Master's in Public Administration - Economics and Governance
Supervisor: Dr. Madalina Busuioc
Second reader: Dr. Dovile Rimkute
Juliette Dubois (s2006308)
10/01/2020
Why do EU agencies involve non-state stakeholders in their
decision-making process?
CHAPTER 1 – INTRODUCTION ...p.4 CHAPTER 2 – THEORETICAL FRAMEWORK ...p.8 I- EU agencies and non-state stakeholder involvement...p.8
1- The emergence of EU agencies...p.8
2- Non-state stakeholder involvement and its legitimacy challenge...p.10
II- Reasons for non-state stakeholder involvement...p.12
1- A range of reasons...p.12
2- A focus on agency needs...p.14 3- Hypotheses...p.15
CHAPTER 3 – METHOD...p.17 I- Operationalization: functional and symbolic involvement...p.17
1-The extent of involvement: from symbolic to functional...p.17 2- The link between the extent and reasons for involvement...p.21
II- The coding method...p.22
1- Codebook...p.22 2- Segmentation...p.24 3- Strengths and weaknesses of the chosen method...p.25
III- Case selection...p.27
1- The ESAs: an important case...p.27 2- The ESAs: making the link with EU agencies...p.28
CHAPTER 4 – ANALYSIS...p.30 I- Empirical findings...p.30
1- Reputation as primary reason for involvement...p.30 2- Slight differences between the ESAs and evolution through time...p.32
II- Theoretical implications...p.37
1- Comparing the results with Arras and Braun's research...p.37 2- Theoretical implications of cross-agency and cross-time differences...p.39
III- Social implications...p.40
1- A lack of input legitimacy...p.40 2- Further implications from a lack of input legitimacy...p.41
3- Implications for other EU agencies...p.41
CHAPTER 5 – CONCLUSION, RECOMMENDATIONS AND FUTURE RESEARCH
...p.43
CHAPTER 1 – INTRODUCTION
An increasing number of European Union (EU) agencies have been established in the last decades. Whereas the first EU agency was developed in 1975, the number has since then increased to a total of 31 agencies in 2019. This proliferation of EU agencies, with some having extensive powers (Busuioc, 2013), echo what Vibert (2009) names the "rise of the unelected". It refers to the increasing delegation of powers from elected officials to non-majoritarian institutions. The latter fulfil public functions while not being directly accountable to elected officials and voters. Several factors explain the "rise of the unelected" at EU level in the last decades. According to Majone (1999), the main driver of this phenomenon has been the increasing need for policy credibility. In a context of increasing interdependence between member states, growing policy complexity and privatisation, officials from EU member states have been in need to make credible commitments at domestic and EU levels. Delegation of powers to non-majoritarian institutions, such as EU agencies, brings an opportunity for EU elected officials to enhance their policy credibility. Contrary to these officials, non-majoritarian institutions are not subject to the electoral process. Therefore, they can commit to objectives that need to be achieved on the long-term.
Delegation of powers to EU agencies brings opportunities but also challenges (Majone, 1999). Their relatively independent nature favours policy consistency, public participation as well as flexibility in policy formulation and application that the government could not necessarily achieve otherwise. However, delegation also raises legitimacy concerns in case these independent bodies operate without strong accountability. Twenty years ago, Majone (1999) pointed out that EU non-majoritarian institutions lacked independence in practice and that their accountability mechanisms were not well developed. As of today, this issue persists and entails a lack of legitimacy of EU institutions (Schmidt, 2015). Majone (1999) suggests a set of solutions to improve accountability mechanisms in order to avoid such a "democratic deficit". Among others, he recommends that non-majoritarian institutions should be monitored via several mechanisms, including public participation. Non-state stakeholder involvement, as it is also called, should hence have the potential to improve legitimacy. The study of the reasons why EU agencies actually involve non-state stakeholders is tied to understanding the legitimacy challenge. The specific reasons why an EU agency involves non-state stakeholders, and its resulting conduct with such stakeholders, should determine the agency's degree of legitimacy. The academic literature has identified different reasons for non-state stakeholder involvement in regard to EU agencies. In line with the literature, this
research will assess the relative importance of expertise, organizational capacity and reputation as reasons for EU agencies to involve non-state stakeholders in their decision-making processes. To assess their relative importance, the minutes of the meetings that the three EU agencies had with their stakeholder bodies will be analysed in-depth through coding. This thesis aims to answer the following question:
Why do EU agencies involve non-state stakeholders in their decision-making process?
This research will first identify the extent of non-state stakeholder involvement in each of the minutes, by distinguishing between "symbolic" and "functional" involvement. A stakeholder is considered to be involved symbolically when he listens to staff merely as a spectator or when his input is disregarded or not commented on. This research will argue that symbolic involvement takes place when EU agencies involve non-state stakeholders only to maintain or improve their reputation. On the other hand, a stakeholder is deemed to be involved functionally when the agency staff specifically asks for his input or when this input is approved or included in the agency’s output. It will be argued that functional involvement takes place when EU agencies involve non-state stakeholders mostly to maintain or improve expertise and organizational capacity.
This research method has been chosen to deliver complementary insights to past research. The literature has identified two broad categories of reasons for involvement. On one side, there is external pressure on agencies, including the formal requirement from principals, pressures from the regulatees to get involved, and crises (Arras & Braun, 2018; Braun, 2012; Groenleer et al., 2014; Emerson, Nabatchi & Balogh, 2011). On the other, there is the reason of agency needs, which can follow functional or symbolic rationales (Arras & Braun, 2018; Groenleer, 2009; Wood, 2018; Carpenter & Krause, 2012; Carpenter, 2010; Busuioc & Lodge, 2016). This paper studies the second stream of reasons, as it enables the identification of the actual purposes for involvement once the structures are put in place. This relatively under-researched topic has previously been analysed by scholars through the spectrum of interviews, analysing the whole population of EU agencies (Arras & Braun, 2018). On the contrary, this research will codify meeting minutes, analysing a few case studies in-depth. The EU agencies selected for the analysis are the three European Supervisory Authorities (ESAs): the European Banking Authority (EBA), the European Securities and Markets Authority (ESMA) and the European Insurance and Occupational Pensions Authority (EIOPA). These agencies were picked because of their far-reaching supervisory and
regulatory powers. The degree of non-state stakeholder involvement entails important democratic implications, taken the important tasks of the ESAs. In addition, these three EU agencies are highly similar. This similarity enables to trace back diverging results between them to some of their specific differences. Furthermore, the EU financial institutions' failure to involve the civil society in their decision-making process has been argued to partly explain their lack of legitimacy (Schmidt, 2015). The study of the ESAs, themselves EU institutions operating in the field of finance, makes it possible to further test this finding. In case involvement would be identified as mostly symbolic, it would suggest a failure of consistent involvement of civil society in the decision-making process, in line with Schmidt's argument. In addition, these cases should be exemplary for EU agencies that share a similar structure of involvement. The three ESAs all involve non-state stakeholders in their decision-making process via stakeholder bodies. This similarity is of relevance to the research question, as past research highlighted that EU agencies with similar structures for involvement of non-state stakeholders shared comparable reasons for this involvement (Arras & Braun, 2018). This similarity, coupled with the significant data that will be gathered for three within-case studies, provides the opportunity to extrapolate, to a certain extent, to other EU agencies that involve non-state stakeholders in a similar way. Being able to extrapolate exclusively to EU agencies with stakeholder bodies is definitely a limiting factor. However, non-state stakeholders are deemed to have more potential for influence on decision-making via this structure than via public consultations (Ibid). It makes it a relevant choice for focus, as social implications should be more important where the stake of involvement is bigger.
As for the research period, this thesis will look at the reasons for involvement from 2011 until 2018. The collection of data from the year of establishment of the ESAs until the year of the last available minutes provides a solid time frame to draw conclusions.
This research finds that maintaining and improving reputation is the main reason for involving non-state stakeholders in decision-making processes, as the type of involvement turns out to be mostly symbolic instead of functional. Expertise and organisational capacity only explain non-state stakeholder involvement to a limited extent, as functional involvement has not been identified often. The results are quite similar between agencies, indicating a potential trend regarding the reasons why EU agencies with stakeholder bodies involve non-state stakeholders in their decision-making processes. Similar results between the ESAs are also identified over time. This research has observed an increase of symbolic involvement over time, hence reputation increasingly explaining stakeholder involvement in the latter
years. These results indicate that a lack of input legitimacy is likely to be an important root cause of the legitimacy challenge.
The next section of this thesis will elaborate on the three EU agencies, explain the concept of non-state stakeholder involvement and its implications concerning legitimacy. It will also engage with the literature on the main reasons for involvement. The third section will present the research design, justifying the link between the types and reasons for involvement and the case selection. Moreover, the chosen time frame and coding method will be explained in this section. The fourth section will present the findings of this research and will describe the theoretical and social implications of these findings. The final section will be the conclusion, summarising the research and proposing policy solutions and an agenda for future research.
CHAPTER 2 – THEORETICAL FRAMEWORK I- EU agencies and non-state stakeholder involvement 1- The emergence of EU agencies
The first EU agency, the European Foundation for the Improvement of Living and Working Conditions (EUROFOUND), has been established in 1975. Since then, EU agencies have grown in numbers, reaching 31 entities as of today. They have increased in numbers partly due to the EU’s aim to stop the legitimacy crisis that started in the 1980s (Schout, 2018). These agencies share several characteristics that will be presented here to give a better understanding of the research and its implications. First, they operate under European Public Law (Groenleer, 2009). Second, they are specialized in specific fields of expertise that are important to the EU. As such, they cover various policy fields such as financial markets, disease control, food security and medicinal products (Perez Duran, 2018). Third, EU agencies have monitoring and supervisory tasks. In addition, some of them also have regulatory tasks in their field of expertise (Busuioc et al., 2012). Fourth, EU agencies all involve non-state stakeholders in their decision-making process, through at least one of the following structures: 19 agencies involve them via stakeholder bodies, 12 via management boards, and 10 via public consultations (Arras & Braun, 2018).
Finally, they can all be classified as non-majoritarian bodies (Busuioc, 2013), which means they are formally independent public institutions. This term characterizes them as
"governmental entities that (a) possess and exercise some grant of specialized public authority, separate from that of other institutions, but (b) are neither directly elected by the people, nor directly managed by elected officials" (Thatcher & Sweet, 2002). In line with this definition, EU agencies are separated from the Community institutions (Council, Parliament, Commission), have their own legal status (Groenleer, 2009), are not elected by the people, and some EU agencies are relatively formally independent from EU institutions' elected officials (Wonka & Rittberger, 2010). However, in spite of their relative formal
independence, EU agencies are actually dependent on a broad range of actors (Groenleer, 2014). If they do not manage to build a good reputation in the eye of political officials, other EU agencies, interest groups, the media and industry, their lack of legitimacy will render them less autonomous from these actors (Groenleer, 2014: 261). It is on this broad audience that EU agencies' survival depends.
As earlier indicated, this thesis focuses its analysis on the three European Supervisory Authorities that were established in January 2011. These agencies share the characteristics mentioned above. To start with, their field of technical expertise is the financial sector. As such, they are part of the European System of Financial Supervision (ESFS), "a decentralized multi-layered system of micro- and macro-prudential authorities established by the European institutions in order to ensure consistent and coherent financial supervision in the EU" (ESMA, 2019a). Each ESA is commissioned with a specific field of expertise within this sector. First, the EBA focuses on banking. It aims to maintain financial stability within the EU and to protect the integrity and efficiency of the banking sector. Second, the ESMA focuses on the securities sector. It aims to improve investor protection and to promote stable and orderly financial markets. Finally, the EIOPA focuses on the insurance and occupational pensions sector. It aims at ensuring financial stability, the transparency of markets and financial products, and the protection of policyholders, pension scheme members and beneficiaries. The EIOPA is different from the other two agencies in that it has two separate stakeholder bodies, for the insurance as well as the occupational pensions sectors.
In addition, the ESAs all have supervisory and regulatory competences. They are in charge of contributing to the European Single Rulebook in their field of expertise, in order to create harmonized standards between member countries. They are also in charge of harmonizing supervisory practices, and of assessing risks in the sector. Busuioc (2013) highlights that they have far-reaching power in this respect. For example, their supervisory competences include the possibility to overrule supervisory authorities. Concerning their regulatory powers, they can draft quasi-rule-making binding technical standards (BTS). BTS are legally binding and directly applicable to member states once adopted by the Commission. They can also draft strong non-binding documents such as Guidelines, Recommendations and Opinions.
Despite the heavy emphasis on their independence in their founding regulations (Busuioc, 2013), the three ESAs seem to be strongly accountable to a broad audience. Their founding regulations still mention some mechanisms that make them quite dependent on their
supervisors. As such, the approval of their draft technical standards depends on the European Commission. The European Parliament and the European Council can revoke the existence of the ESAs. Furthermore, national authorities are part of the Board of Supervisors of ESAs (Regulation (EU) No 1093/2010, 2010; Regulation (EU) No 1094/2010, 2010; Regulation (EU) No 1095/2010, 2010). As regulatory EU agencies, the ESAs have to uphold or improve their reputation in the eye of a broad audience, such as media, interest groups, NGOs and the public at large (Busuioc & Rimkute, 2019: 1-2).
Finally, the three ESAs all involve non-state stakeholders in their decision-making processes via stakeholder bodies. These bodies form the focus of the current study. The
Banking Stakeholder Group (BSG) is the EBA's stakeholder body and the Securities and Markets Stakeholders Group (SMSG) is the ESMA's. As named earlier, the EIOPA has two
stakeholder bodies. One is focused on pensions and the other on insurance, respectively the
Occupational Pensions Stakeholder Group (OPSG) and the Insurance & Reinsurance
Stakeholder Group (IRSG). These bodies meet at least four times a year to give advice to the
ESAs. Occasionally, the ESAs also involve non-state stakeholders via public consultations. The next section will elaborate on the notion of non-state stakeholder involvement. It will explain its meaning and the way it relates to the legitimacy crisis.
2- Non-state stakeholder involvement and its legitimacy challenge
Non-state stakeholders are increasingly involved in EU decision-making processes (Grabosky, 2013). Each of the 31 EU agencies involve them, via stakeholder bodies,
management boards or public consultations, and in some cases via several of these methods combined (Arras & Braun, 2018). In the case of EU agencies, non-state stakeholders usually refer to three groups impacted by the agencies' decision-making process. They are industry representatives, citizen representatives - such as consumer groups and NGOs -, and
employees' representatives - such as trade unions (Perez Duran, 2018). In the case of the ESAs, each stakeholder body includes 30 non-state stakeholders, representing in a balanced manner the following parties, as stated in article 37 of the ESA's founding regulation (Regulation (EU) No 1093/2010, 2010; Regulation (EU) No 1094/2010, 2010; Regulation (EU) No 1095/2010, 2010). Some of the non-state stakeholders represent institutions, market participants and undertakings operating within the EU. This includes credit and investment institutions for the EBA, financial market participants for the ESMA, and insurance and reinsurance undertakings and insurance intermediaries for the EIOPA. Some other
stakeholder body members represent employees and SMEs representatives. In addition, some members represent consumers and users of services. This consists of banking services for the EBA, financial services for the ESMA, insurance and reinsurance services for the EIOPA. Finally, academics are also counted among the stakeholder members.
Stakeholder involvement refers to the possibility for non-state stakeholders to take part in
the agencies' decision-making processes, and the mechanisms through which this is taking place. Perez Duran (2018) identifies that interest groups are represented in different ways within EU agencies: the agencies' management boards, stakeholder groups and expert forums.
Management boards "formally represent the highest level of responsibility in EAs", with "making overall management and budget decisions". Management boards usually include representatives of the Commission and the member states, and sometimes of the EP, countries outside the EU, other European agencies and stakeholders (Perez Duran, 2018). Stakeholder groups or bodies are "permanent entities within the agency where a substantial number of the members are stakeholders, who meet on a regular basis" (Arras & Braun, 2018). They are commissioned to give advice on the agencies' output. Management boards and stakeholder bodies are closed instruments through which only selected non-state stakeholders can participate. On the other hand, expert forums, also named public consultations, are specific consultation in which anyone can participate. This generates "possibilities for a wide variety of stakeholders to get involved" (Arras & Braun, 2018). This is an open instrument for involvement. In this case, the three ESAs involve non-state stakeholders via stakeholder bodies and via public consultations (Regulation (EU) No 1093/2010, 2010; Regulation (EU) No 1094/2010, 2010; Regulation (EU) No 1095/2010, 2010). The choice to focus only on involvement via stakeholder bodies will be elaborated in the research design.
Various reasons have been identified in the literature for the establishment of stakeholder bodies. Busuioc (2013) makes a similar account to Majone (1999) on the positive effects on legitimacy that can be yielded by non-state stakeholder involvement. Busuioc (2013) identifies that EU agencies were established by their principals partly to "ensure the better involvement of affected social and economic actors [non-state stakeholders] in the segments of regulation of which agencies are in charge". She highlights that in turn such involvement "can be crucial to improving agency legitimacy, serving as a potential legitimizing source, even more so given agencies’ low democratic credentials". As such, following the 2008 financial crisis, the establishment of the ESAs and the involvement of non-state stakeholders were partly aimed at restoring legitimacy of the financial sector and of the institutions governing it throughout the EU (Groenleer et al., 2014). The involvement of the relevant stakeholders in the decision-making process was considered a good strategy to improve the legitimacy of decision-makers (Ibid). In spite of these measures, the legitimacy crisis is a continuing challenge and has even become a bigger one in recent years (Schout, 2018). In this case, Schmidt (2015) identifies that EU financial institutions lack legitimacy partly because of their technocratic nature, failing to consistently involve the civil society. Attention should be given to see whether input legitimacy has been successfully achieved in the case of the ESAs, in spite of EU institutions' commitment to do so (Busuioc, 2013: 34; Commission of the
European Communities, 2002). The study of the ESAs' actual reasons for and manners of involving non-state stakeholders might be helpful in that regard.
II- Reasons for non-state stakeholder involvement 1- A range of reasons
Few scholars have, as of today, researched the reasons why EU agencies involve non-state stakeholders in their decision-making processes. These studies will be presented and engaged with. In addition, this thesis will draw upon the literature that focuses not only on non-state stakeholder involvement regarding EU agencies, but in relation to agencies in general. The broader literature may suggest other reasons for involvement that have not yet been identified by the relatively nascent literature on EU agencies.
One of the main reasons for non-state stakeholder involvement is external pressure. This refers to formal requirement from principals, the pressure from regulatees to get involved and crises. First, the literature on EU agencies identifies that formal requirement from principals is an important reason for non-state stakeholder involvement. Arras and Braun (2018) indeed found that agencies’ principals, the Community institutions (the European Commission, Parliament and Council) required EU agencies to involve non-state stakeholders in order to help regulate these agencies. This requirement was set because, according to the principal-agent theory, the principal-agent has incentives to deviate from the principal's instructions. The agency is in a good position to do so, as it holds more information than the principal. In order to keep sufficient control over the EU agencies, the Community institutions required the involvement of non-state stakeholders within the founding regulation of the agencies. This is a fire-alarm mechanism of control, through which agencies are held accountable to regulatees. In turn, these regulatees are expected to inform the Community institutions of any misbehaviour by the EU agencies.
A second form of external pressure that might induce EU agencies to involve non-state stakeholders refers to regulatees' interest in participating in the decision-making process. They can exert pressure to get involved in the decision-making process. The literature has not yet studied in-depth this reason for involvement in the case of EU agencies. However, academics did suggest that there are strong incentives for regulatees to get involved in EU agencies' decision-making process (Arras & Braun, 2018). In particular, stakeholder mobilization should especially be high in the case of EU agencies with regulatory competences, such as the ESAs, as they are at a higher level of government activity (Ibid). In
addition, the broader literature on public agencies stresses that regulatees' pressure to get involved in the decision-making process is a main reason for involvement. Braun (2012) indeed explains that public agencies involve regulatees in their decision-making process partly because of these regulatees' "incentives to consider agencies an important venue to exercise influence".
Threats, such as crises, are a final form of external pressure that can induce EU agencies to involve non-state stakeholders. As such, academics studying the ESAs identified that the financial crisis has incentivized them to increasingly involve non-state stakeholders (Groenleer et al., 2014). The broader literature on public agencies also identified external consequential incentives as determining factors for involvement (Emerson, Nabatchi and Balogh, 2011). They refer to situational or institutional crises, threats and opportunities. However, external pressure in itself does not tell in which manner and for what purposes EU agencies actually involve non-state stakeholders once structures for involvement are in place. The literature focusing on agency needs as a reason for involvement suggests several rationales. Functional reasons for involvement are emphasized, referring to the need for expertise and organizational capacity. In addition, symbolic reasons for involvement are also pointed at, such as the need to improve reputation. The study of Arras and Braun (2018) is the first to systematically review the importance of these reasons for involvement in the case of EU agencies.
Arras and Braun (2018) find that EU agencies involve non-state stakeholders to respond to their needs, identified as organizational capacity, expertise and reputation. Organizational capacity refers to an agency's effectiveness in implementing rules, achieved via its capacity of maintaining compliance that is imposed to regulatees. Non-state stakeholders help agencies achieving high levels of organizational capacity, as they can "serve as intermediate or help facilitate the implementation process" (Ibid). In addition, non-state stakeholders can also be valuable for their expertise in the field. This includes technical, legal and administrative information as well as political intelligence. This expertise gathered by non-state stakeholders helps EU agencies in designing and implementing legislation. Finally, EU agencies place value in the role that non-state stakeholders can play in improving their reputation in the eye of the broader public. Arras and Braun (2018) argue that agency reputation can be improved by involving stakeholders strategically in the networks that they develop and maintain, and by being held horizontally accountable to the stakeholders. In addition, Arras and Braun (2018) also find that these three reasons for involvement are more or less central to EU agencies depending on their structure for involvement. They argue that EU agencies aim to gain
expertise and reputation by involving non-state stakeholders via public consultations or management boards. On the other side, the authors say that EU agencies mostly aim to gain organizational capacity and expertise by including stakeholder bodies. Stakeholder bodies are permanent structures within agencies that facilitate long-term interactions with stakeholders. This is of particular use for EU agencies aiming to improve their organizational capacity, as the latter benefits from long-term relations. In addition, expertise can easily be reached with this structure, but not only specifically via stakeholder bodies.
Although less extensively than Arras and Braun (2018), other studies on EU agencies also point to the importance of similar needs in determining non-state stakeholders involvement. Groenleer (2009) mentions that non-state stakeholder involvement helps EU agencies in encouraging "the European-wide exchange of information", a similar term to expertise. This involvement also helps EU agencies in "spurring further administrative integration in Europe", which relates to the strengthening of organizational capacity. Groenleer also indicates that this involvement should increase "the perceived quality and acceptability of their [EU agencies'] work", relating to the agency need of reputation. In line with this argument, Wood (2018) mentions reputation as the main reason for EU agencies to involve non-state stakeholders. Researchers also mention how important reputation is for public agencies. As such, Carpenter and Krause (2012) stress reputation brings agencies autonomy from their broad audience. Carpenter (2010) also stresses that reputation determines the power of agencies, such as their ability to enhance their legal authority, avoid budget cuts or lobby for policy change. Busuioc and Lodge (2016) further emphasize that reputation enables agencies to convince their audience that they are performing competently their accountability role. The next section will explain why this thesis engages only with the reason of agency needs, leaving aside external pressure.
2- A focus on agency needs
This thesis will solely focus on the perspective of EU agency needs as a reason for involvement. In particular, it will test Arras and Braun’s (2018) finding that EU agencies involve non-state stakeholders via stakeholder bodies for expertise and to improve their organizational capacity, not for reputational reasons. The focus on this particular reason for involvement does not undermine the relevance of the other reasons just mentioned. In fact, these reasons are likely to be complementary in explaining this involvement, as highlighted by various scholars. Arras and Braun (2018) identifiy that EU agencies involve non-state stakeholders mostly to fulfil their needs, but also because their principals require them to do
so. In a similar fashion, Braun (2012) mentions that agencies involve regulatees both because of agencies’ and regulatees’ interest in doing so. Finally, Emerson, Nabatchi and Balogh (2011) explain that agencies involve non-state stakeholders both because of external pressure and internal pressure, the latter referring to agency needs. The focus on EU agencies' needs, and on this study in particular, has been chosen for various reasons.
First, agency needs are a central reason for involvement, as suggested by several studies. A wide part of the literature that focuses on the broader population of public institutions identifies agency needs as the main reason for the involvement of regulatees in the decision-making process (Fung, 2006; Bell & Hindmoor, 2009; Peters & Pierre, 1998; Braun, 2012). It is therefore logical to assume that this reason is also of relevance in the specific case of EU agencies. Second, external pressures do not tell in which manner and for what reasons EU agencies use these structures for involvement. The focus on agency needs is therefore useful to understand the specific reasons for involvement, which appear to vary between EU agencies (Arras & Braun, 2018). Finally, in spite of the likely importance of this reason for involvement, this topic has been relatively under-researched. The before-mentioned work of Arras and Braun (2018) is thus far the most comprehensive study on the reasons for EU agencies to involve non-state stakeholders in their decision-making process.
3- Hypotheses
In continuity with Arras and Braun's research, this thesis will test the relative importance of organizational capacity, expertise and reputation in explaining why EU agencies involve non-state stakeholders via stakeholder bodies. From this research angle emerge two hypotheses that are non-exclusive, meaning that the relevance of one reason does not rule out the relevance of the other. Therefore, the hypotheses could also be complementary. These hypotheses are framed based on the findings of Arras and Braun (2018). Firstly, it is expected that EU agencies involve non-state stakeholders with stakeholder bodies to improve their organizational capacity and expertise. Unfortunately, the method of data collection does not make it possible to investigate whether each of these two reasons matter individually. In spite of this, the insights from this hypothesis should remain valuable. Secondly, it is expected that reputational reasons do not explain EU agencies’ involvement of non-state stakeholders with stakeholder bodies in their decision-making processes. In addition, it is argued that findings on the ESAs can to a certain extent be extrapolated to the broader population of EU agencies that have stakeholder bodies.
H1: EU agencies involve non-state stakeholders in their decision-making process through stakeholder bodies predominantly to improve their organizational capacity and expertise.
H2: EU agencies do not involve non-state stakeholders in their decision-making through stakeholder bodies to improve their reputation
CHAPTER 3 – METHOD
I- Operationalization: functional and symbolic involvement 1-The extent of involvement: from symbolic to functional
This thesis aims to assess the relative importance of the agency needs that trigger EU agencies to involve non-state stakeholders. To do so, it studies the minutes of the meetings the ESAs had with their stakeholder bodies (European Banking Authority, 2018; European Securities and Markets Authority, 2018; European Insurance and Occupational Pensions Authority, 2018a; European Insurance and Occupational Pensions Authority, 2018b). Every year, the ESAs are required to meet at least four times with their stakeholder bodies (Regulation (EU) No 1093/2010, 2010; Regulation (EU) No 1094/2010, 2010; Regulation (EU) No 1095/2010, 2010). Throughout each meeting, several topics of relevance to the agency's work are discussed, such as its current work programme, its achievements, its strategy for the future, trends in its field of expertise as well as administrative matters. The agency staff takes the lead of the discussion, and from time to time non-state stakeholders give advice on topics of discussion based on their expertise (European Banking Authority, 2018; European Securities and Markets Authority, 2018; European Insurance and Occupational Pensions Authority, 2018a; European Insurance and Occupational Pensions Authority, 2018b).
A challenge faced in this research was that the minutes rarely explicitly mentioned the reasons for EU agencies to involve non-state stakeholders. However, these minutes documented in details the interactions between the agency staff and non-state stakeholders. It makes it possible, for each interaction of the minutes, to identify the extent of non-state stakeholder involvement. The extent of their involvement has been classified in two types, namely symbolic and functional. By quantifying the number of times symbolic and functional involvement took place, this thesis aims to identify what is the main extent of involvement. This section will explain what is meant by the extent of involvement, whereas the next section will explain how this should be a useful tool to identify the reasons for involvement.
First of all, EU agencies can involve non-state stakeholders in a symbolic way during the meetings. This implies that non-state stakeholders do not have an impact on the agency's decision-making. There are different extents of symbolic involvement, as pictured by Figure 1. The figure points to a deeply symbolic involvement at the left of the spectrum. As we move further right of the spectrum, this symbolic involvement becomes less intense, up to the point where it turns into functional involvement. From that point onwards, non-state stakeholders
are considered to have an actual impact on the agency's decision-making process. These different extents of involvement have been identified from analysing the minutes.
Figure 1
Most left of the spectrum, non-state stakeholders simply listen as spectators to the agency staff. Non-state stakeholders are least impactful when they are ‘involved’ in this way. With such a type of interaction, the agency staff presents the agency's output without any contribution from non-state stakeholders. Stakeholders asking questions for clarification purposes only falls also within this category.
Moving to the right of the spectrum, non-state stakeholders are second-least impactful when their expertise is disregarded. In this case, non-state stakeholders interrupt the presentation of the agency staff to give advice or opinion on the agency’s output. As a reaction, the agency staff disregards the advice given. This can take two forms: the agency staff either explicitly tells the non-state stakeholders their advice is not relevant or staff tells
implicitly that this is not a relevant advice, by pointing out that the agency was already aware
of it and already acted upon it.
Moving further to the right of the spectrum, the final type of symbolic involvement identified is non-state stakeholders' expertise not commented on. This type can take two forms: first, non-state stakeholders interrupt the agency staff to give advice, which is consequently not commented on by staff. Second, different non-state stakeholders debate with each other and give an advice to staff, which consequently is not commented on by staff. This is not a rejection of non-state stakeholders' advice, but is not likely to impact the agency's decision-making process.
Moreover, EU agencies can involve non-state stakeholders in a functional way during the meetings. Contrary to symbolic involvement, this implies that non-state stakeholders do have
an impact on the agency's decision-making process. Different extents of functional involvement have been identified: first, non-state stakeholders start to be impactful on the agency's decision-making process when they get their expertise asked for further input. In this case, the agency staff asks non-state stakeholders to elaborate on the advice they provided after the meeting. Whether staff will actually take into account or implement this advice in terms of output is not known. Such a decision would then be taken outside the formal meetings, as none of the analysed minutes talks about points of discussion that were raised in previous meetings. However, we can assume that it will be taken into account and therefore impacts decision-making, as staff specifically asked for further input from the stakeholders. Stakeholder dissatisfaction would also grow if staff would disregard the advice at a later stage, to which the stakeholders spent time and effort developing in written format. The potential for influence of the stakeholders should be made clear from an early stage in order to avoid dissatisfaction (Bremere, Indriksone & Strigune, 2018).
Moving to the right of the spectrum, non-state stakeholders would have a bigger impact on the agency's decision-making if they get their expertise acknowledged and approved. In this case, non-state stakeholders give an advice to the agency staff, which will be approved. Generally, in such a case, the staff acknowledges and thanks the non-state stakeholders for their provided advice. At the very least, there is a formal recognition of the advice.
Finally, non-state stakeholders are most impactful on the agency's decision-making process when they get their expertise included. In this case, the agency’s staff not only approves the advice, but also informs the non-state stakeholders that it will take the advice into consideration for the agency’s final output. This is the most functional extent of involvement identified in the minutes. Overall, Table 1 presents examples of the six above-mentioned extents of involvement, taken from the minutes. In addition, Figure 1 complements this table by summarizing what are the various extents of symbolic and functional involvement. The figure also shows how the extents of involvement are linked to the reasons for involvement. This link will be elaborated on in the next section.
Table 1 - Examples of the different extents of involvement
Codes Examples
Listen as spectator - EU agency staff (Staff) overview without involvement from non-state stakeholders (Members)
Example: Whole interaction point throughout which the members do not intervene. (15/02/17 IRSG p.2)
- Members only asking questions for clarification, not any contribution:
"BSG members had a few questions on the interactions between the EBA’s 2019 work programme and the nearly finalized CRR2-CRD5 legislative package" (13/12/18 BSG p.2)
Expertise disregarded
- Staff explicitly says stakeholder contribution is not important:
"Verena Ross [Staff] noted that there is very little that ESMA can do on the topic […]. Significant work on the issue at SMSG level should be carefully considered vis-à-vis other priorities." (19/06/13 SMSG p.3)
- Staff implicitly says their contribution is not important, by giving justifications on the agency's action:
"BSG members urged the EBA to also take into account consumer protection issues, to the amount possible. [...] The EBA noted that this is regarding these BTS not within the remit of the EBA’s mandate." (14/05/13 BSG p.4)
Expertise not commented on
- Agency staff does not comment on members’ advice given to the agency
Example: Whole interaction point throughout which the last point is an advice given by the members, without any following comments from staff. (08/05/15 OPSG p.3) - Agency staff does not comment on the discussion between members, relevant to the work of the agency
Example: Whole interaction point throughout which the last point is an exchange of expertise between the members, without any following comments from staff. (28/02/17 OPSG p.3)
Expertise asked for further input
- Staff asks for member’s input after the meeting:
"The three Working Group Coordinators were asked to initiate the preparation of a possible BSG opinion". (10/12/13 BSG p.4)
Expertise
acknowledged and approved
- Staff approves input provided by members:
"Verena Ross [Staff] thanked for an interesting report and noted the strong links to cross-sector issues like PRIPs." (19/06/13 SMSG p.3)
Expertise included - Staff says it will include the relevant comment into the agency output:
"The following CfA were discussed and, when needed, drafting changes were made on the document" (19/12/11 OPSG p.4)
2- The link between the extent and reasons for involvement
Conceptualizing and operationalizing stakeholder involvement into symbolic and functional categories not only makes it possible to analyse the extent for involvement, it can also help us identify the ESAs’ reasons to involve stakeholders into their decision-making processes. To reiterate, this research takes account of the following three reasons that agencies can have to include non-state stakeholders into their decision-making process: reputation, organizational
capacity and expertise.
As functional involvement is described as a type of involvement where non-state stakeholders have an impact on EU agencies' decision-making process, EU agencies aiming to maintain or improve expertise are likely to involve non-state stakeholders in a functional way. In addition, functional involvement is likely to have influence on an agency’s effectiveness in implementing rules and legislation, therefore, its organizational capacity (Arras and Braun 2018). Stakeholder advice should be taken into consideration and have a proper impact on the agency’s decision-making process in order to enhance compliance with EU agencies' rules. Garcia Martinez, Verbruggen and Fearne (2013) support this argument: they mention the advantage of co-regulation to improve compliance within the context of the EU. Their results indicate that "businesses and organisations feel more committed to those rules that they consider to be of their own". This is because self-made rules should be easier to understand and seem more attainable. Hence, in line with the literature, this research argues that EU agencies involve non-state stakeholders functionally in order to improve their organizational capacity.
Unfortunately, this research method cannot differentiate between the relevance of the reasons of organizational capacity and expertise. They relate both to functional involvement and no other tools have been identified to make a distinction between them. However, it allows us to distinguish functional reasons from other forms.
On the other side of the spectrum, symbolic involvement relates to reputational reasons for involving stakeholders. As earlier described, symbolic involvement is the type of involvement in which non-state stakeholders do not have an impact on EU agencies' decision-making process. Their input is ignored or not even asked for. In line with, among others, Busuioc and Lodge (2016) and Papadopoulos and Warin (2007), it follows logically that EU agencies involve non-state stakeholders in a symbolic way when their aim is to improve their reputation. Busuioc and Lodge (2016) identify that to be seen as reputable is to give the impression of performing competently one's tasks. The word "impression" implies that EU agencies can improve their reputation without non-state stakeholders having a real impact on
their decision-making process. Papadopoulos and Warin (2007) point out that institutions that involve non-state stakeholders for reputational purposes do not include them systematically in their decision-making process. The authors explain that this is because the agencies find them often burdensome. Indeed, policy-making requires more time, effort and resources when the contribution of additional actors must be taken into account. Such agencies deem symbolic involvement sufficient to achieve their goal to enhance their reputation.
II- The coding method 1- Codebook
The previous section explained the meaning of functional and symbolic involvement and the link to the reasons for involvement. This has been summarized into Figure 1. This figure actually presents the codebook, which is a list of "words or short phrases that symbolically assign a summative, salient, essence-capturing, and/or evocative attribute for a portion of language-based or visual data" (Saldana, 2009). These words or short phrases enable the classification of data from a document. In this case, the codes classify the extent of involvement ranging from listen as spectator to expertise included. Each interaction reported in the meeting minutes of the ESAs will be classified as pertaining to a certain code.
This codebook has been constructed following the method of Protocol Coding, which is "the collection and, in particular, the coding of qualitative data according to a pre-established, recommended, standardized, prescribed system" (Ibid). With this method, the researcher must accept the assumptions of the protocol codebook and insure that the codes are clear and inclusive of all scenarios (Ibid). This codebook is based on the codebook of Fung (2006), as illustrated in Figure 2. The figure shows the different modes of communication and decisions that can arise in stakeholder meetings, based on a scale of intensity of involvement. In this way, Fung characterises the ways in which regulatees participate in public decision-making.
Figure 2
In order to better fit this research, the protocol codebook has been slightly adapted. Saldana (2009: 130-132) explains that one must "assess a protocol critically and, if necessary, adapt the guidelines to suit your own research contexts". The first modification made to the protocol codebook is to ignore Fung's differentiation between expressing and developing preferences and decisions. Fung separates those subcategories to explain that the development of preferences and decisions is a more advanced form of understanding. In this research, the differentiation is excluded, because it refers to individual understanding rather than decision-making. This modification has thereafter been made to the codebook.
The second modification to the Fung’s model is to end the depiction of expertise as a separate category. It has been observed that, throughout the ESA's meetings with their stakeholder bodies, expertise was the only way stakeholders contributed to the agency’s output. In this case, expertise was not the most ‘influential’ form of involvement, but the only type of true involvement from non-state stakeholders. It therefore made more sense to associate expertise to different extents of involvement. Overall, this research has identified that non-state stakeholders either listen as spectator or give their expertise. From this given expertise, either the outcome is limited to the expression of their preferences, or a decision is actually made based on their input. This update has thereafter been made to the codebook. The last modification made to this codebook is the redefinition of different levels of involvement. This modification has been made after analyzing the minutes and has given more nuance to the different extents of involvement from symbolic to functional. As such,
expertise given with the outcome of preference expressed was divided into expertise disregarded and expertise not commented on. Moreover, expertise given, with the outcome of decision made was divided into the more precise codes of expertise asked for further input, expertise acknowledged and approved and expertise included. This codebook is illustrated in
According Saldana (2009: 130-132), "some protocols, depending on their transferability and trust-worthiness, may also contribute to the reliability and validity of the researcher's new study". Although assessing the true trustworthiness of Fung’s framework is beyond the scope of this thesis, the transferability of this framework to this research can be assessed. First, Fung’s (2006) figure on the modes of communication and decision aims to characterize the intensity of involvement of regulatees in a meeting, from deeply symbolic to deeply functional. This is exactly what this thesis aims to characterize. Second, Fung's paper focuses on collaborative governance, a similar field of research to the one of this thesis. Fung refers to the cooperation between policy-makers and all the relevant stakeholders, including civil society, in decision-making (Ibid). Lastly, this research agrees with Fung’s notion that decision makers involve stakeholders because they "may lack the knowledge, competence, public purpose, resources, or respect necessary to command compliance and cooperation" (Ibid). This relates to the idea of this thesis that agency needs are a driving factor for non-state stakeholder involvement.
2- Segmentation
In addition to the codebook, segmentation is another important aspect of the coding method. A segment is found in a document (such as meeting minutes) and is a unit of analysis to which a code is attributed (Saldana, 2009). It can take the form of a word, sentence or paragraph. In order to have method consistency, only one form should be chosen to analyse documents (Ibid). In this case, the unit of analysis is a paragraph, grasping an interaction between members and staff. Each segment has a single outcome in terms of functional or symbolic involvement. In this thesis, the segment is referred to as the discussion point. The persons that were in charge of writing the minutes already broadly delineated these discussion points with bullet points or paragraphs. They report on the different topics discussed by the stakeholder group, chronologically organizing them from the first to the last topics that were discussed during the meeting. In addition, these discussion points have been further broken down into more specific discussion points when several outcomes resulted from an interaction.
The following discussion point exemplifies this segmentation process: the agency staff starts with a presentation of a certain topic concerning the work of the EU agency. Members comment on the relevance of the given presentation. Consequently, the agency staff acknowledges the relevance of some comments and ignores some others. In that case, two final outcomes can be identified: expertise acknowledged and approved and expertise not
commented on. This paragraph is therefore separated in two segments, which describe two
outcomes from the interaction.
This unit of analysis has been chosen because it avoids over and under accounting for certain codes. One single code should be identified from the outcome of an interaction between members and staff. On the one hand, under-accounting for certain codes would likely happen if one code was associated with a whole paragraph, regardless of the number of interactions and outcomes it comprehends. On the other hand, assigning a code to any sentence would likely over-account for certain codes. In particular, the code listen as
spectator would probably be over-accounted for. This is because most interactions start with a
presentation from the agency staff, during which non-state stakeholders listen as spectators. This code is often not an outcome from interaction but simply a step towards a situation in which interaction could be taking place.
Certain parts of the meeting minutes are not analyzed. There is a focus on interactions that result in agency output. Thus, interactions between members and staff on internal rules of procedure are left out. This choice has been made because what matters for this research is the actual involvement of non-state stakeholders, not the supposed involvement defined by internal rules. Internal rules of procedure include the following: discussions on renewal of the members’ participation in the meetings, changes to the rules of procedure, remunerations for staff, ideas to let the agency perform better administratively and, lastly, the sharing of tasks. The introductory update that the EU agencies' staff give at the beginning of the meetings is also excluded, as there is no ‘chance’ for interaction and it has no impact on the agency's output.
3- Strengths and weaknesses of the chosen method
The method of coding has however a few weaknesses. To start with, qualitative coding has the limit of being interpretative. The codebook and segmentation are built from text analysis and hence cannot be presented as purely objective. Second, coding is based on incomplete data. Though the minutes are quite detailed, some information shared during the meetings could be missing due the potential confidential nature of the information or because the note taker consciously or unconsciously left out certain interaction between staff and members. In addition, the interactions between ESAs and the stakeholders that take place outside the four, formal yearly meetings are not reported about. In several occasions during the formal meetings, members and staff plan to meet informally later in the year. These informal meetings are not transcribed into minutes and therefore could have a moderately limiting
effect on the trustworthiness of the data. A third limit of the coding method is that the information relies on the writing style of the person in charge of taking the minutes. His or her writing might have resulted in emphasizing different power dynamics between members and staff throughout the discussion. Certain note takers might have put more emphasis on detailing the interactions between members and staff, while some others might have left out some interactions that they deemed not essential to the final minutes. Such a ‘bias’ would have an impact on the results and data consistency.
However, the coding method has also some important advantages. First, this method delivers complementary insights to the interviews conducted by Arras and Braun (2018) on the same topic. Arras and Braun conducted 27 interviews with the stakeholder relations officer of EU agencies or a person regularly interacting with stakeholders, in order to "examine the motivations behind EU agencies’ interactions with stakeholders" (Arras & Braun, 2018). The coding of the meetings between the staff of the ESAs and their stakeholder bodies overcomes the ‘weaknesses’ of interviewing officials. First, coding presents the advantage of not being dependent on a person's subjective answer to a research question. This reduces the bias in the provided information whereas the interview method increases the likelihood of biasness, as EU agencies’ officials could have an interest in hiding (confidential) information. In addition, officials might not recall all the information discussed during the meetings. Most officials only have a maximum of two mandates of 2.5 years each, such as officials in management boards as well as in the stakeholder bodies (Regulation (EU) No 1093/2010, 2010; Regulation (EU) No 1094/2010, 2010; Regulation (EU) No 1095/2010, 2010).
Second, the coding of the minutes makes it also possible to identify the reasons for involvement over a period of time, as the reasons for involvement might have evolved over the course of the period 2011-2018. If this was to be the case, it could help identify and explain why stakeholders are involved in the decision-making process of the ESAs as well as the consequences this could have. Hypothetically, if this thesis were to find that stakeholders are significantly more involved in a symbolic than functional way in 2018 than they were in 2011, it would imply a negative development for the input legitimacy of the agencies of discussion. On the contrary, the interviews conducted by Arras and Braun do not account for these aspects.
Third, the publicly available coding documents make it possible to indicate what reasons should be central for which specific EU agencies in deciding to involve stakeholders in their decision-making process. On the one hand, if there are similar reasons for involvement found
between the different ESAs, this suggests that some of their similarities might be at the root of these same reasons. These similarities include their involvement via stakeholder bodies, but also include their expertise in finance, their far-reaching supervisory and regulatory competences, and the conflicting demands between formal independence from principals and actual dependence from a broad audience. On the other hand, slight different results between the ESAs can also help to understand the roots of the reasons for involvement. For instance, the slight difference between the ESAs in ‘field of expertise’ could have an impact on the reason for involving the stakeholders. Contrary to this research, Arras and Braun's interviews could not determine what reasons for involvement were mentioned by the specific EU agencies. For confidential matters, they were only able to state which reasons were mostly indicated by the broad population of EU agencies.
III- Case selection
1- The ESAs: an important case
The ESAs have been chosen out of the wider population of EU agencies for various reasons. First, the implications of this research regarding the legitimacy crisis are relevant for these agencies operating in the financial sector. Following the 2008 financial crisis, the ESAs were established and the requisite to involve non-state stakeholders was partly aimed at restoring the legitimacy of the financial sector and the European institutions that oversee the sector (Groenleer et al., 2014). To involve relevant stakeholders in the decision-making process was considered a good strategy to improve the legitimacy of decision-makers (Ibid). This refers to input legitimacy, achieved when the representation of who participates in decision-making is improved (Schmidt, 2010). In spite of these measures, the legitimacy crisis is a continuing challenge that has intensified in recent years (Schout, 2018). In this case, Schmidt (2015) identified that EU financial institutions' technocratic nature partly explains why they lack legitimacy, failing to efficiently involve the civil society. Therefore, it is important to identify how and why these institutions (alas, the ESAs) play an important role in the ongoing legitimacy crisis. The ESAs' actual reasons for non-state stakeholder involvement, and their resulting conduct with such stakeholders, might explain why a lack of legitimacy persists. Second, they have been chosen out of the wider population of EU agencies due to their high similarity. As earlier presented, they all operate in the field of finance, have far-reaching supervisory and regulatory competences, are strongly held accountable to a broad audience in spite of their founding regulations emphasizing their independence, and involve non-state
stakeholders via stakeholder bodies. Such similarity makes it possible to identify the small differences between these agencies that might explain diverging results. This is what will be studied with the cross-agency comparison.
In addition, due to their far-reaching competences (Busuioc, 2013) and their stakeholder bodies, the ESAs place greater emphasis on the input of non-state stakeholders than other kinds of EU agencies (Arras & Braun, 2018). There is therefore more at stake with non-state stakeholder involvement than with EU agencies that have less far-reaching competences. Important democratic implications can hence be drawn depending on the extent of involvement by the ESAs. Functional involvement could yield democratic benefits by reducing the gap between EU regulators and regulatees concerning their influence on the decision-making process. It is in line with Majone's argument (1999) that non-majoritarian have the potential to reduce the democratic deficit by having strong accountability mechanisms, including through public participation in decision-making. These democratic benefits can be achieved in case of fair representation and participation of all regulatees impacted by the EU agency. On the contrary, symbolic involvement should have negative democratic implications, as the gap between EU regulators and regulatees remains in spite of the measures put in place.
2- The ESAs: making the link with EU agencies
The ESAs also offer the opportunity to extrapolate, to some extent, the findings to EU agencies that share a similar structure for involving stakeholders in their decision-making process. Thus, agencies that involve non-state stakeholders in their decision-making process via stakeholder bodies. This similarity is of relevance to the research question, as past research highlights the close link between these structures and the reasons for involvement (Arras & Braun, 2018). Therefore, according to this study, other EU agencies involving non-state stakeholders via stakeholder bodies are likely to have similar reasons for involvement. EU agencies with stakeholder bodies are a useful group to focus on for two reasons. First, as explained earlier, EU agencies with stakeholder bodies and regulatory power appear to involve non-state stakeholders more actively and seriously in their decision-making process. This means that non-state stakeholder involvement is of significance to these agencies and that democratic implications arising from their involvement should hence be more substantial. Second, EU agencies involving non-state stakeholders via stakeholder bodies are the most numerous. In fact, Arras and Braun (2018: 1265) identified that there are 19 EU agencies that involve non-state stakeholders via stakeholder bodies, while only 10 involve them via public
consultations, and 12 via management boards. The focus on stakeholder bodies makes it therefore possible to extrapolate to more EU agencies than a focus on public consultations would offer.
The potential for extrapolation is enhanced by the in-depth analysis of this thesis. First, the in-depth analysis of three within-case studies should strengthen the reliability of the findings. According to the law of large numbers (Toshkov, 2016), the larger the data, the higher the likelihood of the results. Three withcase studies is a relatively large x in the field of in-depth qualitative research as well as in proportion to the total population of EU agencies. Overall, 142 meeting minutes have been analysed, which represent 1822 codes. This is a relatively significant amount of data over a multi-year time period. Second, data consistency between the analyzed minutes of the three ESAs also strengthens this claim. Nearly a similar amount of stakeholder body meetings have been organized by the ESAs, with about four meetings held during each year.
As for the research period, this thesis has analyzed the meeting minutes of the ESAs for the years 2011 until 2018. The collection of data from the year of establishment of the ESAs until the year of the last available minutes provides a solid time frame to draw conclusions. All minutes available have been included in the analysis, because "the danger is that the portions deleted might contain the as yet unknown units of data that could pull everything together, or include the negative case that motivates a rethinking of a code, category, theme, concept, theory, or assertion" (Hruschka et al., 2004: 15).
CHAPTER 4 – ANALYSIS I- Empirical findings
1- Reputation as primary reason for involvement
This section presents the findings of this thesis. The coding of the minutes had made it possible to quantify for which reasons stakeholders have been involved in the decision-making process of the European Supervisory Agencies from 2011-2018. As explained in the research design, the different extents of involvement relate to different reasons for involving stakeholders. The quantification of the different extents of involvement therefore enables the identification of the main reasons for involvement. Figure 5 presents the variation in the extent of involvement in each ESA's stakeholder body, regardless of the years. The horizontal axis presents the different extents of involvement, from most symbolic to most functional. The codes are represented by numbers from 1 to 6, which are the numbers used to codify the minutes. These numbers stand for the following: listen as spectator (1); expertise disregarded (2); expertise not commented on (3); expertise asked for further input (4); expertise
acknowledged and approved (5) and expertise included (6). As explained earlier, the first
three codes relate to symbolic involvement, while the last three relate to functional involvement. The vertical axis presents the number of times each of these six extents of involvement were identified in the minutes. This has been translated in percentages. The original numbers are presented in Table 2.
0 5 10 15 20 25 30 35 40 45 50 0 1 2 3 4 5 6 N u m b er o f t im es an e xt en t o f in vo lv em en t a p p ea re d in th e m in u te s (in % )
Extent of involvement
Figure 5. Cross-agency variation in the extent of involvement
EBA (BSG) ESMA (SMSG) EIOPA (IRSG) EIOPA (OPSG)
