Julian Arato is a J.D. candidate and Institute of International Law and Justice Scholar at the New York University School of Law . His article on Subsequent Practice and Evolutive Interpretation: Techniques of Treaty Interpretation over Time and Their Diverse Consequences, 9 Law & Prac. Int’l Cts. & Tribs. (forthcoming 2010) is available here.
Editor’s Note: Readers may be interested in previous EJIL:Talk commentary on the Lisbon Judgment discussed below. For pieces by Joseph Weiler see here and here . See also Piet Eeckhout’s ” The European Union and International Law Under the Treaty of Lisbon“. For analysis of the legal guarantees offered to Ireland to secure ratification of the Lisbon treaty, see Laurent Pech’s “The European Union’s Lisbon Treaty: Some Thoughts on the Irish Legal Guarantees“.
On first reading the 2009 Lisbon case of the German Constitutional Court appears to hew quite closely to the Court’s reasoning in 1993, reviewing Germany’s accession to the Maastricht Treaty. Both cases declare that European integration must respect the inviolable core of the German Constitution (Grundgesetz). (Specifically, in these cases, Article 20, entrenching democracy and the rule of law. See Zitierung: BVerfG, 2 BvE 2/08 vom 30.6.2009, ¶233 [hereinafter Lisbon]; Decision concerning the Maastricht Treaty, 33 I.L.M. 388, 422 [hereinafter Maastricht]). In both cases the Court declares that under the Treaties it retains final say over whether European Law is compatible with the Grundgesetz and is thus applicable in Germany (judicial Kompetenz-Kompetenz).[i] Finally Lisbon, like Maastricht, finds that the Treaty ultimately passes constitutional muster. Thus, at first blush, the Court of Lisbon seems to basically restate its 1993 reasoning. I want to argue, however, that the Court has substantially sharpened its challenge since Maastricht, elevating much of the Court’s earlier state-centric interpretation of the status of integration under the Treaties to a statement of German constitutional principle.
I will focus on three ways in which Lisbon represents an advance on Maastricht. The Court announces: 1) that the Grundgesetz entrenches an absolute and unamendable limit on integration, that State sovereignty as such is inalienable, and thus forbids the delegation of excessive competences, especially Kompetenz-Kompetenz; 2) the Grundgesetz requires the German Constitutional Court to retain final review over the actions of German and European public authorities for possible alienation of, or encroachment on, German State sovereignty (judicial Kompetenz-Kompetenz); and 3) the Court goes about rigorously reviewing the Lisbon Treaty for infringements of German sovereignty in a far more searching manner than it had done in the past. Leaving little to implication, the Court spells out the consequences of its decision: in the exceptional case where European institutions overstep their enumerated powers, even with the interpretive blessing of the ECJ, the German Court will exercise review and may instruct German authorities not to apply the European law, even if it means engaging Germany’s international state responsibility.
1. Constitutional limits to integration: Germany must retain substantial competences.
The rhetoric of Lisbon suggests that, like Maastricht, it concerns a democracy review on the model of Solange. However, I want to suggest that Lisbon is really, at its core, about protecting state sovereignty in light of the expansion of competences at the Union level. In this regard, in 1993 the Court held only that under the Treaty of Maastricht, integration would not yet reach the point of a federal state. In 2009 the Court went further, holding that full integration into a supranational federal state (federalization) would be in principle forbidden by the Constitution.
The plain text of the Grundgesetz is open-textured and somewhat ambivalent on the question of limits to integration. At the same time it expresses boundless openness to integration in Article 24, while entrenching a core constitutional “identity” inviolable even by constitutional amendment under Article 79(3).
Article 24(1) provides that “the [German] Federation may by law transfer sovereign powers to international organizations,” and, according to Article 24(2), it “shall consent to such limitations upon its sovereign powers as will bring about and secure a lasting peace in Europe and among the nations of the world.” However, in Solange the Court made clear that Article 24 “cannot be taken literally”: the phrase “The Federation may by law transfer sovereign powers” must be understood to mean by formal constitutional amendment – sovereign powers may not be validly alienated through ordinary German legislation, or through the autonomous actions of the European institutions. (Solange I, ¶22). Thus integration must be accomplished through formal amendments, which are subject to Article 79.[ii] Herein lies the tension between integration and identity.
Under Article 79(3) of the Grundgesetz, the “identity” of the constitution is considered unamendable, entrenching Article 1 (Human Rights) and Article 20 (Constitutional Principles), which provides:
(1) The Federal Republic of Germany is a democratic and social federal state
(2) All state authority is derived from the people. It shall be exercised by the people through elections and other votes and through specific legislative, executive and judicial bodies.
(3) The legislature shall be bound by the constitutional order, the executive and the judiciary by law and justice
(4) [Right of resistance]
This is very open-textured language. If Article 79(3) should be understood as limiting integration, two questions must be addressed: first, precisely what rights and principles do Articles 1 and 20 protect? And second, whether Article 79(3) contain conditionally limits integration – permitting indefinite integration so long as it respects and guarantees the rights and principles enshrined in Articles 1 and 20 – or whether those Articles set absolute limits on integration? The Constitutional Court has navigated the tension between integration and identity in the Grundgesetz in a series of famous cases on European integration, yet it has only come to a principled stance in Lisbon.
The Solange cases only cautiously addressed constitutional limits on integration. The Court seemed to suggest that while in principle the Constitution sets no limits on how far integration may go, it must occur in such a way that respects and protects the constitutional identity of the Grundgesetz. Insofar as a rule of Community law conflicts with a rule of the Constitution relating to fundamental rights the Court may exercise review – not as to the validity of the Community law but as to whether the rule can “be applied by the authorities or courts of the Federal Republic of Germany.” (Solange I, ¶28). Yet while stating, on the one hand, that it would review Community legislation for compliance with basic rights under the Grundgesetz, the Court indicates that it would no longer exercise such review when the Community took responsibility for protecting those basic rights. (Solange I, ¶28). Thus the Court could foresee a full centralization of rights protection consistent with the Grundgesetz. As for Article 20, the Court hints that further integration may be constitutionally impermissible if not accompanied by democratization at the central level, but leaves open the possibility that integration could proceed to federalization if central democratic legitimation commensurate to the Community’s central powers were guaranteed.
“Article 24 of the Constitution does not without reservation allow [the inalienable aspects of the Constitution] to be subject to qualifications. In this the present state of integration of the Community is of crucial importance. The Community still lacks a democratically legitimated parliament directly elected by general suffrage which possesses legislative powers and to which the Community organs empowered to legislate are fully responsible on a political level…As long as this legal certainty…is not achieved in the course of the further integration of the Community, the reservation derived from Article 24 of the Constitution applies”. (Solange I, ¶23).
Thus, under the Court’s interpretation in Solange, Article 20 entails a conditional limit on integration – that competences may not be delegated to the center without commensurate central democratic legitimation – but not necessarily any absolute limits on the delegation of competences.
In Maastricht the Court seems to follow the Solange approach to interpreting Article 20 as a conditional limitation. Indeed the Court states that the question of absolute limits to integration does not arise and reviews Maastricht only for the delegation of competences in light of the current level of democracy at the European level. (Maastricht, 33 I.L.M. 388, 422 & 423). Although the Court finds that Maastricht passes constitutional muster, it makes clear that so long as the central democracy deficit is not addressed there are limits to just how much competence may be constitutionally delegated to the European level. (Maastricht, 33, I.L.M., 388, 395). “The important factor is that the democratic foundations upon which the Union is based are extended concurrent with integration, and that a living democracy is maintained in the Member States while integration proceeds.” (Maastricht, 33, I.L.M., 388, 421 (my emphasis)). Tellingly, the Court expressly avoids the question of just how far integration may proceed, i.e. whether Articles 1, 20, and 79 entail absolute limits. (Maastricht, 33. I.L.M., 388, 423 (finding that the question “as to whether or not the Grundgesetz allows or excludes German membership in a European State does not arise”)).
Lisbon pushes the ball with respect to the integration/identity tension in two ways. First, the Court interprets Article 20 as a robust (if hazy) limit on of transference of sovereign powers, entrenching not only democracy but also State sovereignty and the constituent power of the people as inalienable. (Lisbon, ¶247). Secondly, and more importantly, the Court explicitly states that Article 20 set an absolute limit on integration: the Grundgesetz permits the transference of limited sovereign powers to supranational entities, but it does not permit the alienation of state sovereignty as such.[iii] German authorities may not, by normal constitutional amendment, join a federal state of Europe and in consequence may not delegate too many competences to European institutions, especially not the competence to decide upon their own competence (legislative Kompetenz-Kompetenz). (Lisbon, ¶233).[iv] Maastricht held that in light of the degree of democracy at the European level in 1993, the “German Federal Parliament must retain functions and powers of substantial import.” (Maastricht, 33 I.L.M. 388, 421). Lisbon translates this conditional limitation on integration into an absolute limit. For all its language echoing Maastricht on reviewing for democracy, the Court admits in an aside that its review of European legislation according to the principle of democracy-qua-participatory/voting-rights is only a secondary constraint – a conditional limitation which sets “limits to the transfer of sovereign powers…which do not already result from the inalienability of the constituent power and of state sovereignty.” (Lisbon, ¶247).
After Lisbon, the conditional limitations on integration – respect for human rights (Solange) and democracy (Maastricht) – still stand. Yet Lisbon goes further, in announcing that the Grundgesetz entrenches state sovereignty as an absolute limit on how much competence may be transferred to Europe. Although the issue of full federalization is probably not particularly sensitive today, the question precisely how much competence may be transferred under the 27 national Constitutions is much more volatile, not least because the precise boundary may differ in each state. The obvious question, then, is who decides whether the constitutional limits have been breached: the ECJ or the national Courts?
2. Judicial Kompetenz-Kompetenz.
In Lisbon the Court finds that under the Grundgesetz the power to decide when European institutions are acting beyond their competences, judicial Kompetenz-Kompetenz, must always remain with the German Constitutional Court. Herein lies the Court’s second major advance since 1993: not only do the Treaties transfer no judicial Kompetenz-Kompetenz to Europe, (as it found in Maastricht), but any such transfer would be in principle impermissible under the Grundgesetz. In other words the Court now considers itself constitutionally bound to determine in last instance both whether the Treaties confer too many competences to Europe, and whether Europe acts beyond its enumerated competences. The Court’s stance is clear: if exercising such review means violating an ECJ judgment, let German state responsibility be engaged. (Lisbon, ¶340).
Maastricht was the Court’s initial foray into the fight on Kompetenz-Kompetenz. The Court placed great emphasis on the interpretation of Article F(3), Maastricht’s “Necessary and Proper Clause”, insisting that it could not be interpreted as giving exclusive competence for jurisdictional conflicts to the European Court. (Maastricht, 33 I.L.M 388, 428). In other words, the Court denied that F(3) gave the Court the exclusive power to determine whether the Community has or has not acted outside of its enumerated powers (judicial Kompetenz-Kompetenz). (Maastricht, 33 I.L.M. 388, 428-429). Going further, it warned the ECJ: If European institutions were to interpret and administer F(3) as delegating either legislative or judicial Kompetenz-Kompetenz, “such conduct would not be covered by the Act of Accession and would therefore not be legally binding within Germany.” (Maastricht, 33 I.L.M. 388, 428). Under this interpretation of the Treaties, national authorities could bring suit to the ECJ against the European institutions for over-extending their competences; yet, if the ECJ were to rule in favor of the European body, the national court would retain the right to declare the “valid” Community law incompatible with the Grundgesetz and inapplicable in Germany. Maastricht’s stance toward the ECJ has thus been called a classic “heads I win, tails you lose scenario.” (J.H.H. Weiler, The Autonomy of the Community Legal Order, in The Constitution of Europe, 307, 309 (1999)). Yet this stance was based exclusively on a narrow interpretation of how much has been delegated by the Treaties.
Here too, with regard to judicial Kompetenz-Kompetenz, Lisbon not only confirms the earlier jurisprudence, but extends it from a restrictive interpretation of the Treaties to a statement of German constitutional principle. In 1993, the Court stated that the Treaties had not transferred judicial Kompetenz-Kompetenz to Europe. Lisbon, in contrast, explicitly stated that Article 20 of the Grundgesetz absolutely forbids the transference of Judicial Kompetenz-Kompetenz of any kind to supra- or international bodies. (Lisbon, ¶338). The Court declares that “it must be possible for the German jurisdiction…[determine] if obvious transgressions of the boundaries take place.” (Lisbon, ¶240). The Federal Constitutional Court reviews whether the inviolable core content of the constitutional identity of the Basic Law…is respected” by the Treaties, or by secondary European law. (Id.). Likewise the Court exercises “ultra vires review” in the last instance, to determine “where Community and Union institutions transgress the boundaries of their competences.” (Id.). Finally, the German Court retains the power to review whether “the Community jurisdiction interprets the Treaties in an extensive manner that is tantamount to an inadmissible autonomous Treaty amendment.” (Lisbon, ¶338). Even if the ECJ has authority to declare a European law valid or invalid on the basis competences, The German Court states that it alone must retain the power to review in last instance whether the European law goes beyond the scope of the German “Acts of Accession” to the Treaties, or whether these Acts (themselves constitutional amendments) violate Article 79(3). (Lisbon, ¶337).
What’s more, Lisbon goes further than Maastricht in making clear the Court’s solution to encroachments on German sovereignty: the Court states frankly that “There is no contradiction to the aim of openness to international law if the legislature, exceptionally, does not comply with the law of international agreements – accepting, however, corresponding consequences in international relations” if this is the only way to avoid violating the fundamental principles of the constitution. (Lisbon, ¶340). If it finds such encroachments by Community action, the Court will bar German authorities from implementing the offending European measures even if it means engaging Germany’s state responsibility under international law.
3. A thorough review, and preemptive warnings on interpretation.
Finally, the Court goes beyond simply claiming the power of review in the last instance for encroachments on the German Constitution by primary or secondary European law – it actually reviews Lisbon Treaty competence by competence. Here too it, in terms of rigor and extent of actual review, the Lisbon Court goes beyond Maastricht.
The Court’s approach to review is to examine new and modified competences incorporated in the treaty to determine whether they violate the absolute limitations of Article 20, (and in some cases the conditional limit of democracy). The Court announces two prongs to its review. First, it examines each new or amended competence to determine whether they, jointly or severally, impermissibly reduce German sovereignty – i.e. whether the German State retains substantial sovereign powers. (Lisbon, ¶246; Maastricht, ¶4). Second, it examines whether any of these competences are sufficiently broad as to transfer legislative Kompetenz-Kompetenz to Europe – if Europe cannot be delegated unlimited powers, the Court reasons, it follows that it cannot be given the legislative power to decide autonomously the limits of its own competence. “The Basic Law does not grant the German state bodies powers to transfer sovereign powers in such a way that their exercise can independently establish other competences for the European Union. It prohibits the transfer of competence to decide on its own competence (Kompetenz-Kompetenz)”. (Lisbon, ¶233).
The Court exercises this bipartite review in several areas: Judicial cooperation in criminal matters, (¶¶352-366); judicial cooperation in civil matters (¶¶367-369); external trade relations (¶¶370-380), common defence (¶¶381-391), and with regard to social concerns (¶¶392-400). Of these, its review of the provisions on European competences in criminal law is the most interesting because the Court actually brushes up against several potential conflicts with the Grundgesetz. The Court determines, first, that on their face the new powers don’t impermissibly reduce German State Sovereignty. However, the Court expresses concern that certain aspects of the new competences in criminal law may be interpreted expansively, so as to grant European powers a legislative Kompetenz-Kompetenz, at least in the area of criminal law.[v] Notably, however, the Court does not strike anything down, because according to the Court, because the Treaty is susceptible of a narrow interpretation whereby these powers would not imply a delegation of Kompetenz-Kompetenz in Criminal Law. (Lisbon, ¶¶361-364). The German Court is satisfied, however, because European competence in criminal law remains subject to unanimity, and the Court simply instructs the German member of the Council to hew to the its favored strict interpretations. (Lisbon, ¶¶360-364). The Court states that if the Council wants to shift the area of criminal law to majority voting, through the “bridging” procedure, then the German member must get prior authorization from the German Bundestag and Bundesrat, which authorization would itself be subject to national judicial review. (Lisbon, ¶366). The Court leaves unsaid the consequences of the German member voting to extend European competences in contravention of the Court’s narrow interpretation, or if after validly “bridging” the area of criminal law to majority voting, an impermissibly expansive interpretation should be adopted over the vote of the German member. Yet the tone of the case makes it easy enough to guess – the German Court claims the power to declare inapplicable in Germany any European actions resulting from an overly expansive interpretation of the Treaty that encroaching on Article 20.
Throughout its review of the various competences incorporated in the Treaty of Lisbon, the Court does not actually declare anything constitutionally inapplicable, pure and simple. However, it frequently flexes its muscles, warning the German representatives to Europe and by implication even the ECJ to adopt narrow interpretations of specific articles of the Treaty, or risk incompatibility with the Grundgesetz.
In several respects Lisbon represents an elevation of the implications of Maastricht to the level of constitutional principle. The Court finds that under the Grundgesetz Germany cannot join a federal European State, nor can it transfer unlimited competences to the supranational level – especially not legislative Kompetenz-Kompetenz. (Lisbon, ¶233). Moreover, the Court adamantly denies that the European Courts have the judicial Kompetenz-Kompetenz to determine, exclusively and in the last instance, when the European legislative bodies have gone beyond their delegated competences. (Lisbon, ¶240). In consequence, the opinion represents a definitive statement that when push comes to shove the German Court considers the European Union to be no more than an international organization which may be ignored in order to safeguard the all-trumping domestic constitution – perhaps at the mere expense of engaging German state responsibility.
In one of the most uncomfortable passages of the case, the Court flatly states that its identity-based line of reasoning should not be unfamiliar to the ECJ. After all,
“The [ECJ] based its decision of …Kadi on a similar view according to which an objection to the claim of validity of a UNSC Resolution may be expressed citing fundamental legal principles of the community. The [ECJ] has thus, in a borderline case, placed the assertion of its own identity as a legal community above the commitment that it otherwise respects.” (Lisbon, ¶340).
The subtext is that European law, for all its supremacy and direct effect in theory, amounts at bottom only to public international legal obligations that must be flouted when the Grundgesetz demands.
The passage on Kadi is particularly uncomfortable because it reveals the sharp division between the ECJ and German Court’s visions of Europe. Lisbon’s identity-based reasoning seems to undercut Kadi’s own identity-based reasoning – the European constitutional-community rationale. In the words of the ECJ,
“…the obligations imposed by an international agreement [UNSC Res. 1267] cannot have the effect of prejudicing the constitutional principles of the EC Treaty, which include the principle that all community acts must respect fundamental rights, that respect constituting a condition of their lawfulness which is for the Court to review in the framework of the complete system of legal remedies established by the Treaty.)” (Kadi, ECJ, ¶285).
The German Court seems to suggest that ultimately the EU is at bottom no more than a community of states like the U.N., with all the binding force of state responsibility under public international law – in other words, by this view the only real atomic constitutional community is the State, and in Europe as under public international law, the Member States remain “Masters of the Treaties.” (Lisbon, ¶235).[vi]
[i] J.H.H. Weiler, The Autonomy of the Community Legal Order, in The Constitution of Europe, 312 (1999) (distinguishing between two senses of Komptenz-Kompetenz: the legislative Kompetenz-Kompetenz (“the power [of a body] to extend its own jurisdiction”) and the judicial Kompetenz-Kompetenz (“the power to be the ultimate arbiter of disputes concerning the extent of those limited competences”).
[ii] Article 79 is the normal amendment rule of the Grundgesetz. The Constitution also provides, in Article 146, that “This Basic Law, which since the achievement of the unity and freedom of Germany applies to the entire German people, shall cease to apply on the day which a constitution freely adopted by the German people takes effect.” As the Court notes in Lisbon, the limitations of 79(3) may be understood as limitations placed by the constituent power on the ability of the constituted bodies to change the constitution through amendment. In contrast, Article 146 provides that the constituent power may itself always remake the constitution as it pleases. Lisbon ¶179 (“Article 146 of the Basic Law confirms the pre-constitutional right to give oneself a constitution from which the state authority founded on the constitution emerges and by which such authority is bound.”) However Article 146 has never been used, and contains no legal procedure of any kind – its use remains an entirely open question, and its meaning entirely speculative.
[iii] The precise boundary between when transferring some “sovereign powers” turns into transferring “sovereignty as such” is of course extremely vague, and is hardly illuminated by the Court. The Court’s vagueness on this subject should not, however, be surprising in light of its interest in maintaining judicial Kompetenz-Kompetenz to decide for itself when alienation of sovereign powers turns into an impermissible alienation of sovereignty. See infra §2.
[iv] The Court does, admittedly, hint that such integration could perhaps proceed under Article 146 – access to the constituent power. Lisbon ¶217. However, as noted above, Article 146 includes no procedure and has never been utilized – its use remains entirely speculative, if not a dead letter. See supra note 2. And even then, in what is perhaps a slight overextension of its jurisdiction, the Court suggests that “it may remain open whether, due to the universal nature of dignity, freedom, and equality alone, this commitment [to guarantee Articles 1 and 20] even applies to the constituent power, i.e. for the case that the German people, in free self-determination, but in a continuity to the Basic Law’s system of rule, gives itself a new constitution.” Lisbon, ¶217.
[v] See, e.g., ¶¶359-360 (expressing concern about the expansive potential of the power to combat “particularly serious crimes” with either a “cross-border dimension resulting from the nature or impact of such offences (Art. 83.1(1) TFEU) or from a special need to combat them on a common basis (Art. 83.1(1) TFEU)”); ¶361 (expressing concern with the “threatening boundlessness” of the “annex competence which makes the approximation of criminal law possible in policy areas which have been subject to harmonisation measures (Art. 83.2 TFEU)”); and ¶363 (the possible breadth of the “general empowerment concerning the definition of criminal offences and sanctions pursuant to Art. 83.1 TFEU). All of these competences must, the Court states, “be interpreted in a correspondingly limiting fashion.” Lisbon, ¶363.
[vi] “What corresponds to the non-transferable identity of the constitution (Article 79.3 of the Basic Law), which is not amenable to integration in this respect, is …the constituent power of the Member States as the masters of the Treaties. Within the boundaries of its competences, the Federal Constitutional Court is to review, if necessary, whether these principles are adhered to.”