsberg, T and Versteeg, M, ‘Why Do Countries Adopt Constitutional Review?’ (2014) 20(3) Journal of Law, Economics, & Organization 587Google Scholar.

87 Ortiz, P J C, EU Treaties and the Judicial Politics of National Courts: A Law and Politics Approach (London: Routledge 2015), p 8CrossRefGoogle Scholar; see also Wendel, M ‘Lisbon before the Courts: Comparative Perspectives’ (2011) 7(1) European Constitutional Law Review 96CrossRefGoogle Scholar.

88 Ex post constitutional review is (potentially) a form of treaty breaking rather than treaty making although we note that 11 Member States allow for ex post review. They are: Austria, Belgium, Croatia, the Czech Republic, Denmark, Estonia, Finland, Italy, Lithuania, The Netherlands, and The United Kingdom.

89 See the Council on Legislation in Sweden, whose advisory opinions are followed as a constitutional norm, and compare the position in Belgium, where the opinions of the Council of State are not necessarily followed by government.

90 Austria, Belgium, Croatia, Italy, Malta, The Netherlands. See eg the decision of the Austrian Constitutional Court in Case G 62/05 Constitutional Treaty, 18 June 2005; Case SV 2/08-3 et al Treaty of Lisbon, 30 September 2008; and Case SV 1.10-9 Treaty of Lisbon II, 12 June 2010. See also Wendel, note 87 above, p 111. On Belgium, see M Claes, The National Courts’ Mandate in the European Constitution (London, Bloomsbury 2006), p 218. On Croatia, see Cases U-I-1583/2000 and U-I-559/2001 (decided in 2010) and U-I-2236/2017. On Italy, see M Claes, ibid, p 622. On the Netherlands, see Art 120 of Constitution of the Kingdom of the Netherlands. Note however, that an advisory opinion can be sought. See Art 73 of the Constitution of the Kingdom of the Netherlands and note 93 below.

91 Cyprus (Article 1a of the Constitution gives EU law supremacy over the Constitution removing any ex ante review of treaties since 2006); Greece (constitutional review does not extend to treaties, see M de Visser Constitutional Review In in Europe: A Comparative Analysis (Bloomsbury, 2013), p 13; Denmark (Danish Supreme Court Case I-361/1997, 6 April 1998 Carlsen. Rasmussen. See also Claes, note 90 above, p 490. Luxembourg (Art 95, TER, 2, Constitution of Luxembourg (1868, as amended to 2009); Malta (Art 95, Constitution of Malta (1964, as amended to 2014)); The Netherlands (Art 120, Constitution of the Netherlands (1815, as amended to 2008)).

92 Bulgaria, see E Tanchev and M Belov ‘Constitutional Gradualism: Adapting to EU Membership and Improving the Judiciary in the Bulgarian Constitution’ (2009) 14(1) European Public Law 3; Arts 87(2), 97(2) Constitution of the Czech Republic. See also the Constitutional Court Act 1993; the Estonian Constitutional Review Court Procedure Act 2002 allows for such review as does the Hungarian Act on the Constitutional Court 1989. Constitutional review is possible in Ireland either via a reference to the Court by the President under Article 26 of the Constitution or by a citizen. See Crotty v An Taoiseach [1987] IR 713; Latvia Sec 16(3) Constitutional Court Law; Lithuania Arts 105, 107 Lithuanian Constitution; Poland, Art 133(2) of the Constitution; Portugal, Arts 134g, 278 of the Constitution; Romania, Art 146(2) of the Constitution; Slovakia, Art 125a(1) Constitution; Slovenia, Art 160 of the Constitution; Spain, Art 95(2) of the Constitution.

93 Opinions are advisory in Belgium, Luxembourg (although constitutional change was only required once and followed, see Art 83bis Constitution of Luxembourg); the Netherlands (Art 73 of the Constitution); followed as a constitutional norm in Sweden (Instrument of Government, ch 8, Art 18). They are binding in Finland (Sec 74 Constitution of Finland, Ojanen 2004, p 205); France (Art 54 of the French Constitution).

94 See eg German Federal Constitutional Court, BvE 2/08, Judgment of the Second Senate of 30 June 2009; Manfred Brunner and Others v The European Union Treaty, Cases 2 BvR 2134/92 and 2159/92; German Federal Constitutional Court, Case BVerfGE 89, 155, Treaty of Maastricht, Decision of 12 October 1993.

95 Germany, Ireland, Latvia, Slovakia, and the UK.

96 For an analysis of the question of constitutional amendment, see Lutz, note 55 above. On treaty amendments, see Conca, K, An Unfinished Foundation: The United Nations and Global Environmental Governance (Oxford University Press, 2015)CrossRefGoogle Scholar. On EU treaty amendments, see Closa, C ‘Constitutional Rigidity and Procedures for Ratifying Constitutional Reforms in EU Member States’ in Benz, A and Knupling, F (eds), Changing Federal Constitutions: Lessons from International Comparison (Verlag Barbara Budrich, 2012), pp 281–310CrossRefGoogle Scholar.

97 For a detailed discussion of the regression analysis, see Hodson and Maher, note 10 above, pp 221–40. On the long shadow of referendum, see Hooghe, L and Marks, G, ‘A Postfunctionalist Theory of European Integration: From Permissive Consensus to Constraining Dissensus’ (2009) 39(1) British Journal of Political Science 22CrossRefGoogle Scholar.

98 See Hodson and Maher, note 10 above, pp 221–40.

99 International Organization, ‘European Defence Community’ (1954) 8(4) International Organization 599Google Scholar.

100 Baker, D, Gamble, A, and Ludlam, S ‘The Parliamentary Siege of Maastricht 1993: Conservative Divisions and British Ratification’ (1994) 47(1) Parliamentary Affairs 37Google Scholar.

101 Supreme Court of Ireland, Crotty v An Taoiseach [1987] IR 713 (April 1987).

102 Hodson, D and Maher, I ‘British Brinkmanship and Gaelic Games: EU Treaty Ratification in the UK and Ireland from a Two Level Game Perspective’ (2014) 16(4) The British Journal of Politics & International Relations 645CrossRefGoogle Scholar.

103 French Constitutional Court, Decision 92–308 DC, Treaty on European Union; Decision 97–394 DC, Treaty of Amsterdam Amending the Treaty on European Union, the Treaties Establishing the European Communities and Certain Related Instruments; Decision 2007–560 DC, Treaty of Lisbon Amending the Treaty on European Union and the Treaty Establishing the European Community.

104 BVerfGE 89, 155, 12 October 1993 (Maastricht); 2 BvE 262-3, 30 June 2009 (Lisbon).

105 Krunke, H ‘From Maastricht to Edinburgh: The Danish Solution’ (2005) 1(3) European Constitutional Law Review 342CrossRefGoogle Scholar; van Ooik, R and Curtin, D, Denmark and the Edinburgh Summit: Maastricht without Tears (Wiley Chancery Law, 1994), p 354Google Scholar.

106 See K Gilland, Ireland's Second Referendum on the Treaty of Nice (October 2002) Referendum Briefing No 1; Laffan, B and O'Mahony, J, Ireland and the European Union (Palgrave Macmillan, 2008), p 108CrossRefGoogle Scholar.

107 See Paris-Dobozy, M-L ‘The Implications of the “No” Vote in France: Making the Most of a Wasted Opportunity’ in Laursen, F (ed), The Rise and Fall of the EU's Constitutional Treaty (Brill, 2008), p 510Google Scholar. The Czech Republic, Denmark, Ireland, Poland. and Portugal all cancelled their referendums. See Hodson and Maher, note 10 above, p 232.

108 J Chaffin, ‘Van Rompuy Draws Up Fast-Track “Fiscal Compact”’ (Financial Times, 7 December 2011).

109 Art 126(14) TFEU.

110 Pantazatou, P, Constitutional Change through Euro Crisis Law (European University Institute, 2014), p 38Google Scholar. See Hodson and Maher, note 10 above, pp 87–120.

111 J Cienski, J Smyth, and P Spiegel, ‘Sinn Fein Legal Threat Hangs over Fiscal Deal’ (Financial Times, 20 January 2012).

112 H Stewart and H McDonald, ‘Ireland Set for Referendum on Eurozone Fiscal Treaty’ (The Guardian, 28 February 2012).

113 Putnam, R, ‘Diplomacy and Domestic Politics: The Logic of Two-Level Games’ (1988) 42(3) International Organization 430CrossRefGoogle Scholar. See Hodson and Maher, note 10 above, pp 246–67.

114 Morascsik, A ‘Armaments among Allies: European Weapons Collaboration, 1975–1985’ in Evans, P B, Jacobson, H K, and Putnam, R (eds), Double-Edged Diplomacy: International Bargaining and Domestic Politics (University of California Press, 1993), p 128Google Scholar.

115 See Putnam, note 113 above, p 441.

116 See Mendez, Mendez, and Triga, note 2 above, p 203.

117 See Closa, note 1 above, p 13.

118 See Hodson and Maher, note 10 above, pp 34–49.

119 Lane, J E, Constitutions and Political Theory (Manchester University Press, 1996), p 114Google Scholar.

120 H Bribosia ‘Revising the European Treaties: A Plea in Favour of Abolishing the Veto’ (Notre Europe, 2009), Policy Paper No 37, p 14.

121 See Korontzis, note 11 above, p 677.

122 Vienna Convention on the Law of Treaties, note 13 above, Art 40.

123 See Korontzis, note 11 above, pp 744–48.

124 Ibid, p 749.

125 Sinnott, R, Attitudes and Behaviour of the Irish Electorate in the Second Referendum on the Treaty of Nice (Institute for the Study of Social Change, 2003)Google Scholar.

126 de Búrca, G, ‘If at First You Don't Succeed: Vote, Vote Again: Analyzing the Second Referendum Phenomenon in EU Treaty Change’ (2009) 33(5) Fordham International Law Journal 1475Google Scholar.

127 Tierney, S, Constitutional Referendums: The Theory and Practice of Republican Deliberation (Oxford University Press, 2012), p 165CrossRefGoogle Scholar.

128 Art 96 Treaty Establishing the European Coal and Steel Community (Paris, 1951).

129 Finch, G A, ‘The Need to Restrain the Treaty Making Power of the United States within Constitutional Limits’ (1954) 48(1) The American Journal of International Law 57–82CrossRefGoogle Scholar; McNair, B A D M, The Law of Treaties (Oxford University Press, 1961)Google Scholar; Lobel, J ‘The Limits of Constitutional Power: Conflicts between Foreign Policy and International Law’ (1985) 71(1) Virginia Law Review 1071CrossRefGoogle Scholar; Kumm, M and Comella, V F ‘The Primacy Clause of the Constitutional Treaty and the Future of Constitutional Conflict in the European Union’ (2005) 3 International Journal of Constitutional Law 473CrossRefGoogle Scholar; Hathaway, O A ‘Presidential Power over International Law: Restoring the Balance’ (2009) 119(2) The Yale Law Journal 140Google Scholar.

130 Four treaty revisions were launched during this period. They concerned: (1) the transitional arrangements on the number of Members of the European Parliament, which were carried out via the ordinary revision procedure; (2) the revision of Article 136 TFEU, which permitted the establishment of the European Stability Mechanism and was carried out via the simplified revision procedure; (3) the accession of Croatia to the EU, which fell outside the simplified and ordinary revision procedures and was instead governed by Article 49 TEU; (4) the ratification of Ireland's Lisbon protocols via the ordinary revision procedure. A treaty change concerning the Czech Republic's ‘opt out’ out from the Charter on Fundamental Rights—an eleventh-hour concession secured by President Václav Klaus in December 2009 before he agreed to sign the Lisbon Treaty—was postponed as a result of domestic wrangling.

131 Christiansen, T, ‘Institutionalist Dynamics behind the New Intergovernmentalism: The Continuous Process of EU Treaty Reform’ in Bickerton, C J, Hodson, D, and Puetter, U (eds), The New Intergovernmentalism: States and Supranational Actors in the Post-Maastricht Era (Oxford University Press, 2015), p 95Google Scholar.

132 Franklin, M, Marsh, M, and McLaren, M, ‘Uncorking the Bottle: Popular Opposition to European Unification in the Wake of Maastricht’ (1994) 32(4) Journal of Common Market Studies 455CrossRefGoogle Scholar.

133 See note 86 above.