School of Law, Liverpool John Moores University, Liverpool, UK
The driving idea of subsidiarity is that public functions should be exercised as close as possible to the citizen. Only if the ‘closest’ authority is not in a position to perform a function or to do it effectively will this function be allocated to a ‘higher’ level of government. Subsidiarity is based on the assumption that ‘closer’ authorities are better suited to respond to certain social demands stemming from their community. Only those demands that are not limited to a given community or that require action on a wider scale or that can be better fulfilled by another authority shall be exercised by other (higher) tiers of government. Ideally, each, ‘higher’, tier of government should only perform a ‘subsidiary function’ in relation to other tiers of government ‘closer’ to the citizen.1
The Committee of the Regions’ White Paper on Multilevel Governance envisages a close link between multilevel governance and the principle of subsidiarity. According to the White Paper, subsidiarity fulfils two fundamental tasks: first, to prevent decisions from being restricted to a single tier of government and, second, to ensure that policies are decided and implemented at the most appropriate level. The White Paper further suggests that respect for subsidiarity and multilevel governance are ‘indissociable’; one indicates the responsibilities of the different tiers of government, whilst the other emphasises their interaction.2
Subsidiarity presupposes coexistence and interaction between multiple tiers of government, which, in turn, is a feature of multilevel governance (cf. supra Chap. 3). At the same time, multilevel governance requires the protection of a degree of autonomy of state and sub-state entities without which the system would become ‘centralised’, i.e., the opposite of a ‘multilevel polity’. However, the specific terms of the relationship between subsidiarity and multilevel governance need further exploration. To address this problem, it is necessary to remember that multilevel governance requires the involvement of the sub-national authorities in the EU decision-making process (‘integrating the objectives of local and regional authorities within the strategies of the European Union (…) and encourage their participation in the coordination of European policy’3). This notion will be used as a conceptual foundation to link subsidiarity and multilevel governance.
This chapter will outline the concept and application of subsidiarity in the EU (Sect. B); will analyse the concept and judicial application of subsidiarity in Germany and in Italy, where the principle plays a key role (Sect. C); will deal with the ‘procedural’ approach to subsidiarity in the Lisbon Subsidiarity Protocol (Sect. D); and, finally, will put forward a sustainable concept of subsidiarity in the light of multilevel governance (Sect. E).
It will be argued that subsidiarity is not per se ‘non-justiciable’. However ‘justiciability’ of this principle is not sustainable in the long run due to the practical and political implications of it. Accordingly, the only viable way to enforce subsidiarity is through ‘procedural’ arrangements during the lawmaking process. Subsidiarity emerges as a ‘procedural principle’ that pursues the same fundamental objective of multilevel governance to legitimise the Union’s authoritative decision-making. In this way, it can be considered a manifestation of multilevel governance.
B. The Principle of Subsidiarity in the EU
1. The Origin of Subsidiarity in the EU
The principle of subsidiarity obtained the first important recognition in Community law in the Single European Act of 1986. In the field of environment, Article 130r EEC limited Community action to the objectives that could be attained better at Community than at Member State level.4 With the Treaty of Maastricht, subsidiarity became a principle informing the entire Community action, beyond environmental law, with the exception of the areas falling within the exclusive Community competence.5
The Treaty of Lisbon enhanced the position of regional and local authorities in the EU through their explicit incorporation into the principle of subsidiarity. According to Article 5(3) TEU, ‘(…) the Union shall act only if and insofar as the objectives of the proposed action cannot be sufficiently achieved by the Member States, either at central level or at regional and local level (…)’ (emphasis added).6
This amendment met a request that federal Member States and the Committee of the Regions had repeatedly put forward to have the Treaty to reflect the reality that Union objectives can be attained not only at national but also at regional or local level.7 It would appear that the reference to the ‘regional and local level’ would imply that the Union shall be aware of the decentralised structure of the Member States. This is of great symbolic importance. However, already under the previous definition of subsidiarity it was sufficiently clear that the capacity of the Member States to attain certain policy objectives had to be determined whilst taking account of possible action by each level of government at State or sub-national level.8 The requirement for the Union legislator to bear in mind the role and the point of view of the ‘regional and local level’ is further confirmed by other provisions: for example, the Commission’s duty to consult widely and ‘take into account the regional and local dimension of the action envisaged’,9 the involvement of ‘regional parliaments with legislative powers’ in the early warning system,10 the advisory role of the CoR in relation specifically to subsidiarity.11
2. The Court of Justice’s Approach to Subsidiarity
Despite the importance attributed to subsidiarity, until now no Union act has ever been invalidated by the ECJ for a breach of this principle. It must be taken into account that only in very few cases subsidiarity has been put forward against an act and, also, that it has never been used as the first or principal plea.12 Faced with subsidiarity challenges, the ECJ has systematically upheld Union action. This jurisprudence is explained by legal scholars in different ways. According to some, the ECJ’s approach to subsidiarity is a ‘light’ or ‘low intensity judicial review’.13 For others, subsidiarity is a political or philosophical concept, therefore impossible (or very difficult) to enforce judicially.14 More specifically, according to some, this principle would be perceived by the European Court as a ‘threat to integration’,15 and for this reason its judicial enforcement would be in conflict with the ‘broad ethos’ of the Court.16 Finally, there are those who argue that it would be wrong to describe the ECJ’s approach to subsidiarity as an aprioristic refusal to engage with the complexities of the scrutiny; in reality, in all relevant cases landed before the ECJ, Union legislative intervention was perfectly lawful.17
Certainly, it is not true that the ECJ systematically accepts the position of the Union legislator without further scrutiny. This was probably the case in UK v Council, concerning the Working Time Directive, where the ECJ appeared to accept the point of view of the Council without an independent analysis of the subsidiarity question. According to the circular argument used by the ECJ on that occasion, if the Council found that it is necessary to achieve a certain objective ‘to improve the existing level of protection as regards the health and safety of workers and to harmonize the conditions in this area’, then ‘achievement of that objective (…) necessarily presupposes Community-wide action’.18
By contrast, in other cases, the Court developed a more substantial reasoning in support of Union action. For example, in Netherlands v Council, concerning Directive 94/44/EC, albeit with a rather concise explanation, the Court held that the Directive’s objective to harmonise legislation and practice in the area of protection of biotechnological inventions ‘could not be achieved by action taken by the Member States alone’. In addition, according to the Court, as the scope of the protection had immediate effects on intra-Community trade, ‘it is clear that, given the scale and effects of the proposed action, the objective in question could be better achieved by the Community’.19
The same argument, that harmonisation of laws is required in order to achieve common market objectives, arose also in later rulings. For example, in British American Tobacco, the ECJ held that ‘the Directive’s objective to eliminate the barriers raised by the differences which still exist between the Member States’ laws (…) on the manufacture, presentation and sale of tobacco products’ could not be sufficiently achieved by the Member States individually and called for action at Community level. ‘It follows’, according to the Court, that ‘the objective of the proposed action could be better achieved at Community level’.20
Also in Vodafone, concerning EC Regulation No. 717/2007 on the ‘Eurotariff’ for roaming services, the Court engaged in a thorough analysis of the challenged act before accepting the explanation contained in the Regulation’s Preamble,21 i.e., that the interdependence between wholesale and retail roaming charges renders the choice to impose a ‘ceiling’ on both compliant with subsidiarity. Interestingly, the Court’s reasoning in this case is similar to the German judicial doctrine of the Annexkompetenzen, according to which the Federation can extend its lawmaking activity to issues that are a necessary ‘appendix’ to the principal area of regulation. The regulation of the ‘appendix’ must be logical, necessary, corollary of the regulation of the main area.22
The view depicting the ECJ’s approach to subsidiarity as relaxed must be refuted also for another reason. The ECJ’s judgments take into account the opinions of the Advocates General. In relation to subsidiarity, in no circumstance the Court came to a conclusion different from that envisaged by an Advocate General. Whilst admittedly some Advocates General’s opinions entail a surface scrutiny of subsidiarity,23 others are quite thorough when tackling the same issue.24
3. The Proportionality Test
Article 5 TEU allows for Union action not only ‘if’, but also ‘insofar as’, the objectives of the proposed action cannot be sufficiently achieved by the Member States and can be better achieved at Union level. The expression ‘insofar as’ refers to the proportionality requirement, according to which all Union action should not go beyond what is ‘appropriate’ and ‘necessary’ to achieve the proposed objective.25 More specifically, proportionality demands that Union action be kept to the minimum necessary, i.e., as specified in Article 5(4) TEU, ‘… the content and form of Union action shall not exceed what is necessary to achieve the objectives of the Treaties’. Accordingly, the Amsterdam Subsidiarity Protocol required Union institutions to leave as much scope for national decision as possible, to prefer directives to regulations and framework directives to detailed measures and to minimise the burden, financial or administrative, of Union measures for, inter alia, national governments and local authorities.26
Some scholars argue that Member States’ autonomy could be better protected through a judicial review of proportionality of Union action rather than through the traditional subsidiarity review.27 This opinion pursues the noble aim of protecting Member States’ autonomy through judicial review. However, it relies on the undemonstrated assumption that a focus on proportionality would favour Member States’ autonomy.28 In reality, there is no evidence that, by focusing on proportionality, the ECJ would be more likely to find in favour of the Member States.29
For example, in UK v Council, in relation to the Working Time Directive, the Court held that in order to establish whether a provision of Community law is compliant with proportionality, it must be ascertained ‘whether the means which it employs are suitable for the purpose of achieving the desired objective’ and ‘whether they do not go beyond what is necessary to achieve it’.30 In this case, the Court pointed out that judicial review of the exercise of discretionary powers of Community lawmaking institutions must be limited to examining ‘whether [such exercise] has been vitiated by manifest error’ or ‘misuse of powers’ or ‘whether the institution concerned has manifestly exceeded the limits of its discretion’.31 After comparing the content of the Working Time Directive with its objective (i.e., improving health and safety protection for workers), the Court concluded that ‘(…) the measures on the organization of working time (…) cannot (…) be regarded as unsuited to the purpose of achieving the objective pursued’.32 Furthermore, the Court found that ‘the Council did not commit any manifest error in concluding that the contested measures were necessary to achieve the objective’33 or ‘in taking the view that the objective of harmonizing national legislation on the health and safety of workers (…) could not be achieved by measures less restrictive’.34 What seems to count, for the Court, is not whether an action is really ‘necessary’, or whether, in theory, less restrictive measures are ‘possible’, but whether there is a ‘manifest error’ vitiating the discretion of the Union legislature.35 The described approach does not promise more rigour than the typical subsidiarity reasoning.
This conclusion seems confirmed by the analysis of other cases. In British American Tobacco, for example, the boundary between subsidiarity and proportionality is blurred where the Court, judging on subsidiarity, made reference to the paragraphs of the same judgment examining and rejecting the plea on proportionality: ‘(…) the intensity of the action undertaken by the Community in this instance [as well as being in keeping with the requirements of proportionality] was also in keeping with the requirements of the principle of subsidiarity in that (…) it did not go beyond what was necessary to achieve the objective pursued’.36 Accordingly, it appears unlikely that the Court may adopt a more rigorous approach to proportionality than to subsidiarity.
Another, greater, mistake in seeing proportionality as a possible panacea is that this view considers judicial enforcement as absolutely necessary for the proper functioning of subsidiarity. In reality, multilevel governance, by emphasising multilevel ‘participation’ and ‘cooperation’, pushes towards ‘procedural subsidiarity’, i.e., political consultation and negotiation rather than court action.37 However, before explaining in further detail my understanding of subsidiarity in the light of multilevel governance, it is crucial to determine whether, to what extent and how subsidiarity can be judicially enforceable. This problem will be examined in relation to two national systems in which the principle plays a fundamental role: Germany (cf. infra Sects. C.1 and C.2) and Italy (cf. infra Sects. C.3 and C.4).
C. Comparative Analysis
1. Germany: The Long and Winding Road to the Judicial Enforcement of Subsidiarity
The introduction of a subsidiarity clause in Germany is the result of a vivid debate that took place during the approval of the Grundgesetz (acronym GG), the constitution of the Federal Republic of Germany.38 Article 34 of the draft constitution approved on August 1948 by the Convention of Herrenchiemsee, a conference of regional delegates endowed with the task of preparing a constitutional draft, stipulated that ‘In the fields [in which the Federation has] the right to pass legislation with primacy [Vorranggesetzgebung] … The Federation shall only regulate what must be uniformly regulated’.39 Article 34 of the draft constitution was strongly criticised by the Allied Military Governors, who wished to see the reinforcement of the Länder vis-à-vis the central government. In their Memorandum of 2 March 1949, they proposed that ‘the Länder shall retain the right to legislate in the fields hereinafter enumerated except where it is clearly impossible for a single Land to enact effective legislation or where the legislation if enacted would be detrimental to the rights or interests of other Länder; in such cases […] the Federation shall have the right to enact such legislation as may be necessary or appropriate […]’.40 In the view of the occupying authorities, a more robust vertical division of powers between the Federation and the Länder would reduce the risk for Germany to fall again into totalitarianism.41 At the same time, a weaker central government would prevent the reconstituted German state from becoming a threat for its neighbours. Following the Memorandum, the Parliamentary Council in Bonn (the body entrusted with the task of passing the constitution42) replaced the ‘right [of the Federation] to legislate with primacy’ with the ‘concurrent legislative power’ (konkurrierende Gesetzgebung) of Article 72 GG:
(1) On the matters falling within the concurrent legislative power the Länder shall have the right to legislate so long as and to the extent that the Federation has not used its legislative power by enacting a law.
(2) The Federation has the right to legislate in this field to the extent that there is a need for regulation through federal law[.]
[Such need arises when]:
1. An issue cannot be effectively regulated by the legislation of a single Land, or
2. The regulation of an issue by a single Land through a Land law may compromise the interests of other Länder or of the whole [State], or
3. It is required for the creation of legal or economic unity and especially for the maintenance of uniform living conditions within the federal territory beyond the territory of a single Land.43
The Bedürfnisklausel (need clause) of Article 72(2) GG incorporated quite clearly the idea of subsidiarity. Certainly, in comparison with the original Herrenchiemsee proposal, it set stricter conditions to the intervention of the federal legislator. From the wording of the ‘need clause’, it is apparent that in the fields of concurrent legislation (including key areas such as ‘civil law’, ‘criminal law’, ‘economic law’, ‘employment/labour law’44), the Federation had only a ‘subsidiary’ right to legislate; i.e. it could only legislate if certain issues could not be effectively regulated by the Länder and, consequently, needed to be regulated by the Federation. However, despite the clear wording and purpose of Article 72(2) GG, the Federal Constitutional Court consistently construed it to a large extent as a non-justiciable provision. Since its first judgment on the matter (Ruling of 30 April 195245), the Court considered the assessment of whether a legislative intervention was ‘needed’ as a ‘non justiciable question falling within the discretion of the legislator’ (nicht-justiziable Frage des gesetzgeberischen Ermessens). The only exception envisaged by the Court was the ‘abuse of discretion’ by the legislator (Ermessensmißbrauch),46 provided that, as specified in a later case, the infringement of the ‘need clause’ was ‘unequivocal and manifest’ (eindeutig und evident).47 The self-restraint (or ‘light touch’ approach) of the Court on this matter emasculated the ‘need clause’ and rendered it unable to limit the legislative intervention of the Federation.48 As a result, the Federal Republic of Germany could be defined as a ‘unitary federal state’ (unitarische Bundestaat), i.e., a centralised federation.49
The position of the Federal Constitutional Court can be better understood if one considers that the ‘need clause’, having been imposed by the occupying powers during the post-war military occupation, had no roots in German constitutional history.50 Probably for this reason the Court chose to construe the ‘need clause’ in terms similar to the Bedarfsgesetzgebung (need legislation) of Article 9 of the Weimar Constitution (‘If there is a need for passing uniform regulations, the Reich [the central government] has the right to legislate […]’).51 During the Republic of Weimar (1919–1933), the question of whether there was such ‘need for passing uniform regulations’ was generally considered as a non-justiciable matter falling entirely within the discretion of the legislator. Another important element to better understand the Court’s approach is the participation of the Länder in the federal legislative process through the Bundesrat, the federal lawmaking body in which the Länder are represented. Where a federal law has received the approval of the Bundesrat, it becomes politically difficult for the Court to annul that law with the argument that it is not ‘needed’ and, therefore, it is in breach of the prerogatives of the Länder. This point is clearly sketched out in the Ruling of 30 April 1952, in which the Federal Constitutional Court held that the approval of the contested federal law by the Bundesrat revealed that the majority of Länder had approved a federal regulation in that field.52
With a view to limiting federal legislative action and to overcoming the lack of judicial enforcement of the ‘need clause’, in 1994 a constitutional amendment modified Article 72(2) GG and set more compelling conditions for federal intervention in the areas subject to concurrent legislation.53 The amendment replaced the ‘need clause’ with a more rigorous ‘necessity clause’ (Erforderlichkeitsklausel):
The Federation shall have the right to legislate in this field if and to the extent that the federal legislative regulation is necessary [A] for the creation of equivalent standards of living within the federal territory or [B] for the maintenance of legal or [C] economic unity in the interest of the whole state.54
The term ‘necessary’ translates the German erforderlich. The meaning of erforderlich is deemed to be more stringent than that of Bedürfnis (need) of the previous ‘need clause’.55
Eight years after the constitutional amendment, in the Ruling of 24 October 2002 on the Geriatric Nursing Act (Altenpflegegesetz), for the first time the Federal Constitutional Court held the ‘necessity clause’ to be justiciable.56 In the ruling, the following statement is key: ‘In the constitutional scrutiny of a federal law under Art. 72(2) GG the following aspects shall be covered […]: the prognosis [made by the federal legislator in relation to the existence of the conditions for federal intervention] must rely on factual assumptions, which are carefully determined or which at least can find confirmation in the context of a court scrutiny […]. The prognosis must be supported by an appropriate prognostic method and this [method] must have been followed consistently (in the sense of the “reliability” of the prognosis […]). The result of the prognosis has to be controlled [by the Court] to determine whether the fundamental aspects of the prognostic assessment have been disclosed with sufficient clarity or at least their disclosure is possible in the procedure for the control [of the constitutionality] of legal rules [before the Court] and whether any irrelevant considerations have been incorporated into the prognosis’.57 In even clearer terms in the later Ruling of 27 July 2004 (case Juniorprofessur, junior professorship), the Court held that ‘In order to determine whether the justification grounds of Art. 72(2) GG are fulfilled, the legislator has a margin of appreciation. The decision of the legislator, which shall be assessed in each specific area in the context of an overall view, can however be scrutinised with regard to its methodological foundations and its conclusiveness by the Federal Constitutional Court’.58 These two excerpts highlight that the Court sees the scrutiny in accordance with Article 72(2) GG as an examination of the rationality of the law, i.e., as a scrutiny aimed to determine whether the legislator exercised its discretion properly, in a methodologically appropriate way (‘the prognosis must rely on factual assumptions, which are carefully determined’; ‘The prognosis must be supported by an appropriate prognostic method’; ‘this [method] must have been followed consistently’; ‘[it needs to be checked] whether any irrelevant considerations have been incorporated into the prognosis’; ‘The decision of the legislator […] can […] be scrutinised from the point of view of its methodological foundations and of its conclusiveness’).
In the Ruling on the Geriatric Nursing Act, the Court outlined its interpretation of the criteria established in Article 72(2) GG. In this way, it laid down a general theory of the constitutional scrutiny on subsidiarity. In relation to ‘the creation of equivalent standards of living within the federal territory’, the Court explained that this requirement is not fulfilled where, in addition to uniformity, there is no other added value in a federal legislative intervention. Likewise, a mere improvement of the existing standards of living is not enough in order to justify a federal intervention, which can be allowed only ‘where the standards of living in the Länder of the Federal Republic have developed in significantly different ways, in this way jeopardizing the social cohesion of the Federal Republic, or such a development is concretely emerging’.59
In relation to the ‘maintenance of legal unity’, the Court is of the opinion that legal differences shall be acceptable in a federation and, accordingly, legal unity, as such, would not be jeopardised simply by varying laws across the Länder. A federal law is required only where there is a problematic fragmentation of the regulatory framework and, accordingly, a situation that clashes with the interests of both the Federation and the Länder.60 In relation to the ‘maintenance of economic unity’, the Court argued that a federal intervention must ensure the correct functioning of the economic space of the Federal Republic. The enactment of federal laws is required only where different regional regulations or inaction by the Länder would cause major disadvantages to the national economy.61
The judgment on the Altenpflegegesetz paved a way to a new course in the jurisprudence of the Court in relation to Article 72(2) GG. In the Ruling of 16 March 2004, the Court held the Law on the Prevention of Vicious Dogs (Gesetz zur Bekämpfung gefährlicher Hunde) in conflict with Article 72(2) GG. This federal law introduced penalties for breaches of the regional laws on vicious dogs. However, given that the regional laws were not uniform (more specifically, the notion of ‘vicious dog’ and the outlawed behaviours were not identical in all the Länder), the federal law failed to produce a uniform system of penalties at national level. Accordingly, the Court concluded that the federal legislative intervention did not meet the ‘legal unity’ and the other criteria of Article 72(2) GG (‘equivalent standards of living’, ‘economic unity’).62
In the later Ruling of 9 June 2004 on the Shops Trading Hours Act (Ladenschlussgesetz), the Court held a federal law deciding the trading hours of shops on Saturdays and Sundays non-compliant with Article 72(2) GG. The Court found that the ‘equivalent standards of living’, along with the ‘economic’ and ‘legal unity’ of the Federal Republic, would not be jeopardised if shops were to have different opening times in the different Länder.63
In the Ruling of 27 July 2004 (case Juniorprofessur, junior professorship), the Court held the federal law modifying the recruitment of university professors to be in breach of Article 72(2) GG. The contested law allowed for the appointment to ‘junior professorships’ of candidates with a doctorate (Promotion) but with no Habilitation, i.e., the second doctorate normally required for the bestowing of a chair at German universities. In relation to ‘legal unity’, the Court observed that different regional regulations are a normal feature in a federal system. A uniform regulation is required only where different legal regimes for the same facts of life may generate substantial legal uncertainties and, in this manner, unreasonable obstacles to legal relations across the country (i.e., beyond Länder borders). In the view of the Court, the contested federal law had no justification under Article 72(2) GG, both from the perspective of the ‘legal unity’ and of the ‘equivalent standards of living’. The federal law did not appear ‘necessary’ to overcome a ‘risk situation’ (Gefahrenlage) arising from legal fragmentation or to prevent the living conditions among the Länder from diverging in an unsustainable way or to remove the obstacles preventing an academic worker from moving from one Land to another or making such move more difficult. The Court also found that the federal law did not contribute to the maintenance of ‘economic unity’, given that its objective was to improve the quality of university teaching and/or research, whilst the ‘economic unity’ of the Federal Republic remained relegated to the background.64
In the Ruling of 26 January 2005 (case Studiengebühren, tuition fees) concerning the federal law prohibiting the introduction by the Länder of tuition fees for first degrees, the Court found the law to be unjustified under Article 72(2) GG. The Court held that there was no sufficient evidence that the prohibition of tuition fees was ‘necessary’ for the creation of ‘equivalent living conditions’ in the territory of the Federal Republic. The Court also held that the fee element (€500 per semester) would be probably outweighed by other elements in the students’ evaluation, including variations in the living costs across the country. Admittedly, according to the Court, some people (those who are least well off) may be particularly affected by the introduction of tuition fees, yet ‘The non quantified possibility of such cases does not justify, at least currently, an intervention by the federal legislator for establishing equal living conditions in accordance with Art. 72, para. 2, GG’. In relation to ‘economic unity’, the Court took the view that different fee arrangements in the Länder would not jeopardise the single economic area. Quite the opposite, by levying the fees, the Länder could improve the educational offer and make a valuable contribution to the national economy.65
So far, the federal law on tuition fees is the last federal law to have been found by the Court in breach of Article 72(2) GG. In later cases concerning the ‘necessity clause’, the Court always rejected the claims against the contested federal acts.66 It is interesting to note the Ruling 14 October 2008 concerning the federal law implementing EC Regulation No. 1782/2003. The Court held the contested federal law to be justified against Article 72(2) GG since the correct implementation of the regulation required the enactment of uniform legal provisions at federal level. Accordingly, the ‘necessity’ of a federal law found justification, albeit indirectly, in EU law.67
In 2006, an important constitutional reform came into effect (the Föderalismusreform, ‘reform of the federal system’68). An aspect of the reform concerned specifically the ‘necessity clause’. The wording of the clause was not modified, but its sphere of application was reduced dramatically. The ‘necessity clause’ does no longer apply to legislative activity in areas of key importance, including civil and criminal laws, employment/labour law, land law, but only to a limited number of, according to some, ‘randomly selected’ areas of secondary importance.69 Most areas previously subjected to the ‘necessity clause’ fall now within the ‘core’ lawmaking power of the Federation (for example, ‘civil law’, ‘criminal law’) or, more rarely, within the lawmaking power of the Länder (for example, ‘shops’ trading hours’, ‘university’). The intended, and fully achieved, result of the 2006 reform was a dramatic fall in the number of cases concerning the ‘necessity clause’. It would appear that the proactive approach of the Court in the period 2002–2006 had generated uncertainty and dissatisfaction by the federal legislator. The 2006 reform addressed this concern by taking power away from the Court in relation to the enforcement of subsidiarity and by returning it to the political players at federal and local levels, who are free to decide ‘whether’, ‘when’ and ‘how’ to regulate a given matter in conformity with the Basic Law.70
2. Lessons from the German Pattern
A few lessons can be learnt by contrasting the German experience with the jurisprudence of the ECJ on subsidiarity. In Germany, since 2002, the Federal Constitutional Court has recognised the ‘necessity clause’ of Article 72(2) GG as justiciable. Between 2004 and 2005, the Court invalidated four federal laws for nonconformity with that constitutional provision. What are the main differences with the EU?
The ‘necessity clause’ is much more specific and compelling than the EU definition of subsidiarity, even though the Amsterdam Subsidiarity Protocol had tried to specify the notion of subsidiarity. The ‘necessity clause’ does not contain a generic formula (‘if an objective cannot be sufficiently achieved’, ‘if an objective can be better achieved’) but contains more specific criteria (‘legal unity’, ‘economic unity’, ‘equivalent standards of living’). However, it must be highlighted that, whilst this aspect (specificity) may have contributed to the justiciability of the ‘necessity clause’, for a few years the ‘necessity clause’ remained only on paper (1994–2002). Therefore, the judicial enforcement of the ‘necessity clause’ seems first and foremost due to a change in the approach by the Court. A similar change is yet to occur for the ECJ.
In the case Altenpflegegesetz, the Federal Constitutional Court laid down in sufficiently clear terms its interpretation of the criteria of Article 72(2) GG. In this way, it further specified the content of the ‘necessity clause’. So far, the ECJ has failed to develop a similar ‘doctrine’ in relation to subsidiarity.
The Federal Constitutional Court understands the scrutiny required by Article 72(2) GG as an evaluation of the rationality of the law, i.e., of whether the federal legislator exercised its discretion correctly. The Court also specified the criteria of such scrutiny (‘the prognosis must rely on factual assumptions, which are carefully determined’; ‘The prognosis must be supported by an appropriate prognostic method’; ‘this [method] must have been followed consistently’; ‘[it needs to be checked] whether any irrelevant considerations have been incorporated into the prognosis’; ‘The decision of the legislature […] can […] be scrutinised from the point of view of its methodological foundations and of its conclusiveness’).71 By contrast, the ECJ has never been as specific in relation to subsidiarity. For example, in the Working Time Directive case, the ECJ stated that judicial review of the exercise of legislative discretion must be limited to ‘whether [such exercise] has been vitiated by manifest error’ or ‘misuse of powers’ or ‘whether the institution concerned has manifestly exceeded the limits of its discretion’.72 It is interesting to note a similarity of language between the ECJ and the Federal Constitutional Court prior to the Altenpflegegesetz case, when the Court required (albeit with reference to the old ‘need clause’) an ‘unequivocal and manifest abuse of discretion’ by the legislator.73
In summary, the opinion that subsidiarity is intrinsically unsuited for judicial enforcement emerges defeated from the analysis of the German case law. What also clearly emerges is that, albeit engaging with subsidiarity, the ECJ has not equipped itself with an armamentarium comparable to the Altenpflegegesetz doctrine of the German Court. The impression is that the ECJ is anchored to judicial self-restraint. However, the interventionist approach of the German Court caused a reaction from political players, in that it limited too powerfully the discretion of the federal legislator. The 2006 constitutional reform confined the ‘necessity clause’ to a small ‘reservation’. It would appear that even a consolidated multilevel system like the Federal Republic of Germany could hardly cope with judicially enforceable subsidiarity. This explains why the legislator has chosen to return to less controversial constitutional pathways for the allocation of lawmaking authority (Table 4.1).
Summary of the comparison between Germany and the EU in relation to subsidiarity
Subsidiarity in Germany, Federal Constitutional Court
Subsidiarity in the EU, ECJ
Since 2002 justiciable, four federal laws invalidated (2004–2005); historically, Art. 72(2) GG has not prevented centralisation of powers
Theoretically justiciable, no act invalidated so far, even though ECJ does not passively marry the legislator’s view
Slightly more specific criteria in Art. 72(2) GG, ‘necessity clause’
Generic definition of subsidiarity in Art. 5 TEU (however, some specification in Amsterdam Subsidiarity Protocol)
Further specification of criteria by the Court (cf. Altenpflegegesetz ruling); in later cases, the Court has applied the Altenpflegegesetz pattern consistently
No further specification of subsidiarity clause by the ECJ
Scrutiny of rationality of the law, specific criteria for scrutiny determined by the Court
Scrutiny of exercise of discretion (cf. Working Time Directive case), but judicial self-restraint (manifest error, misuse of powers, etc.) similar to Germany under the ‘need clause’, unlikely declaration of invalidity of an act
3. Italy: From Subsidiarity to Loyal Cooperation
(i) Subsidiarity in Administration
Despite the clear reference to subsidiarity in the European Charter of Local Self-Government (1985) (‘Public responsibilities shall generally be exercised, in preference, by those authorities which are closest to the citizen. Allocation of responsibility to another authority should weigh up the extent and nature of the task and requirements of efficiency and economy’74), until 1997, the impact of the principle of subsidiarity in Italy remained limited.75
Law No. 59 of 15 March 1997 (known as ‘legge Bassanini’, ‘Bassanini Law’, named after Franco Bassanini, the then Minister of the Public Function) initiated an extensive transfer of administrative tasks and functions from the central government to regional and local authorities. This process is often referred to in Italian legal literature as ‘administrative federalism’.76 The ‘Bassanini Law’ and the following Legislative Decree No. 112 of 31 March 1998 (known as ‘Decree on Administrative Federalism’) started the transfer to the regions and the local authorities of all administrative tasks, with the sole exception of those expressly reserved for the central government (for example, ‘foreign affairs’, ‘defence’, ‘scientific research’77). The ‘Bassanini Law’ enumerates the areas belonging to the exclusive responsibility of the central government and allows for the transfer of powers to the regions and the local authorities in all the remaining areas. This technique (‘enumerated powers’ to the central government and ‘residuary powers’ to the sub-state authorities) is similar to the typical method for the distribution of powers in federal systems and explains why the reform is indicated as ‘administrative federalism’.78
In conformity with the subsidiarity principle, the objective of the ‘Bassanini Law’ is to bring administrative tasks and functions closer to the citizen. The ‘Bassanini Law’ stipulates that the ‘generality’ of administrative tasks and functions must be transferred to local authorities and that only those that cannot be effectively exercised by local authorities shall not be subject to transfer.79 However, after 17 years, the system outlined by the ‘Bassanini Law’ has yet to be fully implemented.80
Four years after the ‘Bassanini Law’, a constitutional amendment81 specifically entrenched the principle of subsidiarity in the Constitution as the basic rule, together with the principles of differentiation and adequacy, for the allocation of administrative responsibilities between the different tiers of government:
Administrative functions are attributed to the municipalities, unless they are attributed to the provinces, the metropolitan cities, the regions or to the State with the aim to ensure their uniform exercise pursuant to the principles of subsidiarity, differentiation and adequacy.82
According to this constitutional provision, subsidiarity must be considered jointly with differentiation and adequacy. Accordingly, the allocation of an administrative function to a local authority depends also on its capacity for performing that function (adequacy). Compliance with the Constitution requires something more than mere ‘proximity’ to the citizen; it requires that an authority must be capable of achieving a certain result.83 The lack of ‘adequacy’ may even determine the exercise of a competence, by way of ‘subsidium’, by a ‘lower’ tier of government, in case of failure to act by a ‘higher’ authority. For example, in 2005, an administrative court found that subsidiarity allowed a Municipality to issue a bylaw on the use of organic fertilisers, given that the Province (normally responsible for that matter) had failed to take action.84
Local authorities are far from being a homogeneous group. Ideally, a ‘good’ allocation of administrative functions would require a ‘case-by-case’ approach. This is the role assigned to ‘differentiation’, which implies that the State or the Regions, when allocating administrative responsibilities, have to take into account the characteristics of the recipient authorities. Due consideration must be given to aspects such as population and territorial size, type of organisation, performance of administrative activities in association with other authorities. This is to ensure that an administrative function is allocated to an authority that is capable of performing it in the most efficient and economic way.85
(ii) Subsidiarity in Legislation
Until the 2001 constitutional amendment,86 the Italian Constitution gave the Regions the power to pass legislation only in a few enumerated subject matters (for example, ‘fairs and markets’, ‘town planning’, ‘tourism’, ‘quarries and peat bogs’, ‘hunting’, ‘agriculture’, etc.). Laws passed by the Regions in these areas had to be consistent with fundamental principles established by State law and had to comply with the national interest and the interests of the other Regions. This type of regional legislative power was known as ‘shared’ or ‘concurrent’ legislation (‘legislazione ripartita’ or ‘concorrente’). All non-enumerated legislative powers belonged to the State.
Until 2001, the Italian Constitution followed the ‘principle of parallelism’ (principio del parallelismo) for the allocation of administrative responsibilities to the State and the Regions. According to this principle, administrative functions and legislative powers were ‘parallel’. In principle, the Regions had the right to perform administrative tasks in the same subject matters in which they enjoyed legislative power (for example, ‘town planning’ or ‘agriculture’). However, despite this clear theoretical framework, the central government retained the power to interfere with the exercise of a number of regional administrative functions. Just to mention only two examples, the ‘permission to release a licence for opening a travel agency’, in the area of ‘tourism’, and ‘agricultural market regulation’, in the area of ‘agriculture’, were reserved to the central government.87
The 2001 constitutional amendment changed the Italian regional system fundamentally. In the new regime, the State retains exclusive legislative powers only in those subject matters enumerated by the Constitution (enumerated powers),88 whilst the Regions have the right to pass laws in all areas not enumerated by the Constitution (residuary powers).89 In addition, the Constitution lays down a list of subject matters in which the legislative power is ‘shared’ by the State and the Regions (concurrent legislative power). In these areas, the State dictates ‘fundamental principles’, i.e., it shall lay down the regulatory framework, whilst the Regions shall implement the principles through a detailed regulation.90
A second ‘revolution’ was introduced by the landmark Ruling No. 303 of 1 October 2003. In this case, the Court held that the principle of subsidiarity may justify ‘a departure from the normal distribution of [legislative] powers’. The Court stated that this principle (like the ‘konkurrierende Gesetzgebung’ in Germany or the ‘Supremacy Clause’ in the United States) brings an ‘element of flexibility’ into what would otherwise be a too rigid distribution of powers. According to the Court, a State law can assign administrative tasks to the central government in areas belonging to the legislative power of the Regions and can also regulate the exercise of these tasks. A State law can do so provided that certain criteria are fulfilled: there must be a need for uniform action by the central government (principle of subsidiarity and principle of adequacy), the evaluation of the public interest underlying the allocation of regional responsibilities to the central government must be ‘proportionate’ (principle of proportionality), such evaluation must not be ‘unreasonable’ in the light of a ‘strict scrutiny’ (reasonableness) and, finally, the State law allocating an administrative function to the central government must provide for the involvement of the Regions in the exercise of that function in the form of an ‘agreement’ (intesa) between the Regions and the central government (principle of loyal cooperation). In summary, the key finding of Ruling No. 303 of 1 October 2003 is that the State, in addition to the areas laid down in the Constitution, can also legislate in other areas, if the principle of subsidiarity requires that a specific responsibility has to be exercised by the centre. This construction of subsidiarity paved the way to a (re-)centralisation of lawmaking power rather than to the protection of regional autonomy in conformity with the idea of ‘proximity’ to the citizen.91
In the wake of this landmark judgment, in a number of rulings, the Constitutional Court upheld the existence of State ‘subsidiary lawmaking powers’. In a first group of cases, the Court adopted a notion of subsidiarity similar to that emerging from the European Treaty (even though the Court has never made an open reference to the Treaty). In these cases, the reasoning of the Court is the following: an action of the central government is required, and therefore justified, where an objective cannot be sufficiently achieved by lower levels of government. For example, in Ruling No. 6 of 13 January 2004, the Court held that the central government can authorise the construction of new power plants and the enlargement of existing ones in order to satisfy the national demand of energy. This result, in the view of the Court, can be achieved only by the centre since the single Regions are not in a position to determine the national energy demand, nor can they cover it in full on their own.92
In a second group of cases, the reasoning of the Court in support of the need for uniform State action and the logic underpinning its judgment are reminiscent of the ‘legal or economic unity’ or of the ‘equivalent standards of living’ criteria of Article 72(2) GG (cf. supra Sect. C.1). For example, in Ruling No. 151 of 12 April 2005, the Court held that the payment of an €150 State subvention for the purchase or the hire of a decoder is a task that needs to be performed by the State. According to the Court, this action is actually aimed to implement a constitutional principle, ‘pluralism of information’, and it must be carried out uniformly within the national territory.93 More similar to the ‘equivalent standards of living’ approach is the reasoning displayed in Ruling No. 166 of 23 May 2008 concerning the national plan for social housing, where the Court held that a national plan is required in order to prevent strong regional differences in relation to the standards of social housing.94
In a third group of cases, the Court embraced a purely ‘procedural’ concept of subsidiarity. When following ‘procedural subsidiarity’, the Court does not question, or more precisely it takes as a given, the underpinning need for uniform State action. The Court only examines whether some procedural requirements are fulfilled (for example, the inclusion of an infrastructure in a national plan approved by the State-Regions Conference and/or the approval of the infrastructure by the territorially competent Region). If such procedural criteria are fulfilled, the Court accepts that subsidiarity has been respected. All the rulings within this group concern the regulation and/or authorisation and/or funding of infrastructures of national relevance. For example, in Ruling No. 303 of 1 October 2003, regarding large-scale infrastructure projects, the Court held that, when reviewing legislation, it is not part of the Court’s role to decide whether a given infrastructure is ‘strategic’ or ‘of preeminent national interest’. The Court attached an exclusively procedural meaning to subsidiarity. It required only that the law had to provide the involvement of the Regions in the form of an ‘agreement’ on the classification of an infrastructure as ‘strategic’ or ‘of preeminent national interest’.95 A further example of this approach can be seen in Ruling No. 79 of 11 March 2011 concerning Parma subway development. The Court held that, given the consent of the relevant Region (Emilia–Romagna), the inclusion of Parma subway in the national Plan of Strategic Infrastructures justified the exercise by the State of legislative and administrative powers relating to this infrastructure.96
In a smaller number of cases, the Court refused to grant ‘subsidiary lawmaking powers’ to the State. In some rulings, the Court found that there existed no need for uniform action by the central government. For example, in Ruling No. 219 of 8 June 2005 on the ‘lavori socialmente utili’ (socially useful jobs97), the Court found that administrative tasks concerning those jobs have an exclusively local dimension and, accordingly, should be handled by the Municipalities. Another interesting example is Ruling No. 148 of 7 June 2012, where the Court held that the need to tackle the economic recession, alone, did not justify a temporary exception to the distribution of responsibilities between the central government and the Regions.98
In another group of cases involving the principle of subsidiarity, the Court carried out the proportionality test.99 From the case law it emerges that when testing proportionality, the Court will look at the ‘breadth’ and/or the ‘intensity’ of regulation. For example, in Ruling No. 214 of 1 June 2006, the Court focused on the range of tasks assigned to a certain public body (breadth). The Court held that the law establishing the National Committee for Tourism had gone beyond what is strictly necessary for the promotion of tourism. The law in question, instead of specifying and delimiting the remit of the Committee, had entrusted it with an all-embracing activity of policy coordination for the entire touristic sector. Accordingly, in the Court’s view, the scale of the State intervention was disproportionate.100 When checking proportionality, the Court may also look at the ‘intensity’ of regulation. For example, in Ruling No. 166 of 23 May 2008, it held the national plan on social housing compliant with the principle because the plan consisted of general guidelines leaving sufficient scope for regional implementation.
(iii) The Relationship Between Subsidiarity and Loyal Cooperation
In Italy, unlike in Germany, there is no national lawmaking body that represents the Regions in the legislative process. Therefore, since the landmark Ruling No. 303 of 1 October 2003, the Constitutional Court has maintained that any State law taking a function away from the Regions and allocating it to the central government in accordance with subsidiarity must provide adequate cooperation mechanisms. This means that the Regions need to be involved in the exercise of that function. In this way, by its jurisprudence on subsidiarity, the Court clearly enforces a ‘procedural’ and ‘cooperative’ pattern.101
Every time a ‘subsidiary power’ of the central government entails a decision on policy, or a decision affecting all the Regions, an ‘understanding’ (intesa), perhaps better defined as an ‘agreement’, with the (majority of the) Regions within the State-Regions Conference is required.102 For example, in Ruling No. 6 of 13 January 2004, on the creation of new power plants and the enlargement of existing ones, the Court stated that the national plan concerning the power plants needed to be agreed within the State-Regions Conference.103 Similarly, in Ruling No. 242 of 24 June 2005, concerning a national fund supporting medium and large-size enterprises, the Court held that the criteria for the administration of the fund had to be agreed within the Conference.104 An agreement with an individual Region is required where a decision of the central government has a specific impact on that Region. For example, in Ruling No. 6 of 13 January 2004, the Court held that a power plant may be built in the territory of a Region only with the consent of the Region concerned.105
In Ruling No. 6 of 13 January 2004, the Court held that the opposition of a Region to the building of a power plant on its territory is an ‘insuperable obstacle’. However, elsewhere the Court accepted that the regional opposition can be overcome through a special procedure inspired by loyal cooperation. For example, in Ruling No. 121 of 26 March 2010, the Court dealt with the opposition of the Unified Conference to the implementation of the national plan on public housing. The Court held that such opposition could be overcome only through a special procedure involving further negotiations. Accordingly, it invalidated a provision that established that, in case of failure to reach an agreement within 90 days, the central government could unilaterally decide on public housing. In fact, the invalidated provision downgraded the position of the Regions and of the local authorities to mere consultation.106
In Ruling No. 33 of 2 February 2011 (permission to build nuclear power plants), the Court held that failure to reach an agreement with an individual Region on the identification of sites suitable for building and operating nuclear power plants can be overcome only through a procedure requiring further negotiations with the Region concerned. The negotiations have to take place within a committee comprising an equal number of representatives of the central government and of the Region concerned. If, after 60 days, no agreement is achieved, the central government will be entitled to decide where to build the plants, but the president of the Region will be involved in the decision by taking part in the relevant session of the Council of Ministers.107
No agreement but mere consultation within the State-Regions Conference is required when a decision is not ‘political’ but merely ‘technical’ (i.e., it is based on the application of standards and methods derived from science). For example, in Ruling No. 278 of 22 July 2010 (licence to build and operate nuclear power plants), the Court found that a decision of the central government identifying the types of nuclear power plants that are considered to be suitable for the Italian territory does not require an agreement with the Regions. Such decision is not ‘political’ but merely ‘technical’. Accordingly, simple consultation with the Regions within the State-Regions Conference would be enough for complying with the principle of loyal cooperation and for legitimising the lawmaking activity of the State.108
4. Lessons from the Italian Pattern
A few lessons can be learnt from the jurisprudence of the Italian Constitutional Court. Since its introduction in the Constitution in 2001, subsidiarity has been construed by the Court as a justiciable legal principle. A few State laws have been invalidated by the Court for being in breach of that principle. However, more often, subsidiarity has been used as an ‘elevator’ for lifting legislative and administrative responsibilities up to the central government. Even though this appears somewhat in conflict with the very idea of subsidiarity as ‘proximity’ to the citizen, it must be noted that subsidiarity clauses, despite their different structures, have generated a comparable outcome in Germany and on the EU level.
Unlike the German ‘necessity clause’ of Article 72(2) GG, the Italian Constitution contains reference to subsidiarity without further specifying or defining its content. However, Article 118(1) Const. includes ‘adequacy’ and ‘differentiation’, together with ‘subsidiarity’, among the criteria for the allocation of administrative responsibilities to the different tiers of government. Furthermore, in the landmark Ruling No. 303 of 1 October 2003, the Constitutional Court outlined its approach to the application of the principle of subsidiarity, by laying down a number of criteria: the State can act only if there is a need for uniform State action (‘subsidiarity’ and ‘adequacy’); State action must be proportionate, i.e., limited to what is necessary for the achievement of an objective (proportionality); State legislation must undergo a ‘strict scrutiny’ by the Constitutional Court in relation to its ‘reasonableness’; and, where subsidiarity pushes for ‘lifting’ a responsibility up to State, State action must abide by ‘loyal cooperation’. The emphasis on loyal cooperation led, in some cases, to a merely ‘procedural’ approach to subsidiarity.
Whilst the German Federal Constitutional Court has followed consistently the original Leitmotiv outlined in the Altenpflegegesetz Ruling, the Italian Constitutional Court has sometimes departed from the theme designed in Ruling No. 303 of 1 October 2003. For example, in two cases, the Court failed to require loyal cooperation. In other cases, the Court took the need for uniform State action as a given and checked only whether loyal cooperation had been fulfilled.
‘Loyal cooperation’ plays a fundamental role in the application of the principle of subsidiarity in Italy. Cooperation between State and Regions, typically consisting in an ‘agreement’ between the two, is the most original feature in relation to subsidiarity in Italy. As noted above, in this way, subsidiarity pushes towards ‘cooperation’ and ‘cooperative regionalism’, i.e., it requires the involvement of the ‘lower’ echelons of government in governance in certain fields (energy, economic recovery, etc.). This idea of ‘involvement’ is a good example of multilevel governance at work, in that one of the implications of multilevel governance is the participation of sub-national entities in lawmaking and policymaking.
A comparable emphasis on ‘loyal cooperation’ does not emerge from the ‘necessity clause’ case law of the German Federal Constitutional Court and from the subsidiarity case law of the ECJ. This is probably due to the fact that in these systems loyal cooperation in the lawmaking process, i.e. the involvement of Länder or Member States, is guaranteed by the Bundesrat and the Council, respectively. By contrast, in Italy, there is no chamber representing the Regions at national level, and the only possible form of regional involvement is the ‘agreement’ with a single Region or within the State-Regions Conference or the Unified Conference (Table 4.2).
Comparison between the Italian Constitutional Court, the German Federal Constitutional Court and the ECJ