The formal independence of the 32 European regulatory agencies
This thesis is concerned with attributing formal independent scores to the contemporary 32 European regulatory agencies. It compares those results with an earlier study to examine whether one can observe how these scores have evolved over time. One of the major explanations why some ERAs possess higher scores than others can be accounted to the field of operation of such an ERA. ERAs operating in economic field possess higher formal independent scores than ERAs concerned with social activities. The difference in score is explained by the idea of making credible commitments, that are of especial concern for economic ERAs. However, the method to measure formal independence have been challenged, and slight adjustments backed by the literature results in different answers to the hypothesis. A result that raises doubt about the former operationalisation of formal independence, and are of particular concern for future research in this area.Master Thesis
Xander van der Klaauw S1600052
Supervisor: Dr Madalina Busuioc 2nd reader: Dovile Rimkute
Leiden University
Faculty of governance and global affairs Master Public Administration
Table of Contents
1. Introduction ... 3
2. Theoretical Framework ... 6
2.1 The three waves of agencification ... 6
2.2 Functional reasons for agency creation ... 7
2.3 Political reasons for agency creation ... 8
2.4 Hypotheses ... 10
2.5 Key concepts ... 13
3. Research design & methodology... 13
4. Analysis ... 17
4.1 Formal independence score of the 32 ERAs ... 18
4.2 Answering the hypotheses ... 23
4.3 Critique on scoring index + alternative model ... 26
4.4 A different look on the hypotheses ... 28
5. Discussion ... 30
6. Conclusion ... 31
Reference List ... 34
Appendix 1. Scoring index formal independence of ERAs (by Wonka & Rittberger) ... 37
Appendix 2. Detailed scoring info for each ERA ... 40
1. Introduction
In his classic paper from the positive to the regulatory state described Majone (1997a) the upcoming changes that he would expect in the governance structure of Western countries, Majone presumed that a interventionist style would be replaced by rule making through delegation (Majone, 1997a, p. 139). Seen from the EU perspective appear the expectations from Majone well-grounded. Within the last twenty five years have the amount of European regulatory authorities (ERAs) been increased from two to around forty. One of the principal reason for this delegation towards independent ERAs is that politicians are ought to be better able to “commit themselves to regulatory strategies that would not be credible in the absence of such delegation (Majone, 1997a, p. 163). Politicians operate under constant pressure to publically score, it is assumed that these electoral incentives make it likelier that politicians switch to what “now appears to be a better policy”, which can have a “negative impact on policy continuity” (Majone, 2000, p.3; Kydland & Prescott 1977). In comparison to institutions as the Commission or the Council do ERAs not have electoral incentives and are therefore ought to be better able to accomplish longer term policy objectives.
Making credible commitments by delegating powers to non-majoritarian bodies seem especially relevant at EU level. The deepening of the internal market in the past decades let to spill-over effects and an increase in European Union competences. This not only increased the degree of politicisation about which competences the Community should pick up, but above all did several crises expose the actual mismatch between the administrative structure of the Commission and its activities it performed at the beginning of this century. The Commission was well able to make new regulations but fell short in implementing these regulations. (Vos, 2000; Majone, 2000; Busuioc, 2013, p. 24). Agencification was aimed to restore the credibility of the Community, in order to signal to potential investors that decisions were less likely to be influenced by electoral incentives in the future (Lodge, 2008; Egeberg & Trondal 2017; Richardson & Mazey, 2016).
The last decades can one not only observe a quantitative increase in the amount of ERAs, but foremost a ‘qualitative increase’ (Busuioc, 2013, p. 15). A couple of the latest founded ERAs are empowered with significant (quasi) rule making powers and are formally considered as the most independent ERAs today. Though, within the academic literature there have not been many studies investigating their independence. The academic contributions so far are primarily focussed on measuring the independence of national regulatory authorities (Gilardi, 2005a;
Groenleer 2009; Pavón 2018). So far just one research has examined the formal independence of the ERAs (Wonka & Rittberger, 2011).
However, several reasons render a renewed investigation necessary. First of all, Wonka & Rittberger (2011) only incorporated the ERAs that were part of the first and second wave of agencification, as such Wonka & Rittberger (2011) do not account for the seven new ERAs that were established more recently. Some of these agencies are among the most powerful ones that the EU knows today (Busuioc, 2013, p. 16). The three regulatory agencies that together form the European Supervisory Authority (ESA) are by exception ones that can adopt direct binding decisions to national supervisory authorities, and may also take decisions that have impact on the financial institutions of the individual member states. Secondly, a couple of ERAs have seen their mandate changed since 2008. Frontex and Europol are two such examples (Meissner, 2017; Busuioc, 2013, p. 33). It can be that the changes that are made within the new mandates had also their effect on the formal independence of those agencies. Next to that, other agencies have seen the span of powers were the agencies operate adjusted. In for example the field of internal security have agencies as Eurojust and Frontex been given a more pivotal role, especially in safeguarding the internal security strategy (Busuioc, 2013, p. 7). Next to the changes in the institutional landscape are there also changes our understanding of how to adequately measure the formal independence of regulatory (European) agencies. Since the article from Wonka & Rittberger appeared in 2011, is the operationalisation of formal independence been critiqued by Hanretty & Koop (2012). The article addresses various flaws, but above all makes a couple of suggestions how one is better able to measure the formal independence of (European) regulatory agencies.
The central question in this thesis is therefore: how levels of formal independence vary across European regulatory agencies? Furthermore is this thesis interested in explaining possible differences in formal independent scores. Next, what is the impact is of the third wave of agencification with respect to agency independence? How does the formal independence of the youngest economic ERAs reflect with earlier established economic agencies. In other words, are we seeing a shift towards enhanced independence or a move away to lower degrees of independence.
This thesis aims to sketch a picture of the formal independence of the current 32 ERAs. By delving deeper into agency independence measurement an additional purpose of this study is to
critical reflect on how this concept is measured. This thesis will examine the indicators that scholars use in measuring agency independence, and if necessary will argue how adjustments may improve the operationalization of formal independence in general.
A revision of the formal independence of the current ERAs can enlighten the existing literature about European agencies in more ways. A couple of scholars namely found a relation between the formal independence and their defacto independence. (Yesilkagit & Van Thiel,p. 17; Hanretty & Koop, 2012; Egeberg & Trondal, 2009, p. 685). Therefore the amount of formal independence can indirectly say something about the influence of such an ERA in policy making. That being said, one must acknowledge that the relationship between formal independence and defacto independence is ambiguous and that formal independence alone does not explain defacto independence, this is only a sufficient condition if there are several veto players involved and if it is a relative old agency (Magetti, 2007, p. 282). Magetti (2007) adds that agencies can also have a high defacto independence even if its formal independence is rather low, if they are embedded in coordinated market economies. Furthermore a common reference point of the formal independence for all the ERAs can be useful in order to stimulate the normative debate about the role of agencies in the EU in more general terms. This contribution can add to normative judgements whether ERAs can indeed act independent enough, and whether their role within EU policy making is appropriate given their degree of independence.
The remainder of this thesis is build up in the following way. First of all I will present a theoretical framework wherein this research pays attention to the various waves of agencification, and wherein the reasons for agency creation are laid down. Furthermore this section provides the theoretical tools to establish two hypothesis that will be examined within this thesis. Part three consists of the methodology wherein this thesis provides more detailed information how the scoring index is build up and how I assigned the scores to the ERAs. The results section presents a table with information on the final scores for the 32 ERAs, explores for any explanations why some ERAs score higher than others as why some scores substantially differ from the ones Wonka & Rittberger (2011) assigned in 2008. Furthermore this section sets forth why I ought adjustments the scoring index necessary and does a first proposal how such a new index should look like. This alternative index will also be tested to examine whether the provisional answers to the hypothesis still hold. Section five consists of a discussion wherein the results of this paper are linked back to the existing literature and suggests any possible
directions for future studies in this area. Finally the conclusion will answer the research questions, provides a short summary and will repeat the most important findings.
2. Theoretical Framework
The introduction already revealed that making credible commitments is one of the main reasons to establish ERAs. Apart from this reason can one detect several other reasons that explain the establishment of (European) agencies. This paragraph will describe these in more detail before I eventually will derive two hypotheses from those theoretical insights. But first of all this thesis will shortly describe how the amount of ERAs has evolved over the past decades.
2.1 The three waves of agencification
In contrast to the delegation to regulatory authorities at the national level did this trend at the European level emerge quite recently. Although the first two ERAs were established in 1975, it took until the 1990’s that one can speak of a real kick-off of agencification within Europe. Within three major waves 1990-1994, 2000-2007 & 2009-2011 has to total amount of ERAs been increased to a total of 41. The Court of Auditors (2018, p. 7-9) differentiates the 41 in three categories; of which 32 of are categorised as regulatory decentralised agencies, six of them as executive agencies and three are classified as ‘other bodies’ (Court of Auditors, 2018, p. 7-9).
The latest wave of agencification kicked off in 2009, the wave consisted of seven ERAs of which a couple of them were founded as direct consequence of the Eurocrisis. One of the interesting features of this wave is that it contained a substantial amount of economic regulatory bodies in contrast to earlier waves of agencification. Interestingly because most ERAs before this third wave of agencification were acting in social domains. The latest wave of agencification resulted in the foundations of five ‘economic’ ERAs of which three of them form the European Banking authority (EBA); the European securities and Markets authority (ESMA) and the European Insurance and Occupational Pensions Authority (EIOPA). The two other ‘economic’ agencies in the latest wave are: the Agency for the cooperation of Energy Regulations (ACER) and the office of the body of European regulators for electronic communications (BEREC). Finally one can detect two social ERAs in the latest wave; the European asylum support office (EASO) and the agency for the operational management of large scale IT systems (EU-lisa) (Richardson & Mazey, 2015, p. 240).
the creation of such institutions; functional reasons and political reasons (Groenleer, 2009, p. 100). The functional motives for the creation of decentralised public bodies are about specific benefits that such an independent body can have over more classical forms of governments (Groenleer, 2009), in serving a particular task, whilst political reasons explain agency creation more as the outcome in conflicting supranational and intergovernmental interests (Groenleer, 2009; Kelemen, 2002; Kelemen & Tarrant 2011).
2.2 Functional reasons for agency creation
Groenleer (2009) has done extensive research in the functional reasons for agency creation and distinguishes several of them. One of the most important functional reasons for agency creation is that agencies are foremost able to take independent decisions of highly technical and scientific nature (Vibert, 2007), decisions that the Commission is less likelier to accomplish. On the one hand because of a lack of adequate knowledge and expertise within the Commission, on the other hand because decisions are not influenced by electoral incentives, crisis situation or contingencies. Agencies are insulated from the political process (Majone 1997b, 2002) therefore are ought to be better able to fulfil longer-term objectives within a policy domain (Groenleer, 2009, p. 101). An additional advantage of the delegation process is that it signals that a chosen policy path is warranted (Majone, 2000).
Other functional reasons for agency creation can be found in terms of increased levels of transparency and rising levels of accountability. Scholars have considered the democratic process where the EP, EC and Council are involved as one of the least transparent processes in the democratic world (Shapiro, 1997, p. 291). The procedure whereby European agencies become responsible in terms of explaining and justifying policy choices can therefore enhance the transparency of European decision making (Groenleer, 2009, p. 110). Additionally, ERAs are not located in Brussels but are spread all over Europe. Almost every member state currently hosts one ERA. The geographic dispersion of the ERAs can contribute to bring policy closer to the people and so “may meet the demand for public accountability y, and thereby increasing the legitimacy of EU policy making” (Groenleer, 2009, p.110).
Before agencies were created it was argued that the European legislative procedure was not sufficiently able to incorporate the views of relevant stakeholders and other relevant third parties in the decision making process (Groenleer, 2009, p. 110). Decentralised bodies are ought to be better able to set up these formal and informal networks as to assure that their interest are
represented within policy making. A clear example of this incorporation of stakeholders can for example be found within the new ESA. In the three agencies that represent the ESA are stakeholders incorporated within the management boards, so to assure that their views are represented.
The fifth functional reason for agency creation can be found in the capacity to contribute to the consistent implementation of Community law. These contributions can especially be valuable in order to reduce the informational transaction costs. These transaction costs arise when ‘principals are confronted with a complex policy environment and often require technical information and expert advice in order to craft effective public policies (Pollack, 2006, p. 168). With the help of the networks and more specialized knowledge within the decentralised agencies can the implementing capacities for the union as a whole be improved (Groenleer, 2009, p.110).
As informational transaction costs can be reduced so can other transaction costs be reduced as well. If the setup of ERAs turn out to be successful and contribute to efficient implementation process that enhances cooperation between member states and the Commission, one ERA may eventually replace all the national agencies. It would contribute to a shrink in bureaucratic costs, as well as a reduction in costs that citizens, ngo’s and companies have to make since they now have to deal with only one agency (Groenleer, 2009, p 111).
The final functional reason for agency creation is one that also has an economic character. The capabilities and networks that agencies possess make that they are able to diffuse regulatory practices that aim to enhance the proper functioning of the internal markets. Groenleer (2009, p.111) ascertains that past events such as the enlargement of the EU and privatisation exposed the need for a solid functioning of the internal market. Especially in light of the past Eurocris can agencies in cooperation with its networks play a vital role in the diffusion of best practices that aim to an optimal functioning of the internal market.
2.3 Political reasons for agency creation
Next to the functional reasons for agency creation are there also various political motives that explain agency creation. Among the most influential scholars are the contributions from Keleman & Tarrant (2011) and from Thatcher (2011). Whilst Keleman & Tarrant (2011) explain agency creation via a rationalist institutionalist perspective, Thatcher (2011) uses a
historical institutional approach to account for political motives that determine agency creation.
Kelemen & Tarrant (2011) describe the agencification process as the results of a political battle between the Commission, the Parliament and the member states that all seek to expand their influence. Kelemen & Tarrant (2011) suppose that the relevant parties act rationally and aim to maximize utility given the impediments of the institutional frameworks. Kelemen & Tarrant (2011, p. 925) argue that “inter institutional battles over regulatory authority and conflicts over the distributional effects of regulation play a decisive role in the design of EU regulatory bodies”. Therefore the authors conclude that the establishment of ERAs can not only be justified by functional reasons, but should more be seen in a wider context of conflicting supranational and intergovernmental interests which play a decisive role in the actual design of ERAs (Kelemen & Tarrant, p. 943).
From a more historical institutionalist perspective adds Thatcher (2011) that the decision to establish an ERA is also conditioned by prior delegation choices to other (independent) institutions. For instance, prior delegations decisions to either the European Union, national agencies or European networks influence why ERAs are created in a specific area, and may also explain why current networks or national agencies still reign in a particular area (Thatcher, 2011, p. 795). A strong institutional network for example may prevent the creation of an ERA, as these networks are able to exert sufficient amount of influence in a policy area, and enjoy their reputation as the one body that should be responsible for making regulations in a policy domain (Thatcher, 2011, p. 805). For the same reasons are ERAs creation conditioned by the strength and power of national agencies.
Among other political reasons for the creation of ERAs can one find decisions taken by the incumbent cabinet. De Figueiredo (2002) found empirical evidence that showed that cabinets have an incentive to lock in their status quo trough the establishment of regulatory agencies, as it impedes future cabinets to easily reverse policy making. Furthermore, political ideology seems to affect agency creation in a way that populist governments are more eager to create agencies than other cabinets (Ennser‐Jedenastik, 2016).
Finally Groenleer (2009) has two additional political reasons for agency creation. First of all considers Groenleer (2009, p. 106) the hosting of an agency as an prestigious project. Groenleer found evidence that the ERAs have been set up in the past in order to distribute them equally
over Europe, so that every European country can claim that they host an ERA, without a real functional purpose of the ERA in question (Groenleer, 2009, p. 106). Secondly has the establishment of ERAs been proven to be the adequate response in order to restore the credibility of the Commission and restore the confidence in the functioning of the Community (Groenleer, 2009, p. 103). Politicians have proven willingness to tackle urgent crisis situation at EU level by establishing European agencies where national administrative structures were not able to respond appropriately. The establishment of the EFSA after the BSE crisis in 1996 and more recently the creation of the ESA in 2011 are two concrete examples of those.
2.4 Hypotheses
ERAs have the special feature that they operate insulated from politics. The issue of credible commitment seemed one of the principal reasons of why ERAs enjoy this degree of independence. The concept of credible commitment will be the basis for one of the hypothesis, as such I will describe this concept a little bit more in detail.
The issue of credible commitment derives from a classic normative discussion about the role of the government in its monetary policy. The normative debate dealt with the question whether the government should adjust its economic policies because of contingencies, or that it by any time should stick to its pre-announced longer term monetary objectives. In other words should governments enjoy a certain amount of discretion, or should it work rule based (Majone, 1997b, p. 143).
People that resist government discretion argue that discretion through time inconsistency has an adverse effect on policy continuity. Time inconsistency occurs when politicians at one particular point in time use their discretion to adjust the status quo, but that appeared most optimal solution at that time is not consistent with the most efficient solution in a later stage. It is not because of two further reasons. First there is a risk of the so called time-leg, so that the amendment of the status quo takes too long which implicates that the governmental response is actually too late. Secondly governments run the risk that citizens anticipate for the new policy and therefore adjust their behaviour that prevent policy makers from fulfilling their initial intensions (Majone, 1997b, p.143-144; Gilardi, 2002, p. 825).
government discretion. Elections give politicians a short time horizon and incentives to focus on politically popular policies. These opportunistic incentives open room for discretion, which is not always particularly aimed at fulfilling longer term policy objectives (Majone, 1997b, p.145). Another consequence of having elections is that it can undermine regulatory credibility, as majorities are temporarily and a successional cabinet may change direction of a chosen path (Shepsle, 1991).
In order to overcome the above stated incentives politicians may decide to delegate powers to independent decentralised bodies so that their actual discretion will be constrained and that the chances to commit one selves to longer term objectives is much higher (Gilardi, 2002, p. 875). The capacity of these independent bodies to commit themselves to a chosen path is higher, because they are not exposed to elections/majoritarian incentives. These regulatory agencies have the chance to operate at the background of the political spectrum without the constant pressure to publically score. It is therefore more likely that ERAs can work on longer-term objectives as when politicians would have had a say in a particular policy area. Furthermore politicians provide a clear signal to potential investors that a chosen policy direction is assured, and that their potential investments are not vulnerable to electoral / majoritarian motives.
There is a theoretical difference in how this idea of credible commitment should apply to either economic regulatory authorities as to social ones. The whole purpose of economic regulatory authorities is to provide the conditions so that a market can function efficiently, which basically means providing confidence to potential investors that policies are not easily reversed. According to Gilardi (2005a, p. 143) are “the sunk costs that characterize investment in utilities mean that this confidence is even more important in this context, because the losses associated to an adverse regulatory change are more severe”. Gilardi (2005a, p. 143) adds “that the lack of credible commitment capacity has particularly negative consequences when the goal is attracting investment”, especially the case in economic regulation, whereas ERAs operating in the social domain are way more concerned with the safety and protection of its consumers. Therefore I hypothesize the following:
European agencies operating in economic areas are formally more independent than ERAs operating in the social domain.
Gilardi (2005a) has performed multiple analyses wherein he has measured the formal independence of a substantive set of national regulatory agencies. Groenleer (2009) has researched how often and in which passages references of independence have been made in the regulations of the ERA’s. Wonka & Rittberger (2011) have used the same indicators as Gilardi (2005a) to compute a formal independence score for all the ERA in existence up to 2008. The outcomes of the studies ascertain that the formal independence of the ERAs differs extensively. Groenleer (2009, p. 118) notes for example that ERA’s that formally state their independence within the regulations are more independent than ERA’s wherein this is only named in between the lines of the regulations, or when there is not an single mention in the regulation at all. Above all, Wonka & Rittberger (2011) found severe variance in the formal independence scores of the EU agencies at the time, varying from 0.21 to 0.64 (Wonka & Rittberger, 2011, p. 731-732).
The historical institutionalist contribution from Thatcher (2011) could explain why until the third wave of agencification one could observe a modest amount of ‘economic’ ERAs. The intergovernmental stakes were presumably higher than the supranational interest of the European Union, as such preferred the member states to have more control over these economic sectors. However, at the same time described Kelemen & Tarrant (2011) that continuous distributional battles between those intergovernmental and supranational powers take place, wherein contingencies and focus events may change these proportions and new delegation decisions can suddenly become available. Around 2011 such a focus event took place; the Eurocrisis. A direct consequence of the Eurocrisis was that three economic ERAs were established within the area of the financial sector (Schimmelfennig, 2018, p. 980). Apparently a crisis had to occur before supranational interests became superior over national interests. The member states realized that the delegation to ERAs was no longer avoidable. The ESA and the three agencies that belonged to it were granted more extensive powers (Busuioc, 2013, p. 15), because the member states realized that this was the only way how a future crisis may be prevented. Given the severe context in which the ESA have been established it seems essentially vital as to assure the highest amount of credibility. Therefore I suppose that these three economic ERAs will be formally more independent than other ‘economic’ ERAs.
The latest economic agencies will be formally more independent than the other ERAs operating in the economic domain.
2.5 Key concepts
If one refers to the independence if ERAs one is likely to speak about either formal independence or defacto independence, in terms of the goal of this research it is essential to make a distinction between those. Whereas formal independence refers to the statutory provisions that place the ERAs outside the control of the political bureaucracy, defacto independence is about the extent of political interference in the decision making and execution phase when exercising its regulatory competences. This thesis will make use the conceptualization of Hanretty & Knoops (2013) and Magetti (2007). Formal independence is defined as: “the statutory provisions that provide the agencies which components it can exercise including the competences that can be exercised isolated from political interference” (Hanretty & Koop, 2013, p. 197). Next, defacto independence is defined as: “the self-determination of agencies’ preferences, and their autonomy throughout the use of regulatory competencies, that is, during the activity of regulation” (Magetti, 2007, p. 272).
In order to speak of a wave of agencification, this thesis assumes that it refers to the creation of a relatively substantial amount of agencies within a short period of time. When this thesis refers to the first wave of agencification it refers to the ERAs established between 1990-1994, speaking of the second wave of agencification in this thesis imply the ERAs founded between 2000-2007, whereas the latest wave of agencification refer to the ERAs established between 2009-2013. Consequently I do not consider the establishment of the ERAs in 1975 as a separate wave.
3. Research design & methodology
The overall goal of the research is both descriptive and explanatory. This thesis aims to update and revise the formal independence scores for all the current ERAs in descriptive terms for the year 2019. Besides the descriptive aspect is it concerned what the effect of the latest wave of agencification on the independence of ERAs. Can one see a general trend towards higher scores of independence, or does one witness the contrary. Next, this thesis is interested whether the scores found in this thesis match those in the study of Wonka & Rittberger (2011) and aims to explain divergent patterns. Furthermore it is interested in the differences between the scores of economic and social regulatory agencies, and where possible explain their variances.
independence of the current 32 ERAs. The set is used by Wonka & Rittberger (2010) which draw on the earlier study of Gilardi (2005a), wherein the formal scores of a variety of national regulatory agencies in Western-Europe are compared. The index consist of 27 indictors separated in four components; the formal mandate of the agency, regulations in terms of how the agency head is appointed, regulation on how the members of the management board are appointed and regulations related to the (internal) decision-making of the agency. Each component consist of around eight separate indicators, apart from the first category. The calculation of the scores takes place by dividing the score of each separate component by four. The reasons that although the first component is only measured by one indicator but still counts equally as the other components is one taken by Gilardi (2005a) and Wonka & Rittberger (2010, p. 739), because they asses that component as “the most significant precondition for an agency to act independently”.
In order to give an idea how one indicator is set up I will provide an example. Take variable two that provide the possible scoring options for the appointment of an agency head (Wonka & Rittberger, p. 750). An ERA will be scored an 1.00 (maximum score) if the agency head is appointed for longer than 8 years. An ERA that provides the agency head with a term varying from 6 to 8 years will be scored a 0.8, if the agency head can be appointed for 4 to 6 years it will be scored a 0.6, for 2 to 4 years it will be scored a 0.4, for 0-2 years a 0.2, for no fixed term it will be scored a 0. For a detailed overview for how all the variables are build up, I refer to appendix I of this thesis.
This thesis is interested in the 32 decentralised regulatory agencies, according to a list of the Court of Auditors updated until 2018 (Court of Auditors, 2018). This means that the executive agencies (6), and the agencies that are labelled as ‘others’ (3) by the court of Auditors will not be included in this thesis. This decisions is being taken as to enhance the comparability of the agencies, since the decentralised regulatory agencies in comparison to the executive agencies form a permanent feature of the architectures of the European Union. These executive agencies were also excluded in the study of Wonka & Rittberger (2011). The information to score the ERAs will mostly be derived from investigating their consolidated constitutional mandates. In cases that its mandate will not provide any information with regard to one of the 27 indicators other documental sources or rapports will be consulted. Difficulties in matching the parameters with one of the values in the index are discussed, explained and justified.
In order to avoid any bias and actually be able to compare the results anno 2019 with 2008 this thesis will score the agencies in the same way as Wonka & Rittberger (2011) did. Although the information that Wonka & Rittberger provide how they actually performed the scoring is not extensively, this thesis aims to replicate their work as precise as possible.
Some coding principles of Wonka & Rittberger are of particular interest. Wonka & Rittberger (2011) ascertain the ERAs can have three different principles; the Commisison, the Parliament or the Council. Wonka & Rittberger (2011, p. 739) assign equal weight to the influence of one of those institutions in for example establishing the budget (V27), reporting obligations with regard to the annual report or the annual working programme (V18, V19) and the agency discretion the decide on activities (V21). Equal weight in this context means that it does not matter whether it is the Council, Commission or Parliament that has a formal say in the activities, nor which institution can formally intervene ‘stronger’. Moreover, Wonka & Rittberger (2011) assume that if multiple institutions are involved in for example the appointment of the director (V3), it would increase the independence of the agency. It is assumed that having multiple principles would open opportunities to play those principles off against each other, since these principles are likely to have diverging interests that make it harder for them to agree on issues as the approval of the financial resources, or adjustments in the formal mandate (Wonka & Rittberger, 2011, p. 739; Tsebelis 2002, chapter 10).
The first indicator with respect to the actual competence of the EU agency is one that accounts for 0.25 of the final formal independence score of an ERA. It is the only component that consists of just one variable and is therefore the variable with the most significant impact on the final score. This parameter is related to whether the agency can actually take binding decisions (1), can prepare proposals for decisions (0.5) or can mere issues opinions and information (0). If one takes a superficial look at a random mandate of one of the 32 relevant agencies, one is likely to find an agency that at least performs two of those tasks. In these cases I have classified the agency based on the strongest tasks it can perform. Still, in some minor cases was the mandate of an ERA difficult to interpret with respect to this variable. In those circumstances I relied on a contribution of Busuioc (2013), as additional source to score the agencies (V1). In cases that I ought multiple classifications possible, I verified whether this ERA was classified as one with (quasi) rule making powers by Busuioc (2013, p. 40). If this was the case I assigned the maximum score (1) to the ERA in question.
In order to make any comparisons with respect to the field of operation of the ERAs another important aspect is to classify them as such. Whereas Wonka & Rittberger (2011) differentiated between informational executive bodies, social regulatory bodies and economic regulatory bodies, this thesis prefers to differentiate between ERAs functioning in the social or economic field. This decision is foremost taken for practical reasons. Wonka & Rittberger (2011) had a slight other purpose than this thesis and were in first instance concerned if European agencies with regulatory competences are formally more independent than ones with informational or executive competences. In this thesis however I am particularly interested whether ERAs operating in the economic field are formally more independent than ERAs performing in the social field. It does not matter whether these agencies have informational capacities or regulatory ones, what matters is how the formal independence score of both groups differ.
To make a distinction between ‘social’ and ‘economic’ European agencies, this thesis relies on the classification that has been presented in the article from Thatcher (2011, p. 797). Although the difference between social activities and economic activities is one that is at least debatable, social ones are primary concerned with protecting the health and safety of European consumers, whereas the economic ones are concerned with matters of competition, “such as prices, unfair competition or obstacles to market access” (Thatcher, 2011, p. 796). However does this thesis not agree with two specific classifications of Thatcher (2011); the European railway agency and the Community plant variety office (CPVO). Both ERAs are classified as ones operating in the social field, whilst this thesis classifies them as economic ones.
I have classified the CPVO and the European railway agency as ‘economic’ for the following reasons. First with regard to the CPVO I have looked at the actual purpose of the institution. The purpose of the CPVO is to grant propriety rights for plant varieties, and is therefore “directly involved in defining the terms and conditions of the internal market” (Chamon, 2016, p. 24). This purpose is more likely to be associated with ‘matters of competition’ than with protecting consumers, therefore is the CPVO categorized as ‘economic’. In the case of the railway agency, it has the following purpose (Article 2):
The purpose of the railway agency is a concrete example why the distinction between social agencies and economic agencies is not clear cut, since its purpose discloses that it is actually acting in two domains. However, I regard this as an economic agency since the emphasis of its purpose in my view hints more in the direction towards matters of competition than protecting the safety of European railway consumers. For the above reasons are there also quite a lot of academic contributions concerned with this theme, contributions that display how difficult it is to assign the ERAs into different categories (Chamon, 2016, p. 20-28).
Additionally, I have classified the agencies that form part of the third wave of agencification in the following way: EBA, EIOPA, ESMA, ACER, BEREC-Office as agencies operating in the economic field, and EU-Lisa and the EASO as social agencies (Chamon, 2016, p. 22; Busuioc, 2013, p. 38). For a complete overview of the classification of the 32 ERAs that are relevant for this thesis I refer to table 1.
4. Analysis
In this section of the thesis is room for an extensive discussion of the final formal independence scores of the 32 ERAs. This section is foremost interested how the scores of the agencies are distributed in 2019, and in particular probes for possible explanations why some ERAs score ‘higher’ than others. Furthermore is this section concerned how the scores in this year vary from the scores in 2008, as such will be made a comparison with the study of Wonka & Rittberger (2011). Obviously, agencies that were established after this period cannot be compared with the scores in 2008. This thesis will probe for possible explanations to account for substantial variances in agency scores. Next I will present some regression models that can examine the explanatory power of the hypotheses, and provide an answer to them. Additionally, this thesis will present a second model where I will process my own critique as that from Hanretty & Koop (2012) into an alternative way of measuring the formal independence of ERAs.
The objective of the Agency shall be to contribute to the further development and effective functioning of a single European railway area without frontiers, by guaranteeing a high level of railway safety and interoperability, while improving the competitive position of the railway sector. (Commission, 2016, p. 8)
4.1 Formal independence score of the 32 ERAs
Table one presents an overview of the formal independence score of the 32 ERAs for the year 2019. Besides information on the final formal independence scores (in bold) finds one details about the field of operation, its budget & its staff size and which score Wonka & Rittberger assigned in 2008. The formal independence scores are distributed between 0 and 1, a total score of 1.00 denotes that an agency is completely independent, whereas a score of 0 indicates that an agency is totally dependent. The scores of the 32 ERAs are listed in the seventh column of table 1, ascending from the ERA with the lowest score to the one with the highest. The formal dependence scores vary between 0,27 and 0,70, which implicates that none of the ERAs is for one hundred percent operating independently nor operates one of them completely dependently. Among the most dependent ERAs are; The monitoring centre for drugs and drug addiction, The European environment agency and The translation centre for the bodies of the European Union (CDT), these ERAs score just below the 0.3. The most independent agencies are The European Chemicals agency, The European securities and markets authority and the European insurance and occupational pensions authority, these agencies score around a 0.7.
Interestingly possess none of the agencies that were included in the study of Wonka & Rittberger as in this thesis an equal score. This is unlikely explained by changes in mandates, since for quite a couple of agencies, especially the ones that formed part of the first wave of agencification have its mandates not been changed for years. In some cases does this thesis even rely on the same mandate as Wonka & Rittberger (2011). It is probable that these differences originated in the coding phase. Later on in my thesis will I delve into more detail of how these differences may have occurred, wherein I also express my critique on the scoring index.
If one assumes that the minor differences between the scores for the year 2019 and the study from Wonka & Rittberger (2011) have occurred in the coding phase, there are still of couple of differences that stand out. Among the greatest ascenders one can find the European medicines agencies (EMA) and the European Union agency for safety and health at work (EU-OSHA). Among the biggest descenders can one indicates the European training foundation and CdT, meanwhile possess the European foundation for the improvement of living and working conditions (EUROFOUND) an almost similar descending score. The following paragraphs will zoom in to those differences, and may offer various explanations for the
The EMA was founded in 2004, since then its mandate have been adjusted eight times. One of the major changes that the agency has gone through occurred in 2006 when it was granted individual decision taking powers on medicinal products for paediatric use. Since the mandate of the EMA altered at the end of 2006, the Commission will not authorise an market application for paediatric use anymore, if the EMA has not approved such an application (Commission, 2006). A change in regulation was deemed necessary for paediatric medicines since market forces “have proven insufficient to stimulate adequate research” (Commission, 2006, p1). The absence of extensive studies of high quality and pre-testing in the area of paediatric medicines led to increased risks of adverse reaction if consumers would use such medicines. The Commission therefore decided to interfere and granted the EMA with autonomous decision making powers. The EMA is now responsible for executing extensive studies and pre-testing to assure the quality of the medicines within the field of paediatric use. Although the agency has solely been allowed to take individual decisions in a small area, I classified the agency in this case based on the strongest task it can perform, thus as a decision making authority (1). Next to some minor changes in the appointment of the management board can this explain the relatively high score in my table in comparison to the one Wonka & Rittberger (2011) assigned for. It is likely that Wonka & Rittberger classified this agency in their time period as an one that can only issue information and opinions (V1).
The increase in the formal independence score for EU-OSHA can also be explained by a change in its mandate. Recently have the mandates of EU-OSHA, Eurofound and Cedefop been revised after a Council decision at the end of 2016. These agencies have a tripartite nature operating in the social field aiming to improve the living and working conditions for all European citizens. The new regulations came in place after a call to be better able to respond to societal, institutional and economical needs. The new regulations feature foremost new governmental structures, as well as conflicts of interest policies and anti-fraud measures. The mandate of EU-OSHA have been expanded from seven pages to sixteen. For EU-EU-OSHA specifically is the role how it can support the Commission much more defined than in the latest mandate. Overall seem these developments be the explanation for the increase in formal independence score. Ironically do all the tree mandates currently appear very similar, and do small details explain their differences in scores (European Council, 2018).
Among the biggest descender are the ETF and the GSA. The ETF was founded in 1990 and has seen its mandate substantially been revised in 2009. The revisions however were not directly
related to any of the variables in the scoring index, these were mere changes in the actual activities that the ETF performs (Commission, 2008, p. 82). It is likely that Wonka & Rittberger as this thesis relied on the same provisions in scoring this agency, interestingly however is then the extensive divergence in score. I have had no access to the exact scoring table of Wonka & Rittberger of this particular agency, therefore I cannot designate where these differences may have occurred. However I have provided further details how I came to the final score of this agency in appendix II of this thesis.
For the same reasons I cannot explain the drop in score of the CdT. For an agency as Eurofound however can the recent change in its mandate explain why I assigned it a lower score than Wonka & Rittberg did. As earlier explained has this agency seen its mandate revised in response to societal demands. These adjustments were not entirely focussed on the actual activities and purposes of the agency, but had also to do with its organisational structure. These adjustment can have had an ‘negative’ impact on their final score.
Eight new entries are included in table 1, of which seven of them belong to the third wave of agencification. The European institute for gender equality was not incorporated by Wonka & Rittberger, however categorized as ERA by the Court of Auditors, and therefore included in this thesis. Of the ERAs that belong to the third wave of agencification do the economic agencies that are part of the ESA stand out. The EBA, EIOPA and the ESMA occupy respectively position 29, 30 and 31 in the final list. Contemporarily they are one of the most formal independent ERAs. Their high degree of formal independence can be explained in light of the Eurocrisis in 2011. The crisis exposed how vulnerable a potential breakdown of one Eurozone country can be for the whole Eurozone. Especially northern European politicians were rather reluctant to delegate the competences of the final sector to supranational institutions before the crisis, however they recognised that the crisis situation as the one in Greece could have had serious financial consequences on their own home countries as well. That recognition opened political room to talk about the establishment of new supranational institutions. (Schimmelfennig, 2018, p. 977). At the same declared the European central bank to do ‘whatever it takes’ to restore the credibility of the Euro, a signal that was positively acknowledged by financial market actors and really calmed the market downs. Moreover the ECB even facilitated the establishment of the ESA (Schimmelfennig, 2018, p. 985). The actions of the Euroleaders and the ECB really restored the credibility of the Eurozone. At the same time did the politicians realise that in order to prevent a similar crisis in the future that they had to
delegate severe supervisory powers to the institutions that they established, what potentially can explain the high formal independence score of the EBA, EIOPA and ESMA.
# Agency name Year Classifi
cation* Budget (mil)+ Staff size + Score Score W&R (2011) Delta
1 Monitoring centre for drugs and drug addiction (EMCDDA)
2006 S 16 111 0,27 0,23 +0,04
2 European Environment agency (EEA) 1990 S 70 212 0,28 0,21 +0,07 3 Translation centre for the bodies of the
European Union (CdT)
1994 S 49 215 0,29 0,46 -0,17
4 European union agency for law enforcement training (CEPOL)
2002 S 9 53 0,30 0,29 +0,01
5 European Union agency for law enforcement (EUROPOL)
2002 S 118 804 0,31 0,26 +0,05
6 European Union’s Judicial Cooperation Unit (EUROJUST)
2002 S 48 242 0,33 0,31 +0,02
7 European Union Agency for the Operational Management of Large-Scale IT Systems in the Area of Freedom, Security and Justice (EU-LISA) ***
2011 S 79 152 0,33 na na
8 European institute for gender equality (EIGE) **
2006 S 8 45 0,34 na na
9 European Union agency for network and information security (ENISA)
2004 S 11 70 0,35 0,40 -0,05
10 European fisheries control agency (EFCA)
2005 S 17 72 0,36 0,31 +0,05
11 European Asylum Support Office (EASO) ***
2010 S 79 200 0,37 na na
12 European Foundation for the
Improvement of Living and Working Conditions (EUROFOUND)
1975 S 20 100 0,38 0,54 -0,16
13 European Centre for the Development of Vocational Training (CEDEFOP)
1975 S 18 117 0,39 0,45 -0,06
14 European union agency for safety and health at work (EU-OSHA)
2004 S 15 64 0,40 0,21 +0,19
15 European global navigation satellite system agency (GSA)
2004 S 704 166 0,41 0,56 -0,15
16 European border and coast guard agency (FRONTEX)
2004 S 281 526 0,42 0,45 -0,03
17 European Centre for Disease Prevention and Control (ECDC)
2004 S 58 279 0,43 0,60 -0,17
18 Office of the Body of European Regulators for Electronic
Communications (BEREC office) ***
2009 E - - 0,45 na na
19 European maritime safety agency (EMSA)
2002 S 79 250 0,45 0,59 -0,14
21 European Union agency for fundamental rights (FRA)
2007 S 23 108 0,48 0,49 -0,01
22 Community Plant Variety Office (CPVO)
1994 E 16 51 0,55 0,50 +0,05
23 European food safety authority (EFSA)
2002 S 81 443 0,56 0,62 -0,06
24 European Aviation safety agency (EASA)
2002 S 192 771 0,58 0,59 -0,01
25 Agency for the Cooperation of Energy Regulators (ACER)***
2009 E 13 91 0,61 na na
26 European Intellectual Property Office (EUIPO)
1993 E 401 995 0,62 0,54 +0,08
27 European Medicines Agency (EMA) 2004 S 331 766 0,63 0,33 +0,30 28 European Union Agency for Railways
(ERA)
2004 E 31 164 0,63 0,51 +0,12
29 European Banking Authority (EBA) ***
2010 E 38 190 0,67 na na
30 European Insurance and Occupational Pensions Authority (EIOPA) ***
2010 E 24 151 0,67 na na
31 European Securities and Markets Authority (ESMA) ***
2010 E 42 226 0,68 na na
32 European Chemicals Agency (ECHA) 2006 S 112 563 0,70 0,62 +0,08 *Classification in which policy field an ERA operates, based on contributions from Thatcher (2011, p.797), Busuioc (2013,
p. 39-41) & Chamon (2016, p.22-24).
**Wonka & Rittberger (2011) did not incorporate this agency within their list of 29 ERAs
*** Agencies that were established during third wave of agencification, therefore not accounted by Wonka & Rittberger (2011)
+ Information derived from the Court of Auditors (2017)
Table 1. Formal independence score of the 32 ERAs in the year 2019 in comparison to the scores assigned by Wonka & Rittberger (2011, p. 731-732).
A second table aims to make a comparison how the formal independence in general seems to fluctuate over time, by providing an average score of the ERAs present-day as the ones in 2008. I have done this comparison by selecting the 24 ERAs that overlapped in this study as the study of Wonka & Rittberger (2011). Wonka & Rittberger (2011, p.731) incorporated five other agencies that were not used in my study1, because in this thesis I have incorporated only the bodies that are assigned as decentralised regulatory agencies by the Court of Auditors. Within the academic world there is no consensus what the exact definition of an ERA is (Chamon 2016, p. 9), and are scholars inclined to embrace a definition what they ought the be the most plausible for the purpose of their study. As such Wonka & Rittberger have used a broader definition what can explain why they incorporated some ERAs that are not accounted for in this thesis.
1 Wonka & Rittberger (2011) incorporated the European defence agency, European Union satellite centre,
European Union institute for security studies, European joint undertaking for ITER and the development of fusion energy and the European agency for reconstruction in their paper, which I do not account for.
A closer look at table two displays a minor difference in the average score for the 24 ERAs in this year as for 2008. On average are the 24 ERAs a fraction less independent today than in the year 2008. However one should note that the agencies established in the latest wave of agencification are on average more independent than the 24 agencies founded in earlier waves, these agencies have an average formal independence score of 0,54. This is likely to be explained by the high scores for the agencies that are part of the ESA. From these details is one not able to say that ERAs are becoming more independent, since there is no other reference point for the latest group than this thesis. The future will tell whether the signals that hints towards higher degrees of formal independence for recently established ERAs are correct.
Relevant wave N 2019 2008
Overlapping 24 ERAs in first and second wave 24 0,44 0,45
First wave of agencification 7 0,42 0,48
Second wave of agencification 17 0,45 0,43
Third wave of agencification 7 0,54 na
Table 2. Comparison of the average scores of the 24 overlapping ERA in this thesis as the ones in Wonka & Rittberger (2011)
4.2 Answering the hypotheses
The previous section presented first signs that economic ERAs possess higher formal independent scores than ERAs operating in the social field. Furthermore are there signals that one is witnessing a tendency that the formal independence is increasing over time. However both observations are highly influenced by the establishment of the ESA. The three ‘economic’ agencies that form the ESA, founded during the latest wave of agencification are all among the most formal independent ERAs today.
First of all I have performed an independent sample t-test to see whether an significant difference exist in the average formal independent scores of the ERAs operating in the economic domain, versus the ERAs operating in the social field. Table one presents an overview in which field I have categorized the agency. In total are 24 ERAs classified as ones operating in the social field, whereas eight ERAs are operating in the economic domain. The results can be viewed in table 3.
N Average Score Standard deviation T value
Social ERAs 24 0,41 0,11
Economic ERAs 8 0,6 0,08
-4,655 * *Significant at the 0.05 level
Table 3. The difference of the average formal independent score of economic and social ERAs.
The table provides evidence that a significant difference between the two groups exist. According to this test are economic ERAs formally more independent than social ERAs. Though, one should interpret this result with caution, since this test only checks whether a significant variation exists in the average score, at the same time does this test not account for possibly other reasons that disturb the relationship between the field of operation of an ERA and its formal independence score.
To examine if I could detect any confounders that may disturb that relationship I have created a multiple regression model. Moreover I have established this multivariate regression model to examine the explanatory power of the first hypothesis. It is important to assume however that I regard 32 ERAs that are incorporated in this model as a sample of a possible bigger population in the future. If I would consider the 32 ERAs as the full population, it would make no sense to make use of inferential statistics, since I am simply in the possession of the data of the entire population. In that case it would be sufficient to provide only the descriptive statistics.
The regression model consists of a couple control variables such as the year of foundation, the agencies’ staff size and their budgets. Table two disclosed signs that recently established ERAs enjoy higher formal independence scores than earlier ones, therefore I have added this control variable into the model. I have incorporated the staff size and budgets of the ERAs also in the model, not only because this information could easily be picked up by the Court of Auditors, but primary because I suppose that ERAs that consists of more personnel and possess a higher budget have been attributed these severe amount of resources to assure the credibility in these domains. I suppose that the European institution delegate more resources to ERAs in areas where they prioritize to be credible. I ought it therefore likelier that these ERAs may also possess higher independent scores.
The control variable year of foundation is incorporated in the model in the following manner. The first agencies that were founded in 1975 (Eurofound & Cedefop) have been attributed a
‘0’. The values assigned to the other agencies derive from their particular year of foundation minus ‘1975’. Information about the staff size and budgets of the 32 ERAs originate from the annual report of the Court of Auditors from 2017. The dependent variable in the first model is the formal independence score, as presented in table one. The independent variable for model one will be the field of operation, this dummy variable differentiates between agencies operating in the social domain or either the economic domain.
While keeping other factors constant economic ERAs have a an average formal independence score that is two tenths higher than social ERAs. This is a significant result at alpha level of 0.05 level. One can be with 95% certain that this results is likely to be found in possible increasing populations of ERAs as well. If one takes a closer look at the relevant control variables in the model, only its staff size seems to be significant. An increase of one staff member leads to an increase of formal independence score of 0,0002. Almost 50% of the variances in formal independence score is explained by the field of operation, the staff size and the budget of the ERAs. If all values in model 1 would have a value of 0, ERAs possess on average a score of 0,32. Based on the T-test and this regression model one can assume that economic ERAs are formally more independent than social ERAs, hypothesis one is therefore adopted.
Independent variable Model 1 Model 2
Operational field 0,2 (0,04)** 0,04 (0,03) Staff size 0,0002 (0,001)** -0,0001 (0,0007) Budget 0,0001 (0,001) 0,0007 (0,0012) Year 0,0016 (0,0020) 0,0091 (0,0165) N 32 32 Constant 0,32 (0,06)** 0,42 (0,05)** Adjusted R-square 0,49 0,06
Note: Unstandardized regression coefficients; standard errors in brackets ** Significant at 0.05 level
Table 4. Regression models aiming to explain formal agency independence.
To formulate an answer to hypothesis two this thesis has compared the average formal independent scores of the EIOPA, EMSA and EBA with the other five ‘economic’ ERAs. On average score the three ERAs that form the European supervisory authority one tenth ‘higher’ than the other group of economic ERAs. If I would assume this test as a sample of a potential bigger population, than can one be with 90% certain that the averages of both groups differ
significantly. However with such a small sample, and relatively high standard errors for especially the first group is this evidence statistically too weak to conclude that there is difference. On the other hand, if one regard this as a sample as the entire population there is sufficient evidence that the three ERAs that form the ESA are formally more independent than the other ‘economic’ ERAs.
N Average score Standard Error T-value
Other ‘economic’ ERAs 5 0,57 0,075
ESA (EIOPA, EMSA, EBA) 3 0,67 0,006
-2,26* * Significant at the alpha level of 0.10
Table 5. How formal independent scores differ for the economic ERAs
4.3 Critique on scoring index + alternative model
This final paragraph goes a little more in detail how the index that is used by Wonka & Rittberger (2011) is build up. In this paragraph I will highlight three critiques on the index and will try to come up with some ideas how one is able to improve the index, these ideas are backed by the literature. I have converted my suggestions into a new model of measuring the formal independence of ERAs, one that furthermore will be used to verify whether the answers to the hypothesis still hold.
In this thesis I already referred to the fact that the first variable of the index that aims to measure the formal independence of ERAs used by Wonka & Rittberger (2011) eventually determines 25% of the final score. The first variable distinguishes between agencies that can take binding decisions (1), prepare proposals for decisions (0,5) or can issue opinions and information (0). One can wonder whether this variable measures the formal independence of an ERA since this variable captures what the competences of ERAs are, but is not necessarily interested whether it can perform those competences with or without political interference. Hanretty & Knoops (2012, p. 202) add that “an agency may possess limited powers but exercise them independently; or it may possess a wide range of powers and exercise them with no independence”. Especially seen in the perspective of the substantial impact of the variable on the final score can one be critical in incorporating this model in the index, one can question if this variable is not mere concerned about the defacto independence of ERAs.
Another critique of the index is that for some variables the possible scoring answers are non-excludable, in other words one can attribute different scoring possibilities to the same parameter. Have a look for example at V21 (appendix 1) wherein one has to score the discretion
of ERAs to decide on activities. ERAs are scored an ‘1’ if it can decide autonomously on when to act, a 0.5 if either the governments, Commission or Parliament can ask the agency to act, and a 0.00 if either the governments, Commission or Parliament can commit the agency to act. In most cases it was likely that with regard to some provisions the agency decided autonomously how to act, whereas in other cases it was more likely that a Commission would request an ERA to act. The same goes for variable 20 & 27 for instance, wherein multiple scores could be given to the same agency. In terms of replication was it not always clear how Wonka & Rittberger (2011) approached this, something that eventually can also explain possible variances in agency scores.
Finally a minor critique of this index is related to variable 22. The variable deals with the external consultations practices of private actors by the ERAs. Agencies that do not feature any external consultations procedures are scored an ‘1’, whereas agencies that consult their external private partners before it drafts any proposals are scored a ‘0’. Furthermore are there three possible scores for the consultation practices that fall somewhere in between. My critique is focussed on the fact that most of the 32 ERAs in this study cannot take binding decisions and are mere concerned with the issuing opinions and disseminations of best practices. Therefore these agencies do not have to deal with external private actors and neither have any consultations practices. As ‘reward’ for being an informative agency, agencies are attributed the full score for this variable.
I have processed the above critique in a new index. The basis of the index is consistent with the index that Wonka & Rittberger (2011) used, but I have made a couple of adjustments. First of all I have left out the first variable about the competence of an ERA for two reasons. Firstly, as recently discussed can one be critical whether this parameter is really concerned with measuring formal independence. Secondly, because it is not easy to differentiate which ERAs possess (quasi) decision making powers. Chamon illustrates (2016, p. 21) how the academic contributions differ in their respective classifications, to demonstrate how complicated it is to make such a classification. Even in my own documental study I found out that there is certain room for interpretation in assigning which competences an ERA can perform. Overall spoken, can the incorporation of this variable lead to a higher change of potential bias if there is no consensus about a classification, especially if academics make classifications that solely seem to suit their specific purpose of study.
Furthermore I have left out V22 for the earlier stated critique. Finally I have adjusted V27 which is concerned with the resource dependency of the agency. The potential scoring options were in first instance non-excludable. I scored all the agencies eventually a ‘0’ since the budget is always to be approved by the European Parliament. It would not make sense to incorporate such a variable in the index since it does not really have an influence on the final score, if all the ERAs are scored the same. At the same time though, there is actual variance in how the ERAs are funded. Some ERAs are exclusively funded through its fees for the services they provide, whereas others are entirely financed from the contribution of the EU budget. To account for this political interference in the establishment of the budget, what says something about the formal independence of ERAs (Kruly 2013), I have adjusted the parameter. The altered parameter assumes that agencies that fund themselves are formally more independent than agencies that are subsidized through the EU budget (Kruly, 2013). I have attributed all the agencies with an ‘1’ minus the percentage that the EU contributes to the budget of the agency. For all the agencies I have used the budget of 2018 to define their score for this variable. Eventually have these adjustments resulted in a slightly new index, one than can be consulted in appendix III.
4.4 A different look on the hypotheses
The adjustments to the index have resulted in different scores for the 32 ERAs. These scores can be consulted in table six of this thesis. Instead of the former four components of the original index, consists this index of three components. Logically have the scores been computed by taking the sum of the average score of the three components divided by three.
Agency Score Agency Score Agency Score
EMCDD 0,32 ENISA 0,44 EU-OSHA 0,50
EEA 0,35 EASA 0,44 GSA 0,51
CdT 0,35 EFCA 0,45 EIOPA 0,53
Cepol 0,37 EASO 0,46 Frontex 0,53
CPVO 0,37 EMA 0,47 EBA 0,53
Europol 0,38 ACER 0,48 EMSA 0,55
Eurojust 0,40 EUROFOUND 0,48 BEREC-OFFICE 0,57
EU-LISA 0,41 EUIPO 0,48 ETF 0,58
EIGE 0,41 CEDEFOP 0,49 EFSA 0,58
EMSA 0,41 EU-OSHA 0,50 ECHA 0,58
ERailA* 0,50 FRA 0,60
*European Railway Authority
Table 6. Formal agency independence of the 32 ERAs (adjusted index).
The scores in table 6 are distributed more narrowly, although they still vary considerably. The highest score can be assigned to the FRA (0,6) whilst the lowest score is the one of the EMCDD
