I. Purpose

According to the original scheme of the �Founding Fathers�, the relations between supranational authorities and national administrations were to be characterised by a strict separation of their respective spheres of competence. This did in fact hold true, at least in the early stages of the development of the Communities. Thus, former Art. 100 (now Art. 94) of the EEC confines the power of the Communities to the �approximation of laws�, while it leaves the implementation of Community policies and laws [1] to the Member States. Moreover, the Treaties contain few references to organisations or procedures whose institutional purpose is to provide a link between Community and national bodies. Perhaps one can construct former Art. 5 (now Art. 10) of the EEC as the source of an obligation for mutual co-operation between Community and national authorities, including as regards the implementation of Community law.

Over the years this framework has been so thoroughly modified that it is not overstating the case to say that the recognition of administrative collaboration as a legal  reality has become one of the distinguishing features of the administration of the integrated European order [2] .

As a result, a relatively simple constitutional structure has been rendered slowly but surely more complicated by an increasingly articulated administrative system. This came about as the result of a series of progressive adjustments and ad hoc changes stemming from the completion of the internal market project. There was no clear and systematic institutional design behind this transformation. The present administrative system is characterised by the fact that the powers relating to the Community public functions are conferred upon a number of structurally separate authorities, whose nature is national, supranational or, sometimes, a composite of the two. Such transformation makes it increasingly essential to identify and analyse the whole set of legal instruments that regulate interaction between the various elements of the system, and assure the effective cohesion and integration of the various entities. Moreover, an analysis of the models of administrative collaboration can provide answers to the many questions pertaining to positive law related to the process of administrative integration. Clear examples are found in case law on the legitimacy of the practice of �comitology� and in issues related to the identification of the applicable sets of rules for administrative proceedings in which the joint intervention of national and Community bodies is required. In addition to this, an analysis of the modalities of administrative integration is relevant for a more ambitious legal reconstruction. In particular, it sheds light on  the question of the relationships between public powers in a legally complex and peculiar environment such as the European Community, which has come to assume the form of permanent interaction and even merger of what were originally two distinct legal systems.

This paper focuses on the relatively new, and most impressive and original technique of administrative integration: the adoption of a number of Community regulations each for specific sector, to be implemented not just by a supranational administration (central or peripheral), but by a plurality of national, supranational and sometimes mixed authorities, with a special role assigned to a Community office set up by the same legislation for a given sector, and granting it legal personality . These bodies are as follows: the European Environmental Agency [3] ; the European Training Foundation [4] ; the European Monitoring Centre for Drugs and Drug Addiction [5] ; the European Agency for the Evaluation of Medicinal Products [6] ; the Office for  Harmonisation in the Internal Market [7] ; the European Agency for Safety and Health at Work [8] ; the Translation Centre for Bodies of the European Union [9] ; the Community Plant Variety Office [10] ; the European Monitoring Centre on  Racism and Xenophobia [11] ; and the European Agency for  Reconstruction [12] . To these bodies must be added the European Foundation for the Improvement of Living and Working Conditions [13] and the European Centre for the Development of Vocational Training [14] . While established at an earlier date, their status was modified in the early 1990s [15] .

The purpose of this paper is to verify whether the various regulations by sector ought not to be regarded as variants of an emergent general model of joint exercise of certain Community functions. It will be argued that such general model is still in the making, but it is in the process of becoming consolidated, notwithstanding the variety of approaches adopted by European legislators. Such a pattern is characterised by specific, differentiated organisational and procedural features, which will be illustrated in the paper. If this interpretation is correct, the theoretical and practical implications of the new regulatory structure will have to be evaluated.

The analysis will be articulated in four parts, as follows: Section II, a concise summary of the theoretical background against which the hypothesis has to be assessed; Section III, identification of the common features of the regulations considered in this paper in order to ascertain the unitary regulatory structure common to all, and to clarify their main differences compared with the previous legal frameworks; Section IV, a summary and discussion of the main aspects of the emerging model; Section V, an assessment of the shortcomings and the potentialities of the new legal pattern as a tool to guide the process of administrative integration within the European context.

 

II. The Theoretical Background

For all too long studies by legal scholars about the integrated European system have failed to pay due attention to problems concerning the interaction between national administrations among themselves and the supranational public authorities in the exercise of Community functions.

This lack of attention can be certainly seen as an aspect of the traditional underestimation of the importance of administrative profiles for the establishment and functioning of the Community legal system. Notwithstanding a number of promising intuitions in the early 1960s [16] , European scholarly work in  administrative law has shown a rather clear reluctance to focus on the administrative implications of the integration process [17] . Moreover, Community administrative law has long represented an area �not only unexplored, but whose very existence was unknown� [18] , even for scholars in contiguous fields of law. As for international law, in spite of certain preliminary attempts to define the distinguishing legal features of the new international entities, including from an administrative standpoint [19] , the scope of the discussion has long remained limited to the theme of the constitutional integration of the Community and national legal systems. As for international administrative law scholarship [20] , studies have focussed on explaining international actions taken by the State, rather than the position taken by its offices vis-�-vis their international counterparts. Thus, light has been shed only on some specific questions in connection with the process of the administrative �complication� of the international order, such as the international relevance of State internal organisation and the competence of national bodies to adopt international measures.

The elaboration of a theoretical framework for the relationships between supranational and internal administrations as a way to exercise Community functions is therefore the result of recent thought developed within the context of a renewed attention to the administrative profiles of the European legal system [21] .

A first round of studies considered the various hypotheses in which the Community has not been granted powers of administrative execution. One aspect in particular that emerged is the indirect exercise resulting from the resort to national administrations in order to achieve Community objectives. Yet, it ought to be admitted that such a scheme, commonly known as indirect execution, is not based on a clear definition of the legal relationship between the national and the Community authorities, owing to the fact that it refers to an element � the indirect exercise of the function � that is common to several legal institutions and therefore has a limited explicatory capacity. In addition to this, it could be convincingly argued that there is no formal relation at all between the national administrations and the Community authorities. This is because the power of direction exercised on  the national administrations by  Community authorities manifests itself only in the course of the informal contacts frequently held between the two levels of public bodies for the purpose of avoiding a follow-up control by the Commission on the implementation of Community law [22] .

The issue of the relations between the national and the Community administrations was further explored in 1992, when research led law scholars to argue that it was possible to identify two distinct legal patterns in the exercise of Community functions [23] .

In the first pattern, the Community legislation by sector establishes that certain national public bodies can avail themselves of powers provided under national law, but only in the interest of achieving Community objectives.

In contrast, the second pattern is based on shared administrative competence, in the sense that a given Community function is formally assigned by supranational legislation to both the Commission, which is responsible for decision-making, and the national administrations, which are allotted the task of carrying out an action deemed indispensable for achieving a Community goal. In order  for such a pattern �felicitously defined as �co-administration� [24] -- to be present, a number of elements are required. First of all, the administrative competence of the Community function must be assigned to two distinct subjects acting jointly. Secondly, this shared administrative competence must be for the purpose of achieving a common Community goal. Thirdly, a formal provision must assign the necessary tasks to a specific national body, which is thus characterised by a mixed legal nature in that, although structurally part of the organisation of a Member State, its powers are authorised by Community legislation. Finally, it is necessary that the national and the supranational bodies act in close co-operation, something which excludes any possibility of independent action; in addition, the co-operation must constitute a permanent organisational arrangement.

A decade after its first formulation, the theoretical relevance of the �co-administration� model is still recognised and has even been further extended, to the point that the entire gamut of forms of administrative integration within the Community tends to be interpreted in terms of this model [25] . However, several developments in positive law have caused the multiplication of hypotheses theorising collaboration between the national and the Community administrations, with a consequent complication in terms of administrative differentiation in the supranational order.

Thus, it is no longer possible to avoid the question of whether the rapid and intense evolution of Community law considered in this paper � the so-called �agencification process� - actually provides new elements for reflection on the issue of the relationship between the national and supranational administrations. In other words, it is necessary to assess whether the regulations establishing the so-called European agencies ought still to be interpreted in terms of the co-administration model, meant as the typical scheme of joint administrative action; or whether we are instead faced with the development of new and autonomous patterns of the joint exercise of Community functions.

 

III. A Unitary Regulatory Structure

In order to address this question, it is first of all necessary to consider whether a regulatory structure common to all the relevant regulations by sector can be identified as underlying every single legislative framework. This will make it possible to identify a core of principles, rules and institutions common to each, which might be said to constitute the distinguishing features of a unitary regulatory scheme [26] .

 

(A) From this standpoint, it should first be observed that in all cases the provisions of Community law envisage a regulatory function characterised by specific objectives, sectors, tasks and addressees [27] . The regulatory structure, in other words, is not constructed as a set of legal institutions independent from the features of the administrative action, but as a way to exercise a specific function. This is in sharp contrast with the co-administration model, which does not require the exercise of a specific activity, but simply the pursuit of a Community goal.

As for objectives, in some instances they can be identified as market efficiency and the promotion of competition, while in others they are more tied to socio-economic aims. Regarding the former, Community law establishes a regulatory framework under which the interaction of economic actors is intended to strengthen, not undermine, the basic framework of a competitive market. Regarding the latter, public intervention aims not so much at constructing the basic market conditions, as  at guaranteeing a number of specific social objectives that not even a functioning market is likely to safeguard. This specifically occurs when public authorities are called upon to control the activity of economic operators for the purpose of ensuring that general objectives are met pertaining to the quality of the lives of individuals, such as in the area of public health or environmental protection.

The social nature that characterises the regulation of sectors where Community law has set up bodies provided with legal status has been highlighted in various studies. In attempting to define a feasible theoretical notion of social regulation, these studies have identified its specific objective as the solution of the �probl�mes cr��s par des cat�gories pr�cises de d�fiance du march� telles que les externalit�s n�gatives et les d�fiances en mati�re d�information� [28] . Having granted that, it must be pointed out that the mentioned investigations tend to overlook that the legal framework under consideration might have a purely economic dimension, in addition to its social one. Moreover, such research tends to assume an excessively broad definition of regulation, encompassing all kinds of public interference in the economy (ranging from planning to health to safety in the workplace) [29] , while the concept of regulation could be more accurately used to indicate �all kinds of interference that are part of the market, in the sense that they contribute to its formation establishing an equilibrium that produces interests� [30] .

The distinction between social and economic regulation [31] tends to blur when one considers that forms of intervention designed to promote competition, including because of their stable and permanent nature [32] , tend to favour consumers. On the other hand, it should be noted that the social objectives are not conceived as limits on or corrections of market efficiency, but rather as the constitutive elements of the market. This is based on the principle of harmonious, balanced and sustainable development, which aims at permanently integrating the objectives of environment, health, and consumer protection with the promotion of economic competition. The two kinds of regulation tend to overlap, coexisting in a stable manner for the achievement of both economic and social objectives.

The definition of the objectives of the function established by the relevant Community legislation also allows the identification of the sectors for intended application. These areas may be identified in reference to both the market and the social integration thereof [33] . This last reference is to those sectors that are legally relevant when developing and implementing other public policies. An example of this is environmental protection, which has been designed increasingly clearly over the years as a tool for shaping the market in compliance with the principle of sustainable development.

Such an articulation of the objectives and sectors of the function is reflected by the differentiation of the tasks assigned to the competent administrations.

These can first of all be traced back to a model of conditional regulation [34] , meaning when Community law obliges individuals to abide by specific standards of behaviour when exercising the right to engage in economic undertakings by calling upon the relevant public authorities to guarantee observance of the EC legal framework, a solution preferred to seeking to directly balance the relevant private and public interests.

A different legal structure has been chosen insofar as the hypothesis of intervention for the purpose of reaching objectives of a social nature.  Such structure is characterised by the �functionalisation� of private action in a public interest, according to the more traditional scheme of �finalised regulation�. In this hypothesis, the administrations are not placed in a neutral position with respect to the interaction of individuals, instead being called upon to reconcile the relevant private and public interests. This takes place only within a bilateral relationship of public and private power, without producing any direct effect on relations between individuals.

It must be observed that the public interest, both in the case of conditional and finalised regulation, is the Community public interest, since it is independent of  possible peculiar situations that could arise in national law, and since it is provided for by Community regulations and achieved through Community administrative proceedings. This is so notwithstanding the fact that internal and mixed administrations participate during some of the procedural phases.

The general pattern now being illustrated may, however, be implemented using many different methods. Thus, the tasks that are ascribable to conditional regulation may actually consist of the power to control in some cases, but other, more complex procedural solutions are possible. Analogously, finalised regulation may establish procedure for authorisation, but the same effect can also be achieved through soft law  measures.

Finally, on the basis of what has been observed thus far, it turns out that the objects of the regulatory function are exclusively enterprises active in the home market. It should be noted that, while in the cases of finalised regulation intervention by the public powers is directly addressed to the interested company, the measures of conditional regulation relate to a broader regulation of inter-private relationships, so that the objects of it may be parties to such relationships.

(B) In order to carry out the function examined, all the European regulations under consideration here establish a complex legal structure. Such structure is characterised by a combination of several elements in a complete and original framework.

First of all, there is a Community law provision expressly allocating the administrative competencies necessary to carry out the Community function among a variety of different offices. All such offices are thus competent simultaneously. In this way, the function is distributed on various levels, according to a pattern characterised by a degree of competition caused by the existence of different centres of authority located in either the single Member States� administrative systems or the Community organisation, or with competence jointly assigned to the Community and the Member States� administrations.

It should be noted, however, that such an allocation of attributions does not necessarily follow a single legal pattern. This allows different degrees of polycentrism and differentiation in the administration responsible for the implementation of the new Community regime. Thus, the function can be distributed among a larger or smaller number of offices, having different legal natures, exercising their function in various ways and provided with various powers not always directly identified by the Community legislation. Moreover, the legal patterns through which the allocation of the tasks is accomplished do not necessarily consist of a precise definition of the attributions and of the scope of intervention of any single competent administration. This is proven by the fact that sometimes the legislation for a given sector provides only a general description of the role of the various offices, based on their specialisation and  level of government.  

It is precisely the plurality of patterns of allocation of the competences among offices that makes it necessary to consider in greater detail the other elements of the model. As a matter of fact, the variety of organisational solutions makes it impossible to derive from the relevant legislation a definition of the specific modalities through which a joint exercise of the Community function is realised. This is so to the extent that they depend on the nature and characteristics of the tools employed to make the various subjects and their activities interdependent.

This does not imply, however, that the sharing of a competence among  several administrations is irrelevant for the present task of legal reconstruction and interpretation.

To the contrary, it should be considered that, despite the variety of possible schemes of functional distribution, the solution adopted by Community law is determined by the goals to be achieved. In fact, the choice to allocate the tasks among a plurality of national and Community bodies is determined by the specific features of the activity regulated and ideally aimed at ensuring its best and most appropriate exercise. It is precisely the principle of an efficient accomplishment of the function that predetermines the criterion on which the legislative intervention in the sector ought to be based. In other words, the most efficient way of accomplishing a given function predetermines the choice of the national and Community competent authorities, the distribution of powers and scope of intervention singly and as a whole.

Moreover, highlighting the exercise of the Community function in terms of various distinct administrations puts us in a position to reject the usual interpretation given by the European institutions themselves, according to which the disciplines by sector under consideration are primarily meant as a tool for the devolution of central tasks to decentralised Community offices [35] . Clearly, such an interpretation does not correspond to positive law, since it fails to take into account that the establishment of Community offices provided with legal status takes place within a wider organisational framework, characterized by the original combination of decentralisation and the functional integration of a variety of national, supranational and, in some cases, mixed offices. Besides, the lack of such perspective flaws other common interpretations, including the reading that presents the Community bodies as �regulatory agencies� [36] . This thesis seems quite hard to accept, if one considers that the Community body does not exercise the function autonomously, but rather in combination with all the administrations which have been assigned by Community legislation for a specific sector the attributes required to carry out that function [37] .

(C) The assignment of the Community function to a plurality of subjects is only one of the various elements of the legal structure necessary for the accomplishment of the function itself.

As a matter of fact, in addition to the legislative provision which expressly assigns the powers necessary to carry out the Community function among a plurality of offices, it is also necessary for those offices to be integrated in a unitary administration in functional and structural terms. Put differently, the various competent administrations have to exist legally as the components of a wider administration for any given sector. Such administration by sector consists, on the one hand, of the totality of the activities designed to achieve the objectives set by the Community legislation, and, on the other hand, in the totality of the public authorities  responsible for the exercise of the function.

The unitary nature of the administrative system derives from two distinct legal structures, which it would be more appropriate to examine separately.

According to one possible model, the integration of the competent offices is brought about as an articulation of the function in Community procedural sequences, characterised by the participation of supranational, mixed and internal offices. All of them are interconnected through a network of organisational relationships of variable  legal nature. In this model, Community law is characterised by the combination of two distinct elements: (1) the function is proceduralised, by establishing administrative procedures  that, even if regulated by European legislation by sector, are based on the intervention, in various stages, not only of Community administrations, but also of home and composite administrations; (2) on the other hand, a number of organisational relationships are established between the various authorities assigned the relevant tasks. These relationships are necessary in order to ensure their complementarity and functional integration.

However, Community law may also opt for a different legal structure. Rather than distributing the activity among a plurality of subjects, Community law at times establishes that the activity has to be carried out by a complex organisation [38] , sometimes defined as a �network�. This is expressly designed, in functional terms, as the totality of the activities required to accomplish the Community function. In structural terms, it is defined in reference to the totality of the competent offices. Within the framework of such complex administration, moreover, the contextually established Community body is assigned the task of co-ordinating the different relevant offices, as well as of planning the activities of all the subjects composing the complex organisation. It should be noted, then, that in order to perform this co-ordination and planning function, non-binding legal instruments are provided, aimed at achieving in a non-coercive way expected standards of behaviour for the different competent apparatuses.

Therefore, the unitary nature of the system does not derive from the provision for administrative procedures and a complex set of organisational relationships between the various responsible offices, but from an explicit characterisation of the relevant discipline and from the provisions assigning the Community body specific co-ordinating and organisational tasks, to be accomplished through legally non-binding acts.

The peculiarity of such legal structure lies not so much in the legally non-binding nature of the instruments conferred upon Community bodies for a specific sector, which have nothing more than practical effects on the activity of the offices  they refer to, as in the fact that the use of these instruments gives rise to a series of operative relationships which allow the development of a real practice of functional complementarities and interdependence.

Furthermore, the tools under examination bring into being a set of conditions at once limiting and enabling the subjects involved in the administrative action, for the purpose of making possible �a rational communicative process� [39] , to the extent that the aim is to reach an agreement on  the interpretation of a specific situation or on  the mutual co-ordination of the respective plans of action.

It is on these grounds that it is possible to clarify the reasons for the use of similar legal instruments. They might be needed due to the peculiar nature of the activity concerned, characterised by the totality of the interpretative processes arising among  the offices assigned the task of producing information, and eluding the cohesive force produced by organisational and procedural tools for the purpose.

In addition to this, the complexity of these interpretative processes accounts for the tendency to multiply the hypotheses in which the single office initiates forms of voluntary co-operation. In other words, order and stability in the complex administrative organisation are obtained not only by means of the adoption by the European agency of the non-binding measures provided under Community law itself, but also by means of the development of a panoply of �soft-law� instruments which are not specifically contemplated by Community law and which are aimed at integrating the measures adopted by the supranational body in order to facilitate the effective co-ordination of the competent offices.

The two techniques outlined in the previous paragraphs are not necessarily exclusive, in the sense that they can be used in order to complement each other in the event that this is necessary due to the peculiarity of the activity to be carried out, as exemplified by the new regime for pharmaceutical products. On the one hand, this is structured around the proceduralisation of the function and the establishment of a series of organisational relationships among the administrations assigned to accomplish the relevant tasks, and on the other hand, is based on various  forms of �spontaneous� co-operation involving those administrations, resulting in a partial modification of the �design� established by Community law.

Finally, what has been described thus far augments the previously highlighted perplexities concerning the usual interpretations of the disciplines by sector under examination. It tends to be assumed that what is being considered is a case of the devolution of central tasks to decentralised Community offices. As a matter of fact, not only do such interpretations fail to consider the Community bodies in the context of a more general organisational framework, characterised by the assignment of a competence to a plurality of subjects, but they also ignore the intricate network of relationships that come into existence involving that office, the other supranational authorities and the Member States� administrations. The architecture and dynamics of the network are left out, even though they constitute one of the outstanding features of the emerging legal model.

(D) Last but not least, the third element of the legal structure consists of the establishment of a Community office endowed with legal status and with legal tasks conferred upon it by the supranational legislation.

Such an office is, first of all, auxiliary to the establishing body, being subject to the central Community administration. As a matter of fact, accurate legal analysis reveals that some form of power of the Commission over the body is established in all the regulations by sector, albeit in varying degrees. This circumstance rules out the possibility of singling out any one specific form of expression of the power relationship, as it takes different forms in each case.

The identification of such legal trait characterising the Community body makes clear the inaccuracy of the statement that European agencies fall under the category of independent authorities. This is argued quite frequently in theories advanced by political scientists that regard the Community as a �regulatory State� [40] .

This is crystal clear in cases where powers of the central Community administration are exceptionally great. An example is found in the agency for pharmaceutical products, whose official duties are limited to advising the Commission. The latter is responsible for the adoption of a provision concluding the various administrative procedures pursuant to Community legislation by sector. However, even when the powers of the Commission are not so great, the auxiliary body never enjoys independent legal status vis-�-vis the central Community administration in order to safeguard it against political or bureaucratic interference, as would be the case with  fully independent entities.

Besides, the decision to establish a body auxiliary to the Commission is not so much to be attributed to a need to guarantee such bodies a certain degree of independence from political, bureaucratic and economic power, as to a desire to  discipline the matter in decentralised fashion.

The aim is not really to establish certain organisations capable of acting independently of outside interests able to influence  their final decisions, as independent authorities are typically meant to do [41] . Rather, the attribution of legal status to an office is intended to ensure the performance of an activity which, for political or technical reasons, cannot be directly carried out by the central administration, even though decentralisation could be accomplished in different ways  according to the various hypotheses (for example, by means of a partial decentralisation or a full devolution of decision-making powers to peripheral or non-central Community offices).

On top of that, the element of decentralisation does not complete the functional features of the body under examination, this being an organisation not only decentralised, but also intended to create and manage a plurality of relationships involving the central Community administration, the national governments and, in some cases, the European Parliament.

The aim of co-operation among the different authorities is determined by a specific legal pattern, based on the exploitation of the organisational formula of representativeness. In this formula, certain centres of reference representing interests identified by Community law nominate the subjects composing the various internal offices of the European agency, who in their turn represent to the Community authority the interest expressed by the body in question. This does not imply applying to said body the effects of the measures adopted by the offices of the European agency.

Besides, representativeness can be applied in different ways in the various hypotheses, either by stressing the profile of the composition of the supranational and intergovernmental interests, or by attributing greater relevance to Member States� interests than to the supranational interest.

In addition to this, the internal organisation of the body may also vary from case to case. Although it is always necessary to provide for collegiate bodies representing a multiplicity of interests and composed through the procedures provided by the establishing legislation, the establishment of collegiate bodies instrumental to allowing  discussion among  experts on specific scientific matters is only an eventuality, since such matters cannot be easily interpreted in terms of the protection of national interests, owing to their exclusively technical nature.

Such a picture underscores the partiality of the scholarly interpretations that consider the so-called European agencies as essentially aimed at realising a technical or administrative decentralisation of services, while underestimating the profile of administrative co-operation found within the Community body [42] .

But a similarly limited explicatory power is also found in interpretations which, while stressing the integrative and co-operative function of the bodies under examination, fail to recognise their specific legal content. These interpretations simply acknowledge the emergence of new forms of administrative co-operation in the European system, viewing them as functionally similar to previously existing ones, and as differing from earlier forms only in their major completeness and institutional stability [43] .

Actually, the legal profile of the organisation under examination is characterised both by its status vis-�-vis the establishing body, which is never legally guaranteed, and by its being organised according to the formula of representativeness.   To this corresponds a peculiar functional character, since the Community body established by the Community legislation by sector simultaneously produces a decentralisation of the matter and an integration of distinct authorities, according to the modalities and within the limits set by Community law.

This picture is consistent with the general framework laid down by the European legislation, founded, as previously noted, upon (1) an attempt to address the need for plurality deriving from the nature of the function; (2) the necessity to identify in the supranational organisation the subject of reference; (3) the opportunity, both political and organisational, not to overburden the central Community administration with additional tasks, according to a pattern which presupposes not only a shared allocation of powers to supranational and Member States� administrations, but also a decentralised ordering of the matter and a functional and structural integration of the various competent offices.

In fact, if the distinguishing feature of the resulting model is the development of the hypothesis of decentralised integration for the accomplishment of a specific Community function, the so-called European agency is an essential element of that model, not only because its institution excludes or circumscribes the powers of the Commission�s active administration, but also because its structure is such as to allow a form of integration and collaboration between distinct offices, even if only in the specific sense outlined above.

In this perspective, in addition to the already mentioned interpretations that single out only one of the aspects of the element, and not always adequately highlighting its legal relevance, we can discount references to other forms of administrative integration characterising the European legal system, among which the institution of comitology committees, which do not meet the requirements of decentralisation, and are rather the expression of the joint competence of the Member States� bureaucracy vis-�-vis the Commission in the process of the approval of  Community legislation for implementation.

 

IV. Decentralised Integration as a New Model of Joint Exercise of Community Functions

The discussion thus far, it should be clear that the legal model (�decentralised integration�) fleshed out in this paper deserves to be characterised as a new legal model of the joint exercise of supranational functions. Its novelty is, of course, relative, as it borrows from previous Community and national legal institutions and experiences.

The regulatory structure that emerges is first of all characterised by the shared allocation of powers to the Community and to the Member States� administrations. This is the result of the conferral of administrative competencies concerning the same Community function to a plurality of offices, so that the Community function is distributed at two or more levels, none of which able to handle it autonomously.

Secondly, as a consequence of the various tools for co-ordination and co-operation prescribed by Community law, the competent offices must be seen, from a legal standpoint, as parts of a broader administration by sector. This broader administration can be identified both with the whole set of activities aimed at reaching the objectives singled out by Community law, as well as with the public authorities  responsible for the exercise of said function.

Thirdly, this regulatory structure presupposes the establishment of a Community body granted specific powers under Community law and informed by the organisational formula of representativeness. The purpose of creating the body is at once to realise a decentralisation of the matter and an integration of distinct authorities, according to the modalities and within the limits provided by Community law.

Fourthly and finally, it should be noted that such a legal framework is not used in order to ensure the achievement of any Community goal. It is rather aimed at carrying out a specific regulatory administrative function, with a view to satisfying economic and social purposes, in relation to which the public authorities  are conferred conditional regulatory tasks and teleologically prescribed regulatory tasks.

Such structure is peculiar in two respects.

Firstly, this legal model is based on a specific pattern of exercise of a particular regulatory function. The legal model, in other words, cannot be used in order to carry out any kind of Community action, since it is specifically designed with a view to the �regulatory model� which constitutes a typical aspect of the discipline of economic activities within the European Union.

Secondly, the regulatory structure under consideration is characterised by its peculiar combination of decentralisation and integration. As a matter of fact, the joint exercise of the Community function in this case requires a structure which is at the same time decentralised -- through the establishment of a supranational body which excludes or restrains the administrative powers of the Commission -- and integrated -- through the provision by Community law of organisational and procedural tools for co-operation among the plurality of competent public bodies.

Thus, the resulting model can be described neither as ensuring the supremacy of the Community administration over national bodies, nor as an instance of a permanent centralisation of the exercise of certain powers by bodies or procedures with an associative nature.  Rather, the joint action results in a pattern of decentralised integration.

The identification of such connection between decentralisation and integration provides an a posteriori justification for the new model of joint exercise of Community functions. This is because such model is the result of a political compromise between the Member States and the Commission, rather than the outcome of an intentional institutional design. Moreover, a variety of reasonable criticism can be made concerning the overly-complex and dysfunctional choices intrinsic to the model. And yet, the scheme of decentralised integration reveals, if only a posteriori, the attempt to reach a sustainable balance between (1) the necessity for �pluralisation� deriving from the specific features of the function; (2) the necessity to identify in the Community the subject of reference; (3) the imperative, both organisational and political, not to overburden the central Community administration.

At the theoretical level, the model of decentralised integration provides new conceptual tools to interpret and explain the process of administrative integration between the supranational and the national public powers.

First of all, such model makes it possible to work out a taxonomy of the various legal patterns of the exercise of Community functions. More precisely, in light of the model of decentralised integration it is possible to single out four such patterns.

In one possible pattern, usually known as indirect administration, the Community has no power of administrative execution, with the accomplishment of  Community objectives set by the Treaty or by the relevant secondary legislation being the province of the national administrations.

In other instances, the Community centralised administration is required to act directly, which is what is meant by direct execution. However, it is important to stress that this second pattern is not a completely free of outside influence, since in practice we find increasingly closer  co-operation between the Commission and the national administrations.

Moreover, during the last 15 years, the traditional dichotomy between the direct and indirect exercise of Community functions has been blurred by the emergence of a number of forms of joint execution of supranational laws and policies. Though differing in several respects, the various solutions adopted by Community legislators may be classified under two main legal patterns.

The first such pattern is the so-called co-administration model, developed in the mid-1980s and frequently implemented within the context of the common agricultural policy [44] ; the second one, presented in the present paper as decentralised integration, is a more recent and increasingly more common model of the joint exercise of Community functions [45] .

Despite its widespread use in Community legislation, the model of decentralised integration is far from covering and explaining all the latest forms of administrative integration within the European context.

Let us consider the example of the new food security  regime [46] . At first glance, the administrative system envisaged by the relevant Council Regulation presents a number of clear similarities with the European administration for pharmaceutical products and can be partly referred to the decentralised integration scheme. In fact, many are the analogies with that model: in particular, the Community administrative function is distributed among a plurality of public authorities ; a European agency by sector has been established; a number of tools for co-operation among the various competent administrations are envisaged. However, other elements make it difficult to satisfactorily include this regulatory structure in the decentralised integration model. In particular, this system puts a rather clear emphasis on the independence of the �technical� administrations from the policy-makers, and this is so in two different senses. First, the European agency is defined as a Community body characterised by �independence, excellency and transparency�. Second, the internal organisation of the European agency envisages the institution of an advisory body made up of the national offices responsible for functions analogous to those carried out by the European agency. This is a very important point because, while in the decentralised regulatory structure the Community legislation is neutral concerning the features of the national competent administrations, in this case Community legislators have envisaged a complex system of independent authorities which can be interpreted as limiting the administrative autonomy of the Member States [47] .

The food security  regime is but one of several cases that cannot be interpreted in the light of decentralised integration. In general terms, the problem raised by these recent developments in Community positive law is whether it is possible to identify a third model of joint administrative action in addition to the two models mentioned of co-administration and decentralised integration: a model characterised by the integration of various independent bodies. The existence of this third model of independent integration, however, has not yet been convincingly proven, as the existing literature on the issue has been limited to investigations of certain sectors and founded on legal analyses and reconstructions which are not always impeccable.

In addition to providing the basis for a taxonomy of the various legal patterns of exercise of Community functions, the model of decentralised integration is important for its analytical capacities.

By means of the model it is, in fact, possible to relate certain organisational and procedural tools for administrative integration with more general models of the exercise of Community functions. For example, while the co-administration model normally relies on composite or mixed administrative procedures (that is, administrative procedures not only involving the participation of national and supranational administrations, but also regulated partly by the Community legislation and partly by the national legal regimes), decentralised integration typically relies on Community procedures fully regulated by supranational law in spite of their polycentric structure.

Finally, a third important aspect of the decentralised integration model concerns its relevance for non-legal theoretical approaches to the European integration process.

As a matter of fact, the new administrative reality as described in this paper represents a challenge to traditional theories dealing with the question of how to explain the process of integration in the EU. This is obvious as far as intergovernmentalism is concerned, provided that the new forms of administrative integration unambiguously show that the States are not at all unitary actors pursuing their self-interest, and that bilateral and political relations are complemented by a complex network of partnerships [48] . Moreover, the multi-level character of decentralised integration adds a significant new layer of institutional complexity to the neo-functionalist theories, thus confirming the intuition expressed in the �new governance agenda� as to the need for redefining the simple supranationalism that is embedded in neo-functionalist thought [49] .

Yet, what seems to be the most promising element of the decentralised integration model vis-�-vis the European integration process is its communicative potentiality. As stressed in section III, the legal framework determines the existence of a set of conditions limiting and, at the same time, enabling the subjects involved in an administrative action. These conditions should permit �a rational communicative process� to take root, being intended to facilitate the reaching of agreements on the interpretation of a specific situation and/or the mutual co-ordination of the various subjects� plans of action. If this interpretation is correct, the decentralised integration model could then be of some utility within the context of the emerging lines of research directed at developing a new theoretical frame of reference capable of understanding and interpreting European integration. This is the case, in particular, with certain innovative theories, such as the legitimation through deliberation theory, according to which �integration also occurs through deliberation, or what is commonly referred to as arguing. This type of integration is very important, as stability depends on learning and alteration of preferences. Deliberation, when properly conducted, ensures communicative processes where the force of the better argument will sway people to harmonize their action plans and transfer agreements into binding contracts, with the aid of the legal structure in place� [50] . While a complete reflection on the connection of the model to this theory exceeds the scope of this paper, the line of research sketched is undeniably relevant and could open new horizons in the study of European administrative law, as well as in the exploration of the issue of democracy in the EU.

 

V. �Decentralised Integration� as a Sound and Feasible Institutional Arrangement?

Thus far, it has been argued that a new legal model of joint exercise of supranational functions -- for the sake of convenience referred to as �decentralised integration� -- has been emerging during the last decade. In addition, the theoretical relevance of the identification of such regulatory pattern has been highlighted.

It is now necessary to assess whether this model really represents a sound and feasible option for the administrative evolution of the Community legal system. Actually, the model of decentralised integration is not above criticism, and it may be considered weak and as having merely conceptual relevance. The argument that will be developed in this section is that, although it may not be the best possible administrative institutional arrangement, the model of decentralised integration is nevertheless a rather useful model, since it provides a positive and pragmatic answer to a number of administrative questions arising at the present stage of the maturing supranational legal system.

(A) Let us consider a first possible critique of decentralised integration, concerning the perils of administrative pluralisation. As expressed earlier in this paper, one of the essential elements of the decentralised integration model is that the administrative competencies concerning any one Community function are assigned to a plurality of offices, so that the Community function is distributed on two or more levels, none of which can fulfil it autonomously. This complexity of levels might be viewed as highly problematic. Arguably, such a pluralisation of the function is not always necessary and in some cases it could even produce certain shortcomings in the exercise of the Community action. On certain occasions, full centralisation (from the Member States� administration to the Community � central or decentralised � administration) might be considered more advisable. How can we then be sure that the institutional choice of distributing one function among a variety of national, supranational and mixed bodies is always appropriate? Would it not be more convenient to opt for a different regulatory structure which does not necessarily imply the distribution of the function among a multiplicity of distinct bodies?

It is undeniable that this critique highlights a true weakness of the model. By definition, decentralised integration is very complex, in that it presupposes joint action by many different administrations, while, at least in principle, simpler institutional solutions could be identified.

Nevertheless, an empirical examination of the various sectors in which  Community legislators have decided to implement this model demonstrates rather clearly that the pluralisation of the Community administrative function is always instrumental either to the action to be carried out or to certain specific features of the regulatory structure.

Let us consider, for example, the European trademark administration. Here Community legislation has provided for the participation of the national central administrations for industrial property rights in the exercise of administrative actions in relation to the Community trademark. The choice derives neither from the nature of the protected interests, which have an exclusively supranational nature, nor from the exigency to ensure the full effectiveness of Community law. Rather, it is an institutional arrangement which finds its justification in the opportuneness of using the national administrations to handle specific matters of local relevance, such as investigations into trademarks already registered nationally. Moreover, the involvement of the national administrations is related to the peculiar structure of the Community discipline, which does not harmonise the national legal regimes, but co-exists with them, making it possible for the individuals operating in a single market to choose between two different legal regimes. Such choice requires certain forms of connection between the national and the supranational orders of administrations.

The European administration for pharmaceutical products is a more problematic case. In this sector, in fact, several doubts may arise in relation to the quite complicated overall organisational architecture. One could contest, in particular, the criterion adopted in Community legislation in the allocation of the tasks between the European agency and the Commission (that is, between the decentralised and the central administration). Actually, such criterion is not based on a division between the scientific assessment and the political choice, since both authorities express a purely scientific and technical judgement. Moreover, the centralised authorisation procedure is assigned an excessively limited role, while both the assigned role of the national governments and, more generally, the extremely complex procedural architecture are open to criticism. On the other hand, however, one should not underestimate the peculiar features of the sector regulated by the Community legislation, which requires, as a complement to authorisation activity, continuous and effective control over the negative side effects of the authorised pharmaceuticals. From this standpoint, the organisational architecture is indeed complex, which is only augmented by the decision  to base the current regulatory framework on a series of reciprocal commitments involving various subjects expressing different but complementary points of view within an overall administrative system.

While other examples could be cited, the general point should already be clear: the element of pluralisation (that is, the allocation of the relevant tasks to a multiplicity of bodies) is a necessary feature of a modern administrative framework for action. It is quite unrealistic to imagine the typical Community administrative functions being effectively performed without the involvement of a plurality of administrations (Community, national and mixed). This does not mean, however, that all such cases have been well thought out and structured. It is no easy task to justify the extreme complexity found in some of the existing cases of decentralised integration, so that in the future it will be necessary to design more effective forms of organisational architecture.

(B) A second critique of the decentralised integration model relates to the effectiveness of the techniques envisaged by Community legislators for assuring  functional integration among the several competent offices.

The typology of the instruments for co-operation and co-ordination used in fulfilling various regulations has already been singled out for discussion. In particular, it has been argued that the unitary nature of the administrative system by sector derives from two different legal structures. The first is characterised by the establishment of a unitary administrative system, with a Community agency assigned specific co-ordination and organisational tasks to be accomplished through legally non-binding acts. The second legal framework is characterised by the combination of two distinct elements: one, the proceduralisation of the administrative function through the establishment of Community administrative procedures based on the intervention, at various stages, of internal and composite administrations, as well as Community administrations; two, the setting up of a number of organisational relationships between and among the various authorities responsible for the relevant tasks.

In the perspective of public management and organisational theory, doubts may arise about the effectiveness of this second type of instruments for co-ordination. Such perplexities are reinforced by the circumstance that these instruments are used in order to integrate administrative systems that are really very complex (and certainly more complex than the systems structured according to the first technique for achieving integration). One may therefore wonder whether �flexible� instruments � essentially based on soft-law measures for co-ordination � are really adequate to manage this complexity or whether instead they could eventually lead to the collapse of the administrative system [51] .

A clear example of such risk is provided by EIONET (the European administration for environmental information). EIONET is, in fact, a network made up of more than 600 bodies performing different functions (some of them responsible for the co-ordination of the national competent administrations and for the management  relations with the European agency; others for gathering information in certain geographical areas or with reference to certain specific issues, etc.). These bodies are also identified in different ways: sometimes they are identified in Community legislation; sometimes they are identified by an agreement that has to be negotiated by the European agency and the national governments. Moreover, the polycentric nature of EIONET is intensified by the provision requiring the performance of a number of duties involving co-operation with other external administrations (for example, EUROSTAT) and with third-country administrations.

Can it truly be said that �flexible� tools for co-ordination are adequate for managing such complexity (or turbulence, to use the public management terminology [52] )? Or would such an arrangement lead to the paralysis of the administrative system?

The author�s position on this point is that these tools for administrative integration are much more advanced than one might suppose at first glance. This is so because, as stressed earlier, the use of these tools gives rise to a series of operative relationships which allow the development of a real practice of functional complementarity and interdependence.

To elaborate on the point, consideration must be given to the definitions and concepts commonly used in this type of administrative system, such as, for example, the notion of drug addiction in REITOX (the European administration for drugs and drug addiction), or the definition of water pollution in EIONET. Notions of this kind are not unitary and vary according to the office making reference to them. For instance, the notion of pollution is not the same for an observer, a polluter, an office using probabilistic patterns or empirical models, and so on. Moreover, these notions are, by definition, continuously rewritten in the information chain (because they are formalised, because they are put into the context of other notions, and the like). How, then, is it possible to arrive at a commonly accepted definition? One might say that this occurs only if and when a series of operative relationships are established in order to gather together in a unitary discourse the irreducible multiplicity of concepts. For example, in the REITOX system the emergence of a single notion of drug addiction depends on the possibility of grouping in a unitary system the various relationships and interactions that take place at the different levels and in the different processes: among the various processes of description, articulation and classification; among the different criteria used to assess the scientific validity of a notion; among the relationships between observation, deduction, etc.

The same applies to the object of the administrative action. As such, the object cannot be considered as the unifying factor of the administrative system, because it fragments into a multiplicity of elements. To cite but one example, the theme of environmental protection can be used in a variety of theoretical models, of which the correction of the market and the maintenance of an ecological equilibrium between living beings are only two of the most obvious instances. In addition to this, the conception of environmental protection varies greatly according to the perspective of the actor. EIONET brings together consumers; private bodies, such as laboratories; public bodies, such as the public offices responsible for crisis management or the public bodies responsible for industry, and so on. Each of them unavoidably has a different understanding of what is meant by environmental protection. How, then, can this multiplicity of factors be composed and unified in a unitary architecture? What the author would suggest is that the translation of these factors into comprehensive empirical action can take place only if and when a series of operative relationships are established, such as the relationships between the various filters used in the information chain (in the case of REITOX, for example, this would involve the questionnaire, the scientific evaluation, the police action, the judicial inquiry, etc.).

Most notably, these relationships cannot be legally predefined. They can be originated only in the practice of structural or procedural forms of voluntary co-operation among the various components of the administrative system.

The kind of administrative order and stability resulting from recourse to this technique is therefore at once spontaneous and externally determined. It is spontaneous because the relationships underlying it are freely developed in the interactions between the various offices and cannot be predefined by Community legislation. It is externally determined because such relationships derive from the adoption by the European agency of a series of non-binding measures as provided under the establishing Community regulation. This is true in the case of EIONET, with the various work projects undertaken by the EEA. Moreover, such administrative order is not formal in nature; on the contrary, it is empirical and may be continuously modified and readily adapted to changing conditions.

This reconstruction should help clarify the reasons for the use of similar legal instruments in existing Community regulations. It further makes it possible to state that this technique for administrative integration is not only an original concept, but also rather well-balanced with respect to the features characterising these overly-complex administrative systems. Actually, the peculiar nature of the activity involved , which is characterised by the whole system of interpretative processes arising among the various offices assigned the task of producing information, eludes the cohesive function of the formal legal tools for connection. The only way to achieve some degree of functional integration among the various bodies and their actions is through such flexible tools as those mentioned above.

(C) A third critique of decentralised integration concerns the establishment of a Community office conferred with legal status and assigned specific legal tasks by the supranational legislation.

As previously argued, the choice of establishing a body auxiliary to the Commission is not so much related to the necessity of establishing a body that is independent from political, bureaucratic and economic power, as to the appropriateness of providing a decentralised discipline for the matter.

If this interpretation is correct, one might wonder whether a requirement of independence would not be more appropriate -- in other words, whether an independent Community body would not be a better institutional arrangement than a decentralised body, provided with limited management autonomy but always auxiliary to the Community central administration. In the words of Majone and Everson, �Where administration is buried within or answers directly to governmental departments, political goals may be easily adjusted, or subverted, without any public debate� [53] . Accordingly, it is apparent that granting agencies and similar bodies a degree of independence from the government can, in fact, �enhance the credibility of regulatory commitments. Independent regulators have strong incentives to pursue the statutory objectives assigned to their agencies, even where the objectives are no longer politically popular� [54] .

Clearly, this is a powerful critique of the model of decentralised integration, which might appear to insufficiently insulate the policy-maker from the administration. Nevertheless, it must be recognised that, contrary to the argument made by Majone and Everson, the necessity of independence from politics is not always fully demonstrated.

As a matter of fact, the above authors propose the establishment of independent agencies as the most appropriate answer to a general institutional problem of the Community. In particular, they argue that independent agencies represent an appropriate response to the internal as well as external threats to the credibility of EC regulation. Internal threats arise from the �serious mismatch between the Community�s highly complex and differentiated tasks, and the available administrative instruments� [55] . External threats originate in the social, economic or political context in which the system is embedded: among these threats, strong emphasis is put on the risks inherent to the process of progressive parliamentarisation of the Commission.

Some of these statements could obviously be the object of considerable criticism. What is important to highlight, however, is that in this approach the setting up of independent agencies is considered to be the best possible solution to a general institutional problem of the Community. Assuming this approach, the various spheres of Community action are not differentiated and all receive equal treatment.

In the author�s view, it is not that the case made for independent agencies is wrong in itself. There are obviously many instances when a legal framework so-conceived could prove useful and even necessary. The problem is that the establishment of independent agencies cannot be presented as a universal solution to a general institutional problem, because of its inevitable failure to differentiate among sectors and fields of action which often do differ. A more flexible and analytical approach, based on the consideration of the nature and of the features of the administrative function to be carried out, would therefore be preferable. In this perspective, one is led to identify cases where the establishment of independent agencies could be seen as necessary (for example, as pertains to guaranteeing food security ); cases where independence is not necessary at all (for example, the Community trademark regime); and a number of problematic cases that require further reflection (for example, telecommunications).

Within this context, decentralised integration should not be considered a weak or obsolete model. In several sectors where its utility and effectiveness can be demonstrated, it could and it should be adopted in Community legislation. The critical issue, then, becomes the choice of an appropriate type of decentralisation, meaning one able to strike a reasonable balance between the Commission�s role and that of its auxiliary body, the European agency, which is not true for certain actual cases of decentralised integration.

(D) Last but not least, it is necessary to briefly assess the legitimacy of the decentralised integration model. In this respect, it must to be stated clearly that while the current form of decentralised integration is deficient in several respects, nevertheless, its legitimacy could (and should) be increased by further developing  certain elements already intrinsic to the model, albeit in an unclear and insufficiently elaborated way.

Let us first consider the procedural regulation of the new legal pattern.

It is common knowledge that, according to a fairly widespread school of thought, proceduralisation offers a way to achieve a higher degree of legitimacy for European administrations (in particular, comitology) [56] . More precisely, proceduralisation serves as an instrument for regulating the interface between Community administrations and the general public according to the principles of transparency, openness and participation. Thus, the legitimacy of the post-legislative administrative phase could be bolstered by a process-oriented approach where interested citizens are given a say.

Assessed against this background, the model of decentralised integration still cannot be considered fully satisfactory. In particular, the various administrations by sector tend to be treated in different ways, because of the differences in the procedural schemes of the various sectors, ranging from full codification in law to the absence of law. Thus, at one extreme we find the EC trademark administration, where a panoply  of procedural provisions regulates the exercise of administrative action and the related litigation, leaving virtually no room for soft law. At the other extreme, we find the environmental information system, where action is devoid of procedural requirements. Clearly, the level of protection afforded by procedural guarantees corresponds to the degree of their codification: while the EC trademark administration guarantees extensive legal protection to the applicant as well as to third parties, the pharmaceutical framework is incomplete in many respects, while in its actions EIONET simply ignores the position of the interested parties.

Here a reasonable conclusion is that the existing procedural framework is, as such, unsatisfactory; however, certain advanced regulations, such as those concerning biodiversity and Community trademarks, could be conveniently used by Community legislators as a source of inspiration for developing an adequate set of procedural guarantees applicable to any instance of decentralised integration.

The so-called �procedural avenue�, however, is not the only possible source of legitimacy for the decentralised integration model.

In fact, thought on network management has elaborated its own solutions to the issue of the legitimacy of complex administrative systems. It is deemed necessary to establish mutual trust among the various subjects of the network, which can only be achieved through a pattern of flexibility and efficiency. Law and legal instruments are excluded: self-control is preferred to external control, whether administrative or judicial, and the spontaneous development of codes of conduct in the interest of  fostering loyalty among the partners is encouraged. In this perspective, the network could and should facilitate the development of behavioural standards and work practices that create shared expectations and enhance the effectiveness of  administrative action [57] .

 Interestingly enough, this line of reasoning seems to find a rather clear confirmation in the empirical functioning of the complex administrative networks managed by the European agencies. As a matter of fact, observation of the various systems indicates a tendency to establish a number of relationships that may be considered examples of mutual trust among the different administrative partners. This is particularly true in cases where there is a clear lack of formalised rules of administrative action and where the organisational environment is highly turbulent.

In addition to the preceding, it seems appropriate to refer to a third source of legitimacy, consisting of a set of different institutional arrangements that are typical of the decentralised integration model. Unlike the previous sources of legitimacy, such institutional instruments do not refer either to output legitimacy or to processed-based legitimacy. Rather, they refer to the way in which the relations among the actors are arranged in a given administrative system.

In this perspective, three relevant elements should be taken into account.

To begin with, the administrative systems set up within the context of the decentralised integration model are based on the participation of the Member States, in the form of the representatives of their central administrations. This constitutes one of the various sources of legitimacy in the Union as it presently stands: an indirect source, which derives from the Union�s coming into being through an agreement by a number of States, subsequently joined by others States [58] . In this perspective, it is necessary to clarify that the Member States may participate in these administrative systems in different ways: at times as single States (for example, in administrative procedures where the possible involvement of just one Member State is envisaged); at times as members of a collegiate body (as with the comitology committees, which are sometimes members of the administrative system); at times as parties to negotiations (in all cases requiring administrative co-operation).

Moreover, the Parliament itself participates in these administrative systems through a number of representatives in the internal offices of the European agencies. Despite possible doubts as to its effectiveness, such participation represents an element of direct legitimation of the decentralised integration model.

Finally, there is the legitimacy deriving from the competition of regulatory systems encapsulated in several Community regulations. In several instances, exemplified by European legislation on pharmaceuticals and by Community trademark regulation, individuals are given the opportunity to choose among a variety of national legal regimes, or between the national and the supranational regime. Thanks to reciprocity, in other words, they may compare different national legal regimes and choose the most convenient one. This is a largely unnoticed way toward the legitimation of Europe, perhaps because it pertains to certain sectors. Nonetheless, it is interesting to note that a Europe-wide public arena is created thereby, in which companies and consumers interact with public authorities in a competitive way [59] .

 

VI. Conclusions

The main thesis of this paper is that an accurate legal reconstruction of  Community regulations establishing the so-called European agencies reveals the emergence of a new legal model of joint exercise of supranational functions. This model can be defined as decentralised integration and is characterised by specific and differentiating organisational and procedural features. Decentralised integration is plainly distinct from other models of administrative integration in the Union. In particular, it cannot be collapsed into the contiguous notion of co-administration.

This conclusion is relevant in several different ways.

Firstly, it provides new conceptual tools for interpreting and explaining the process of administrative integration between supranational and national public authorities , in particular by specifying the taxonomy of the patterns through which a Community function can be carried out by two different authorities acting jointly. Thus, to the traditional schemes of direct and indirect exercise of Community administrative functions, the new models of co-administration and decentralised integration can now be added. Yet, decentralised integration should not be considered as a passe-partout for interpreting all recent transformations in the Community administrative system. The regulatory structure foreseen in the new telecommunications regime is just one of several examples of administrative integration which cannot be reduced to the decentralised integration model. In this regard, further empirical and legal research is necessary in order to reconstruct in proper conceptual terms the new administrative evolution of the supranational legal order.

Secondly, the new regulatory structure described in this paper could prove useful in addressing some of the vexing legal issues that inevitably arise in the context of complex administrative systems, such as, for example, the scope and the limits of co-operation between the single units, or the legal principles and rules applicable to the administrative action of the Community system as a whole.

Finally, the decentralised integration model should be considered as a sound and feasible option for the administrative evolution of the Community legal system. As a matter of fact, while some criticism may be advanced to demonstrate its practical weaknesses, the new regulatory structure should be regarded, if not as the best possible administrative institutional arrangement, at least as an acceptable and even suitable model of administrative action, providing a positive response to several administrative issues emerging at the current stage of the supranational legal order.