contingencies that possibly influence
the effectiveness of European
Regulatory networks
By Anouk Ouwehand
S1970208
Jelmer Schalk
Machiel van der Heijden
Foreword
In front of you lies the thesis ‘A comparative case study on the contingencies that possibly influence the effectiveness of European Regulatory Networks’. This research was done through a case study of the European Network for Implementation and Enforcement of
Environmental Legislation (IMPEL). This thesis was written in relation to the graduation from my Master of Public Administration at the University of Leiden. This thesis was written in the period from December 2017 until August 2018.
First, I want to thank my supervisors for their support during the writing of my thesis. Writing my thesis was not always easy because of several reasons, but my supervisors have been patient and supportive, allowing me to complete it. Secondly, I want to thank those who have allowed me to interview them and use this information. Without the interviews getting the results I needed to answer my research question would have been very difficult.
Third, I want to thank my fellow students of the capstone who have helped me when I had questions regarding the thesis and were also there for moral support. Lastly, I want to thank my family and friends for moral support especially during the warmer days of summer.
Studying at the University of Leiden has been a pleasant experience. I have had the possibility to learn from interesting teachers of which I really enjoyed their lectures. I, therefore, hope that this thesis will be an enjoyable read.
Yours truly, Anouk Ouwehand
Abstract
European regulatory networks have become more present in the regulatory regime of the EU. Besides being more in numbers their influence on the implementation and enforcement of EU legislation has also increased. Because they are becoming a bigger part of the EU society we need to find out whether they are effective in fulfilling their purpose. European regulatory networks like other transnational regulatory networks have been said to be so disrupted by political interests and political battle that they are not effective at all. While other scholars argue that these transnational regulatory networks are much more flexible and without the restraints of a constituency than an international organization that they are very promising structures to combat big problems such as global warming. However, in the literature of transnational and European regulatory networks, there is not much about the actual effectiveness of these networks. In this thesis, by testing the theory from an influential article about network effectiveness of public service delivery networks by Provan and Kenis (2008) through a comparative case study of the European Union Network for the Implementation and Enforcement of Environmental law the effort is made to see if this theory can help us better understand network effectiveness of European regulatory networks. While answering the research question: “ To what extent do network governance contingency factors explain the effectiveness of European regulatory networks?” it was found that rather than contingency factors the management of three tensions might be more important for network effectiveness.
Table of contents
Foreword ... 2 Abstract ... 3 Introduction ... 5 Regulatory networks ... 9 Advantages ... 13 Limitations ... 15 Theoretical framework ... 17 Network effectiveness ... 18 The factors ... 22 Research methods... 31 Case selection ... 32 Operationalization ... 33 Data collection ... 36 Results ... 40 Dimensions ... 40 The themes ... 45Analysis of the results ... 59
Conclusion ... 62
Appendix ... 64
Interview data ... 64
Introduction
There are certain problems that have transcended the boundaries of the nation-state. To tackle these transboundary problems, like global warming, networks have been created in the last decennia. Scholars say that networks are needed to reach ambitious and complex goals (O'Toole, 1997). Preventing the disastrous effect of transboundary or wicked problems like climate change is an ambitious and complex goal that can only be solved when all countries try to fight it together. It is a dilemma of immense scale which needs policy coordination and multi-level governance (Esty & Moffa, 2012, p. 777). Governments are working together more in the last decades than ever before but instead of international organizations, the network style cooperation has become more popular, especially in regulatory cooperation (Raustiala, 2002). These regulatory networks have members that are domestic regulatory agencies or executive institutions and cooperate in a less formalized structure than international organizations (Raustiala, 2002). Although less formal than international organizations transnational regulatory networks do differ from each other (Eilstrup-Sangiovanni, 2009). Thus, it seems that these networks have an increasing influence on the way of how we solve wicked problems.
To build more capacity to deal with wicked problems like environmental problems the EU responded with a possible solution that would avoid the increase of centralization. Facilitating the creation of networks between regulatory agencies from each of the Member States (MS) in so-called ‘regulatory networks’ (Bach, De Francesco, Maggetti, & Ruffing, 2016). These European regulatory networks (ERN) are networks that deal with EU regulations and their implementation in the MS, each ERN deals with an overarching topic. The national regulatory authority (NRA) of each MS is part of the respective ERN that concerns their topic of focus. (Van Boetzelaer & Princen, 2012). The NRAs of different MS can then work together by exchanging best practices and implementation guides. The EU hopes to create a level playing field for regulatory harmonization as a result of this type of horizontal co-ordination. (Van Boetzelaer & Princen, 2012). The idea is that these ERN’s will decrease the asymmetry between rule formulation and rule enforcement. This is needed because the consistent implementation is essential for EU policies and regulations to work (Van Boetzelaer & Princen, 2012, p. 819).
However, the idea of using networks as an instrument for legislation implementation has been a topic of discussion. There are scholars that see an advantage in using a network structure and conclude that these networks are effective in reaching their goals, while there are also scholars who have found that there are some important disadvantages and using a network structure is therefore not effective at all. It seems that it is not clear whether ERNs are an effective instrument for the EU to improve the implementation of legislation. Therefore, there is a need to
find a way to measure the effectiveness of networks to find out which weighs more, the advantages or the disadvantages in terms of being able to reach the goal that a network has. Thus, to find out more about the effectiveness of networks in this thesis theory about network effectiveness will be tested in the form of a case study of an ERN.
The reason for choosing a network structure to build more capacity is because of the benefits networks have and there are many scholars who discuss the benefits of network creation. One benefit of Transgovernmental networks like ERNs is that they are more flexible and adaptable compared to international organizations (Raustiala, 2002). Furthermore, networks have a more horizontal structure and are often more informal even when the network is formalized. They are faster because communication of information is quicker between the members of the network compared to traditional hierarchical governmental structures because there is no direct involvement of political principals (Eilstrup-Sangiovanni, 2009). This is also because network members tend to train and hire staff in similar ways, through socialization networks are able to simplify communication which makes it easier to share information. Networks also create an environment that fosters experimentation and innovation (Raustiala, 2002). Moreover, networks are a good way to include non-governmental organizations and governments for solving a problem (Whelan, 2015). Thanks to these characteristics implementation might be faster and more flexible when a new regulation is issued or when there are problems (Newman & Zaring, 2013). In the case of ERNs, these characteristics account for the decision of their creation. The characteristics of networks could rid of the time consuming the legislative process of the EU (Eberlein & Newman, 2008, p. 36).
However, skeptics argue that regulatory networks are just extension of the state which limits them compared to normal networks in their “ability to independently influence international dynamics or provide effective governance solutions” (Newman & Zaring, 2013, p. 246). These scholars are thus skeptical of regulatory networks as they see structural weaknesses in the set-up of the networks (Van Boetzelaer & Princen, 2012). Furthermore, there is another point that we have to take into account. The EU is an organization that facilitates the creation of policy while dealing with the tensions of national interest of the 28 MS and the objectives of the EU, as the EU has intergovernmental and supranational spheres. ERNs thus need to deal with national and international constituencies and their interests (such as EU regulation and international treaties like the Paris climate agreement) (Verdier, 2009). This means that instead of reaching the best solution together each regulator is limited by their constituencies to act according to their preferences which often results in a decision which is sub-optimal for example
for the environment (Verdier, 2009). Therefore, to summarize the argument skeptics argue that regulatory networks do not produce any real output and only serve as platforms for discussion.
Regulators in a TRN are experts on a certain topic and speak about complex problems in an effort to settle these. These regulators are supposed to be free from domestic politics to reach a decision that is in the interest of “all people”, however, this is rarely the case in reality (Verdier, 2009). Regulators are accountable to domestic structures, these structures and their incentives are more influential to the regulator than the incentives they get from reaching the international objective (Verdier, 2009). This is because short-term national interests have more priority than international long-term collective interests, These domestic constraints then shape the strategy of the regulator which they bring to the international decision-making sphere. Therefore, it can happen that domestic interest and international interests are in conflict with each other which could hinder the decision-making process of the network (Verdier, 2009). To prevent (too much) hinder TRNs should address their distributive and enforcement problems to reach international objectives. However, because of the lack of legal jurisdiction and any kind of enforcement mechanisms, it is hard for regulators to make it work. (Verdier, 2009)
If these issues are not taken into account and distributive problems are not tackled, powerful states can use threats and other incentives to steer outcomes to their preferred outcome. Moreover, more powerful states deal with higher costs when they have to change their existing regulation (Raustiala, 2002) and would thus push for outcomes that are similar to or the same as their current regulation. Moreover, the economic market also likes this as they are often also reliant on more powerful states and are also accustomed to current regulation of the more powerful state. Furthermore, the lack of tools for enforcement results in the lack of commitment of participants to implement new regulation or rules set by the TRN (Verdier, 2009, p. 163) creating moral hazard and increasing the possibility for members leaving the group or non-compliance.
As Verdier (2009) suggest doing research on the effectiveness of TRNs is important because we do not know under what conditions these networks can be effective. Therefore, there should be more research on what mechanisms and under what circumstances TRNs could be effective in producing results and cooperation. This is proposed because little research has been done on the effectiveness of regulatory networks. Scholars that are positive about the future of TRNs have not been focusing on systematic analyses of TRNs (Verdier, 2009). ERNs being European centered they also possible deal with problems of network effectiveness. The problem with the lack of network effectiveness literature is probably because effectiveness in a
transnational and public setting is very hard to measure as outcomes are often vague and complex. Available literature on possible factors of the effectiveness of regulatory networks is focused on different possible variables to explain the effectiveness of regulatory networks and do not come to a definite conclusion. Therefore, we do not yet know in which circumstances regulatory networks can be effective. This is a problem because these networks have been proved to be influential for regulation implementation in Europe (Bach et al., 2016) and if these networks appear not to be effective it could have the adverse effect of what the EU wanted to attain.
However, there is literature where networks have been extensively researched such as in public service delivery networks. For example, the many works were written by Provan, Milward, and Kenis and all the works that have been built upon theirs. They have found factors and circumstances in which these networks could be effective. Even when taking into account that public service delivery networks actors are different in their interdependence than ERNs, this literature could provide us with a way to look at regulatory networks and form assumptions into which circumstances and conditions they could be effective in a transnational context. In this thesis, an attempt will be made to analyze an ERN on their effectiveness. This will be done by exploring the contingencies of a network following the theory by Provan and Kenis (2008) where they expand on the contingencies of network governance types and when they should be effective. The research question is thus, To what extent do network governance contingency factors explain the effectiveness of European regulatory networks?
To see what factors are influencing the effectiveness of a regulatory networks a case study will be done on the European Union network for the implementation and enforcement of the environmental law (IMPEL). Environmental legislation can be a sensitive topic as it often affects the industry and the economy of a country, therefore one could expect tensions between national and international spheres. These tensions could possibly show through political conflict in the cooperation between members of the network during the execution of projects or in which projects to give priority to. This network was chosen as the case study because the network allows for comparison within the case. IMPEL has five different themes concerning environmental legislation that they focus on creating variation within the case. Because of the variation within the case testing the theory will possibly deliver different outcomes depending on the theme. Using an information-oriented selection strategy it is possible to gain maximum variation in one case which allows for the exploration of different factors of the theory (Flyvbjerg, 2006). Using a comparative method the different themes will be compared with qualitative data, such as interviews done with Dutch members of the network and project documents. Using different sources of data allows for triangulation and decreases selection bias.
IMPEL is an informal non-profit network that is a collaboration between environmental authorities of the 36 countries including all EU MS and members from other countries which are NRAs from “Republic of Macedonia, Serbia, Turkey, Iceland, Kosovo, Albania, Switzerland and Norway” (IMPEL, n.d.-a). The networks administrative department has its office in Brussel Belgium and is an official non-profit organization under Belgium law. Although the network was created in 1992 as an informal network under the EU it became autonomous in 2008. (IMPEL, n.d.-a) Moreover, IMPEL is an informal network but has a structure which is very formalized compared to the other regulatory networks in the environmental field (Martens, 2008, p. 33). IMPEL has been chosen as a case because it is a network that complies with the functions that ERNs should have. Although they have separated from the EU they still have a close connection with the European Commission (EC). Moreover, IMPEL is an informal network that is very institutionalized because it works so closely with the EU. IMPEL is also very active in a policy regime of the EU to which the MS have delegated quite a lot of responsibilities to (Van Boetzelaer & Princen, 2012), meaning that IMPEL as a network that is focused on implementation and enforcement of environmental regulation can have a lot of influence on the members of the network. Because of its influence, it is interesting to find out whether or not IMPEL is effective in reaching her goals.
The first thing that needs to be done is to look at what literature, theoretical arguments, and empirical evidence are already available to explain regulatory networks and network effectiveness. As said previously most research on the effectiveness of public networks are focused on public service delivery networks, however, there might still be information that could be useful for regulatory networks. Next, literature that does focus on regulatory networks will be explored. After the literature section, some assumptions will be made and hypotheses will be given, thereafter the research methods and data collection of the case will be discussed. Then the results and data analysis will follow. The thesis will be rounded up with a discussion.
Regulatory networks
As mentioned in the introduction the EU has a lack in capacity to deal with the implementation of the legislation of the many policy fields that the EU deals with. Therefore, to create uniformity or to create a level playing field for the MS the EU promoted the idea for national regulators to work together and create networks. Although each network has their own story of how they came into existence, the main argument is that they exist to support the harmonization of EU regulation. These networks are often jointly researched together with EU agencies as they
can be seen as complementary structures of multilevel coordination (Bach et al., 2016). Therefore, ERNs are often studied as agencies, yet in this thesis, they will be looked at as pure networks.
Regulatory networks are similar but also tend to be very different from each other. First of all, regulatory networks exist of regulatory institutions that are embedded into an international organization or individual regulators that decided to make a connection and work together in the form of a network (created by itself) (Eilstrup-Sangiovanni, 2009). Moreover, many of these regulatory institutions and regulators have created networks that were able to grow beyond national borders (Slaughter, 2004) or have started out as a network between governments. However, as they have not been formed through the existence of a treaty they are not considered to be part of the legal system (Slaughter, 2004). Yet, they still have an executive purpose and could be called executive transgovernmental networks (Slaughter, 2004). Verdier (2009) gives a definition of TRNs which summarizes what these networks are: “informal multilateral forums that bring together representatives from national regulatory agencies or departments to facilitate multilateral cooperation on issues of mutual interest within the authority of the participants” (p.118).
To categorize where these networks fit three types can be distinguished: Networks of executive officials that were created within international organizations, networks that were born through an agreement between heads of state, networks that were created on the initiative of NRAs themselves outside of a formal framework (Slaughter, 2004, p. 45). Networks of executive officials are government networks within international organizations. These are created when high state officials come together to create a new institution, after signing treaties, or after discussing general new legislation is their field of work (Slaughter, 2004). Networks that were born within the framework of an executive agreement are networks that have been created to operate outside of a formal institution but heads of state agree that such a network is necessary (Slaughter, 2004). Often these networks were already growing and existing by themselves but get an increased mandate because heads of state acknowledge them as a useful organization (Slaughter, 2004). The last type, spontaneous government networks, are networks that were created without any interference of formal institutions or any agreements. Here national regulators were the ones to give the initiative to cooperate with each other on a certain issue (Slaughter, 2004).
Then these networks can be categorized for their purpose, what do they do? First, there are information networks, where members exchange information such as experiences, best practices, and discuss ideas. Exchanging best practices and experiences makes it possible for
governments to improve their policies or create new more effective policies (Slaughter, 2004). Secondly, enforcement networks focus on “enhancing cooperation among national regulators to enforce existing national laws and rules” (Slaughter, 2004, p. 55). Lastly, harmonization networks, although also focused on enforcing national laws and rules the idea behind it is broader. Harmonization networks main focus is to uniformly implement regulatory standards as the network allows for better-coordinated cooperation between regulators. These networks often come into existence because regulatory standards or laws have changed within a certain framework or agreement (Slaughter, 2004).
In Europe specifically, ERNs have some more defined functions which are: “to advise the Commission, draft implementing legislation, coordinate national enforcement, promote information exchange among national regulators, and make recommendations to the public on emerging regulatory issues. These regulatory groups organize and interact with sectoral stakeholders as supranational legislation is developed and at the same time devise on-the-ground strategies for implementation.” (Eberlein & Newman, 2008, p. 32) ERNs thus exists between the EU and national regulators (NRAs), being an intermediate between these two levels ERNs have to rely on the authority granted by the EU to NRA’s to implement and enforce new directives and agreements that were made at the EU level. When ERNs were set up they were “given tasks of coordinating national regulatory authorities through functions such as providing ‘technical’ advice to the Commission and establishing norms and benchmarks” (Coen & Thatcher, 2008, p. 56). These tasks overlap with the idea of information network category of Slaughter (2004). She also mentions that these categories can overlap which can be seen in ERNs as they need to provide information to the EU and each other, implement and enforce new rules, and make sure that EU regulation is implemented in a similar fashion in all member states. Still, most of the above-described tasks are more information focused than enforcement or harmonization. Looking at the research of Levi-Faur (2011) he gives an overview of the European regulatory space and provides a list of ERNs. This list shows that they are mostly councils, meetings, advisory, or committees showing the preference towards information type networks.
Coen and Thatcher (2008) also categorize ERNs on a scale from fewer powers received from the EU and NRAs to more powers. Or this categorization can be specified as a continuum from ‘hard’ to ‘soft’ “where hard refers to greater powers and formalization of position” (Coen & Thatcher, 2008, p. 57). Another typology is by Van Boetzelaer and Princen (2012) who make a distinction between central type co-ordination and non-central type coordination. The latter is co-ordination governed by a central decision-maker and the other is when participants manage co-ordination among themselves (Van Boetzelaer & Princen, 2012, p. 822). In this spectrum ERNs
use non-central co-ordination as in ERNs there is no formal hierarchy (Van Boetzelaer & Princen, 2012). Within non-central co-ordination there are two types of processes: “co-ordination through mutual adjustment, and coordination through common purpose” where the latter means that participants adjust their behaviour to obtain the desired behaviour instead of sharing common objectives and the co-ordination through common purpose means having agreed upon certain rules of behaviour to reach co-ordination (Van Boetzelaer & Princen, 2012). With these previously mentioned dimensions (as summarized in table 1) of regulatory networks, the position and the IMPEL’s case can be further defined later in the thesis.
Then there are characteristics that regulatory networks share. First of all, they all have members that are agencies or ministries, take “action through rulemaking, a taste for subsidiarity, enforcement through peer review, small secretariats, and informal process” (Newman & Zaring, 2013, p. 245). Regulatory networks also often use consensus as the main modus of decision-making without formal voting procedures, these networks select who can participate, and are only transparent after they release the results of meetings (Verdier, 2009). Moreover, they do not produce legally binding documents as a network and they often do not have a system in place to formally monitor the decisions that were made as a network (Verdier, 2009) and therefore have a lack of enforcement power. TRNs act within a structure of dual-delegation which means that national governments have handed authority to supranational institutions and at the same time to NRAs (Eberlein & Newman, 2008). These TRNs inside of this dual-delegation structure have functions such as participating in “supranational rule-making, enforcement, and implementation” (Eberlein & Newman, 2008, p. 32). While relying on the authority they have received to act out above functions they also “formulate rules, which become binding owing to the two-level mobilization of delegated authority” (Eberlein & Newman, 2008, p. 32) which forms the concept of being ‘double-hatted’, wearing two hats at the same time. ERNs, for example, is in the hands of agencies and the EC who tend to institutionalize these networks as their own agents (Levi-Faur, 2011). ERNs, therefore, have little to say about their agenda and there are only a few ERNs that actually have control of their agenda and financial resources (Levi-Faur, 2011) which shows the extent these networks are ‘double-hatted’.
Already mentioning a possible limitation to regulatory networks the next part will first discuss the advantages that regulatory networks have compared to other methods of aiding regulation implementation. Afterward, the limitations will be discussed which will lead us to the theoretical framework of the thesis.
Table 1 Network typologies Theory/dimensions Type of executive networks (Slaughter, 2004) Networks within international organizations
Networks that were born through an agreement between heads of state Networks on the initiative of NRAs themselves outside of a formal framework Type network purpose
(Slaughter, 2004)
Information networks Enforcement networks Harmonization networks Level of power (Coen &
Thatcher, 2008)
Hard soft
Type of non-central coordination (Van Boetzelaer & Princen, 2012) Coordination through mutual adjustment Coordination through common purpose
Advantages
There are reasons for creating TRNs, as said in the introduction there are societal problems that have grown beyond the national border and TRNs would be a possible solution to these problems. Therefore, countries have increasingly worked together on this issues. However, in the past decades instead of international organizations networks have been the trend of cooperation type to solve these cross-boundary problems and is especially seen for regulatory cooperation (Raustiala, 2002). As Anne-Marie Slaughter in The new world order puts it, TRNs are the solution to what she calls the ‘globalization paradox’. Because of globalization, there is demand for more government on multiple levels (from regional to global), yet, “we don’t want the centralization of decision-making power and coercive authority so far from the people actually to be governed”(Slaughter, 2004, p. 8). Furthermore, there is an accountability problem when there is no centralized power but public actors are involved in solving global issues. These government actors could work together with a wide range of non-governmental (NGO) actors, however, governments have a public role as re-presenters of their domestic constituency while NGOs do not which can be dangerous as their focus is often profit (Slaughter, 2004). Therefore, if government networks become entities that are able to exist alongside or even replace traditional international institutions creating a new world order these networks hold great potential (Slaughter, 2004, p. 10). Regulatory networks are one of the types of networks that are used to tackle the ‘globalization paradox’.
TRNs are organizations that have many experts employed and this expertise is important because sharing collective experiences between regulators creates a good basis to assess regulatory needs, monitor compliance, and developing legislative proposals in general. (Eberlein & Newman, 2008) Networks also have the advantage of the ability to create and mobilize relationships with the private sectors and nationally elected officials, creating the possibility to pool resources and capacity for collective action. The information that is gathered through sharing expertise and through the ties with domestic actors. The EU can use this information to improve their regulations and policies. (Eberlein & Newman, 2008) Thus, through sharing information and best practices the effectiveness of regulation and policies will increase as countries can solve problems they are having with the implementation of regulation and policies (Newman & Zaring, 2013). Moreover, these regulatory networks have regulatory authority, meaning that they are able to reprimand noncompliance of the regulations they are trying to implement. Because ERNs work closely with the EC they can report noncompliance of regulation and EU policy, however, ERNs also exist to work on the reason for noncompliance (the regulation or policy might not be feasible in a certain MS). With this power and their expertise participants in these networks can frame the agendas of their country to reach their goals (Eberlein & Newman, 2008).
ERNs are TRNs but focused on the EU-level and “function to advise the Commission, draft implementing legislation, coordinate national enforcement, promote information exchange among national regulators, and make recommendations to the public on emerging regulatory issues. These regulatory groups organize and interact with sectoral stakeholders as supranational legislation is developed and at the same time devise on-the-ground strategies for implementation” (Eberlein & Newman, 2008, p. 32). The idea of these networks is that they should be able to be more effective as the concept of TRNs is that they enable more speedy and flexible responses (Eberlein & Newman, 2008). This would mean that governance responses, the legislative process of rule development and its implementation could be sped up as it removes parts of the hierarchy and bureaucracy of EU institutions (Eberlein & Newman, 2008).
Because TRNs do not produce legally binding documents they do not need ratification processes like treaties which are often very time-consuming and bureaucratic. TRNs could therefore possible speed up the legislative process by removing alternating between one institutional level to the other like how it happens with policy-making in the EU and shortening the implementation process, (Eberlein & Newman, 2008; Newman & Zaring, 2013) because the network works as an intermediary between the EU and NRA. Within the EU these networks are also concerned with the implementation of directives in member states. The network structure then allows through information and best practices sharing to reach more effective
implementation of EU regulation by the NRA. Because when experiences are shared between countries they can learn from each other and will use the best possible way to implement and later enforce legislation, which leads to a more harmonized legislation of the EU.
Another way of looking at why TRNs are more speedy and flexible is that “individual government agencies and actors negotiate directly with their foreign counterparts and reach informal understandings relating to their areas of responsibility. Their expertise and insulation from domestic political pressures allow them to solve problems that traditional international organizations cannot adequately address.”(Verdier, 2009, p. 115). Thus, through this more informal way of cooperating TRNs can be more flexible than international organizations in reacting to emerging global challenges (Newman & Zaring, 2013). Yet, national regulators are not free from the preferences of their domestic constituencies. These domestic preferences include preferences on political issues, legislature, domestic industry or the domestic market, media, and the public (Verdier, 2009). Regulators are therefore bound to the country by incentives and accountability structures that have much more influence on the functioning of the regulatory agency than any supranational organization (Verdier, 2009).
Limitations
Here we reach a first and important limitation or drawback to these TRNs. You can recognize the aforementioned two-level game discussed by Putnam (2009), although a little different. Putnam (2009) describes the national playing table as the place where different national actors pressure the government to pursue favorable outcomes. Politicians then create coalitions among these actors to gain power. On the international table, national governments try to find a balance between satisfying national pressure and negative consequences for international relations (Putnam, 2009). The point is that the best move that a politician may make on one table might not be a wise decision for the other table, but a politician can’t choose to have many different opinions on either table or their legitimacy will be endangered (Putnam, 2009). TRNs thus play a similar game, on the one hand, they are restricted by domestic preferences ( such as the interests of their domestic markets) but on the other hand, they need to reach network-level goals. Furthermore, in the EU NRAs have been working closely with the EC outside of other institutions or structures. Because of this NRAs start acting in a ‘double-hatted’ manner, serving both the domestic constituency and the EC (Egeberg & Trondal, 2009). As a consequence, NRAs can receive conflicting policy expectations from both the institutions served which could be difficult to unify (Egeberg & Trondal, 2009) which decreases the possibility for successful collective action.
In the case of ERNs specifically, they have to deal as a network with regulations and policies that need to be implemented in all EU MSs. Compared to other TRNs these European versions are more pressured to actually implement directives as there are sanctions in place for MSs that do not comply with directives in a given time period. The number of MSs has increased a lot since the creation of the European Union. Moreover, interdependence between the EU MS have also increased yet they are jealously guarding their sovereignty (Eberlein & Newman, 2008). The MSs are thus expecting a lot from the EU, however as they both want to keep their sovereignty but expect the EU to act out their policies the EU lacks the capacity to meet the demands of the MS. This is called the “governance dilemma” (Eberlein & Newman, 2008), this dilemma makes it a difficult task for EU-law and policies to be effective. As a response to the governance dilemma and the implementation gap transnational networks have been given a mandate in the EU decision-making process (Eberlein & Newman, 2008). Moreover, European Regulatory Networks (ERN) have been created by the European Commission to hopefully increase the effectiveness of EU policies and regulations (Van Boetzelaer & Princen, 2012). Yet, these networks might have trouble coordinating new policies that come from supranational organizations such as the EU.
This brings the first tension between diverse and unified policies and regulations, on the one side TRNs are dealing with distributive problems where members should come to a harmonious decision but prefer different outcomes (Verdier, 2009). Distributive conflicts are consequences of “new standards, as the costs and benefits of alternative proposals fall on different states” (Verdier, 2009, p. 115). Distributive problems make it harder to come to an agreement of cooperation because during the negotiations state A may try to wait for state B to settle for the outcome that state A prefers. It is possible that states will compensate the state B with side payments, which can be money or to agree to help state B in an area where they have a strong preference outcome (Verdier, 2009). In this way, powerful states can use this strategy to their advantage because they have more ways to “compensate”. To solve these conflicts states are forced to make concessions and trade-offs which can lead to making threats and other political ways of showing power during international negotiations which are contradicting with TRNs being apolitical (Verdier, 2009).
Furthermore, TRNs have enforcement problems as their informal nature limits them from monitoring and enforcing rules which lead to noncompliance when individual members can gain by not implementing the solution that was decided on together decreasing the effectiveness of transnational regulation (Verdier, 2009). Enforcement problems, however, arise when states see more benefits in the short-term and therefore give up on implementing decisions for the long-term. It is essentially freeriding which could be tackled by developing strategies to motivate states
to adhere to the rules. The success of these strategies can only be reached when members know what they can and have to do when other members freeride. Only when members are informed trust can be build and cooperation will be easier to attain (Verdier, 2009). Distributive and enforcement problems are not mutually exclusive, states can have both types of problems at the same time. In other words, in this game of implementing new regulation standards, there are clear winner and losers. Meaning that although TRNs or ERNs try to be apolitical, distributive conflicts and lack of tools for enforcement create a situation where states will create a situation of moral hazard, members of a TRN will take more risks with their actions such as non-compliance. Technical solutions that are not influenced by politics seem to be an illusion.
In the previous parts, the advantages and limits of regulatory networks have been discussed. It is clear that although the advantages such as flexibility and horizontal informal structures are needed to tackle today's societal problems, a lot can be said against the use of networks. The number of regulatory networks has been growing and the EU has used them to solve implementation problems but it is not clear if these networks are actually effective in supporting regulation and policy implementation in the EU. In general, measuring the effectiveness of networks and especially government networks have been proven to be very complex, because of multiple reasons such as having to deal with multiple constituencies (Provan & Milward, 2001) and multiple layers of government. Moreover, there is still no consensus on what network effectiveness is and how to measure it (Raab, Mannak, & Cambré, 2015). Therefore, there are two problems. First, it is not clear under what conditions regulatory networks can be effective and second it is not clear how network effectiveness should be defined. Therefore in the next section, a theory will be put to the test that has an answer to both these problems. It will be tested whether or not this theory can help us to understand how the effectiveness of ERNs works.
Theoretical framework
O'Toole (1997, p. 45): “Networks are structures of interdependence involving multiple organizations or parts thereof, where one unit is not merely the formal subordinate of the others in some larger hierarchical arrangement.”. Networks are not a new invention and there are many types of networks which all have different intentions (O'Toole, 1997). In general, all networks have a set of characteristics that make a network a network instead of an international organization (IO), characteristics such as “participation, process, enforcement, and institutionalization” (Newman & Zaring, 2013, p. 245). A network exists of multiple organizations that are interdepend because they cannot reach certain goals on their own and therefore collaborate in a non-hierarchical structure.
Network effectiveness
Research has been done and theories have been constructed that discuss network effectiveness and its determinants. Scholars that have been influential in the field of public service delivery network effectivity are Provan and Milward (2001) who discussed this topic in their work where they proposed a way to evaluate networks in the public field. Public service delivery networks are: “Networks, consisting of three or more organizations that consciously agree to coordinate and collaborate with one another” that “are used to deliver services, address problems and opportunities, transmit information, innovate, and acquire needed resources.” (Kenis & Provan, 2009, p. 440) The problems they saw with evaluating networks is that there is a popular belief that indeed when you work together between interdependent organizations the outcomes will be more effective (Provan & Milward, 2001). This belief stems from the concept of game theory where you receive better result when you work together than when you compete with each other (Provan & Milward, 2001). However, because for public organization network there are few data that can be linked to their outcomes. Therefore one cannot say networks are an effective way to deal with complex problems (Provan & Milward, 2001). Moreover, when assessing the effectiveness of a network you have to take into account multiple organizations which makes the assessment even more difficult (Provan & Milward, 2001).
Provan and Milward (2001) conclude that even though it is hard to evaluate networks, also because of the diverse interests of the multiple stakeholders, is important. To accommodate these diverse interests, they conclude that public networks should be analyzed on three different levels, community, network, and organization/participant levels. Community-based networks are those that are focused on making a contribution to the community they serve, the local area (Provan & Milward, 2001). Network-level focusses on the inter-organizational entity and participant or organizational level analysis is focused on the management of an individual agency (Provan & Milward, 2001). A network is then effective when the network satisfies the interests and the needs of those who support and work for the activities of the network (Provan & Milward, 2001). Furthermore, a network that is built on cooperative inter-organizational relationships that collectively provide services is effective when it is more effective and efficient than an organization that has fragmented funding and services (Provan & Milward, 2001, p. 422). Lastly, they argue that even though there will be problems such as disagreement on goals and methods, organizers of public-sector networks should in the least “satisfy the needs and interests of stakeholders at network and organization levels, while emphasizing the broader needs of the community and the clients the network must serve.” (Provan & Milward, 2001, p. 422)
In a later article written by Provan and Kenis (2008) they go deeper into the network level discussed above. Furthermore, in their research, they go deeper into the functioning of the network and the definition of network effectiveness. Which they define as: “the attainment of positive network level outcomes that could not normally be achieved by individual organizational participants acting independently” (Provan & Kenis, 2008, p. 230).These positive network level outcomes can be reached when network-level contingencies and tensions are taken into account during the creation and development of a network (Provan & Kenis, 2008). This definition and the network-level analysis approach will be used. In the article by Provan and Kenis (2008) they study the governance perspective which had not been well researched. Because there was the common belief that as networks are collaborative instead of hierarchical and controlled which is implied by governance, governance would not make sense to study (Provan & Kenis, 2008, p. 230). Besides networks not being hierarchical they are also not legal entities, therefore, governance is not a legal necessity for networks to have.
However, argue Provan and Kenis (2008), if a network is goal-directed and has a certain identity a form of governance is needed to coordinate the engagement of participant when their activities have to be: collectively and mutually supported, to address conflict, and effective and efficient management of resources (p. 231). Governance is therefore critical for effectiveness (Provan & Kenis, 2008). Provan and Kenis (2008) treat a network as having its own forms of governance as they use the view that a network is a mechanism of coordination which has its own structural characteristics and patterns of relations. They then argue that the different types of network governance depending on certain conditions will have a certain effect on network effectiveness (Provan & Kenis, 2008). Moreover, they argue that besides these conditions management has a significant effect on the effectiveness of networks and they propose certain tensions that need to be well managed for considering the type of network governance to be effective.
Although, it is not feasible for a master thesis to look at a network from all three perspectives as discussed by Provan and Milward (2001) in the case of ERNs the community are the MS of the EU while they are also the ones participating in the network. Therefore, by looking at why they need the network through the lens of ‘the need for network-level competencies’, the community-level besides the network-level is slightly covered. This thesis will cover a theory that focusses on public networks which are mainly public service delivery networks and how they can be effective. This theory will be used because the scholars in this field are one of the few that have focused on whole network effectiveness regarding public organizations as shown above. The
theory by Provan and Kenis (2008) has been scrutinized by other scholars and has been highly cited in the field of public administration. A lot of networks have been established in the last decades, yet there is still a lack in the research in which the network itself is the unit of analysis (Whelan, 2015). Moreover, Provan and Kenis are one of the first scholars that started to focus on whole network performance of which their most known article (Cristofoli, Markovic, & Meneguzzo, 2014) is the one that is discussed above and will be the theory that is tested in the next sections.
Contingencies
First, Provan and Kenis (2008) name four factors which depending on the type of governance should have a different level for that governance type to be effective, they call them the key predictors of network effectiveness. These key predictors together with the governance forms are the contingencies that make up the environment of a network as described by Provan and Kenis (2008) and the ‘correct’ combination and levels of these contingencies are the conditions for a network to be effective. When something changes in the environment of the network, internally or externally, this has an effect on how a network should be governed. If the network remains static and does not amend how they govern the network, the network will have a harder time reaching its goals. Thus to say, when these contingencies change from for example high trust to low trust in the network, this will have a high impact on the effectiveness of the network governance. This impact on network governance can throw the network out of balance and it will become less effective.
The types of governance are divided by the level of how brokered they are, or to what extent they have a central organization that takes the responsibility of the network governance and administration. When a network does not have any kind of broker and the members of the networks are governed by themselves it is called shared governance or participant-governed networks (Provan & Kenis, 2008). The next level is when there is a member of a network that takes responsibility for some of the governance tasks, which would be a lead organization (lead organization-governed networks). Lastly, when a network has a centralized administrative body that governs the resources and activities of the network which especially set up to do so, then we are talking about a network administrative organization (NAO). (Provan & Kenis, 2008) The type of governance and the results could be taken into account depending on the goals you want to reach as a network.
These four predictors depend on structural and relational contingencies which are “trust, size or number of participants, goal consensus, and the nature of the task” (also named the need
for network-level competencies) (Provan & Kenis, 2008, p. 237). First of all, Provan and Kenis (2008) also argue that to understand trust on a network-level we must look at how trust is distributed and if the members of the network trust in each other but also the network. Depending on the density of trust in a network can be effective or not. The trust should be based on a relationship that is more than just “being in the network” but it does not have to go very deeply either So when a network has a low level of trust but is based on shared-governance the network won’t be effective. Therefore when a network has many members and the trust density is low the network should opt for a brokered type of governance to be sustainable. When an organization is fully brokered such as an NAO trust in a network doesn’t have to be as high as the other forms of governance as every member is able to monitor the NAO (Provan & Kenis, 2008). A high level of trust together with a long history of personal contact has also shown a positive influence on effectiveness (Cristofoli et al., 2014; Turrini, Cristofoli, Frosini, & Nasi, 2010). Moreover, trust gives networks internal stability and is a contribution to reciprocity (Turrini et al., 2010). Trust is also a very important factor for ERNs, only when there is trust co-ordination can be successful and would they be willing to exchange information which is important when the structure is non-hierarchical (Van Boetzelaer & Princen, 2012).
Secondly, when a network grows in its number of participants the links between them become more and more complex and so does its governance. Introducing a broker organization and thus centralizing governance might be a solution as the broker can act as a mediator between the participants. However, when the group of participants is small introducing a broker is not the best choice and the network is better off with governing themselves. (Provan & Kenis, 2008). Although the effect of size is not certain in general it has been found that the bigger the network a lower level of effectiveness is perceived (Turrini et al., 2010). Third, it is important to have the consensus on goals as a network, because consensus has better results regarding effectiveness in general than conflict (Provan & Kenis, 2008). Goal consensus can also be a less attainable goal because of political interest and the political salience countries have on certain issues. As discussed previously, when many countries come together with different preferences coming to a decision becomes more difficult.
Lastly, every organization in a network has a reason to participate in it, one reason is that some goals cannot be reached on their own and have therefore need network-level competencies. Thus, the network needs to question itself what the nature of their tasks are and what the external demands are. It depends if these tasks are more needed in an individual organization or network-level. When the tasks and external demands are focussed on the network-level than having an NAO instead of shared governance, for example, will be more effective. (Provan & Kenis, 2008)
Summarizing the criteria for the governance structures on the four predictors the assumptions that are made by Provan and Kenis (2008) are as follows:
Table 2 four predictors of network governance effectiveness (Provan & Kenis, 2008, p. 237)
Governance forms/ key predictors
Trust Number of Participants or size
Goal consensus Need for network-level competencies
Shared High density Few High Low
Lead organization Low density moderate Moderate to low Moderate Network administrative organization Moderate density Moderate to many
Moderately high high
However, there will be one addition to this table as an extra variable to accommodate the political nature of governments working together. As in the literature section regarding ERNs has been discussed, the expectation is that because there are governmental agencies involved they are assumed to be influenced by political interests. In the following section, political interest or political salience will be discussed as an extra possible factor that influences network governance and therefore network effectivity.
The main hypothesis can be derived from this theory section. If a certain regulatory network is described through these contingencies and then scores according to the correct level of their presence such a network would be effective. Therefore the main hypothesis is:
H1: If the conditions of the four key predictors in relation to form of governance are met, a
network is effective.
The factors
ERNs are mainly networks that are reliant on some kind of administration. They will either have their own secretariat or administrative body (Levi-Faur, 2011). However, this secretariat is in some cases part of the EU and in this case, the network will be dependent on the EC (Levi-Faur, 2011). Shared-governance is, in general, an unsustainable option for ERNs as shown in the table the size should be small for the network to be effective and ERNs will have to at least include the 28 MS of the EU. It is also not likely for one country to take the responsibility for the administration of the network considering this would be a political decision and some states would not have the resources to do so, thus a lead-organization form governance is also not sustainable. However, having a lead-organization type governance for an ERN could be the case in practice. Yet, having
found that ERNs often have a secretariat or administrative body, either independent or dependent on the EC, ERNs can be considered as the governance form of an NAO. The four factors also correlate with each other, for example, because the size of a network increases it is harder to reach consensus on the network goals, yet the need for network-level competencies increases but creating trust (or ties) between members becomes more difficult. All the factors explained further below need to be in a certain balance for the type of governance to be effective as has been shown above.
Trust
Most interactions or exchanges between humans are based on trust. In the setting of an organization, trust is important for “communication, conflict management, negotiation processes, satisfaction, and performance” (McEvily, Perrone, & Zaheer, 2003, p. 91). Trust has also been identified as an internal factor for coordination that was found to have an effect on the effectiveness of ERNs (Van Boetzelaer & Princen, 2012). Because without trust actors would not be willing to engage in the exchange of information (Van Boetzelaer & Princen, 2012). The level of trust between members within a network is, therefore, a very important factor for network effectiveness. Trust in as an organizing principle can be explained as “the willingness to accept vulnerability based on positive expectations about another’s intentions or behaviors” (McEvily et al., 2003, p. 92). With the idea that one will not exploit the other decision-making becomes more efficient as one can trust that the information they receive from others is reliable (McEvily et al., 2003).
H2: Trust in an ERN should be at least moderate for it to be effective.
Size
The size of a network has an influence on the density of trust, the difficulty to create goal consensus, and make decisions. When the size of a network increases it becomes more difficult to connect and trust every member that joins the network. Moreover, the more members a network has, the more the network has to deal with different preferences. Networks, especially ERNs are based on unified decision-making. The number of members and opinions will make it more difficult for the network to reach a decision that every member agrees upon, thus affecting the difficulty of goal consensus and slowing down the decision-making process. In general, the bigger a group the more opinions which leads to a less efficient decision-making process and this affects effectiveness. Furthermore, ERNs have members that are based on different cultures, the heterogeneity of the network in size a composition may affect the effectiveness of the network.
However, there is not a lot known about what the effect of size is on effectiveness. Provan and Kenis (2008) suggest that it depends on the form of governance but in general, the bigger the network the less effective it will be if it does not change to a structure that is fit for the number of members.
H3: The more members involved in a project the less effective the decision-making within the project.
Goal consensus
For a network to work effectively the members need to agree on the goals the network is trying to achieve to a certain extent depending on the form of network governance. The main argument is that when members of any type of organization are in consensus and there is “domain similarity” there will be better results than when there is conflict (Provan & Kenis, 2008). In the case of an NAO which is the preferred form of network governance of ERNs, goal consensus needs to be moderately high. Participating in a network means that at least you joined because you have a common goal you want to achieve and are interdependent on others to achieve this goal. To reach goal consensus it is important to have similar interests between the members and a few conflicts. Essentially networks try to foster coordinated “decision-making and joint action among autonomous entities with distinctive aspirations, operational goals, and organizational characteristics” (Saz-Carranza & Ospina, 2010, p. 328). Besides goal consensus, but can be treated as very similar or is part of goal consensus, it is important to consider organizational homophily, which means that actors form network relationships because of similarity between organizations (Provan & Kenis, 2008). Moreover, it needs to be considered that goal consensus is different from trust, because trust depends on previous interactions and experience while consensus is based on how similar one's goals are (Provan & Kenis, 2008).
H4: The members of the network ERN have to agree on the goals of the network in a moderate to a high degree for the network to be effective.
Political salience
Goal consensus is important also for ERNs, however, because ERNs are tied together with the EU, the EC, and the members are nation states political salience or political interests could play a role in the decision-making and participation within an ERN. As discussed in the previous section about regulatory networks the biggest disadvantage of transgovernmental networks is that political interests influence the decision-making, making it less quickly than other types of networks and allows for more powerful states to get their way through bribing less powerful
states. For example, more powerful states can convince less powerful states to vote for projects that are preferred by the powerful state. When this is the case there will be members that do not want to cooperate in the network anymore or do not support it. Furthermore, conflicting political interests can also result in less active members because when a member does not agree with the activities of a network for political reasons it will not participate. When fewer members support the goal which is legislation harmonization will be less effective, because of lack of information and willingness.
H5: If political interests are highly prevalent in an ERN effectiveness decreases.
Need for network-level competencies
Networks are created because there is a reason for organizations or countries to come together, one of these is the need for network-level competencies. These competencies allow single organizations to achieve goals or solve problems that are not possible when they would have tackled these goals and/or problems by themselves. The need for network-level competencies is also highly related with goal consensus. Wanting to participate in a network essentially means you agree on a goal that you want or can only reach by tackling it collectively. To know what competencies are needed for the network in the interest of the members we need to know what the internal and external demands are for a network. Internal tasks have to do with the coordination of members and management while external tasks work on the outside image of a network by lobbying, seeking new members, get funding, etc. Moreover, it means to protect the network from external shocks such as new regulations (Provan & Kenis, 2008). Which is essentially bridging, framing a capacitating as discussed by Saz-Carranza and Ospina (2010) which will be discussed in the section about management.
The first internal task that needs to be considered is interdependence. When the network needs interdependence between members to achieve network goals than the need for network-level competencies is higher, because the members need a network that coordinates their activities and has task-specific competencies (Provan & Kenis, 2008). Although, Van Boetzelaer and Princen (2012) identify interdependence as an external factor for the effectiveness of a network they then proceed to consider it as an effect on coordination. Which is in the eyes of Provan and Kenis (2008) an internal task. Van Boetzelaer and Princen (2012) define interdependence as the dependency between the network and the EU, between the participants and their national governments, and the interdependence between participants. Which in this case interdependence could also be considered an external task of a network.
Interdependence was found to be the most prominent incentive for actors to make coordination as effective as possible (Van Boetzelaer & Princen, 2012). Co-ordination might increase when interdependency between national regulatory agencies (NRAs) and ERNs are higher, furthermore goal consensus or how Van Boetzelaer and Princen (2012) defined it, level of uniformity, also seems to have an effect on the relationship between interdependence and coordination. However, there is a clear positive correlation between interdependence and coordination (Van Boetzelaer & Princen, 2012, p. 830). Coordination should be considered as a network-level competency for ERNs as Van Boetzelaer and Princen (2012) argue that coordination is needed for effective implementation of EU-law. The need for network level competencies is therefore very high as the ERNs’ main goal is to effectively implement EU-law and regulations. Turrini et al. (2010) are naming some tasks of networks that can be categorized as internal or external tasks. They found that there are structural characteristics of a network that has an effect on effectiveness in the literature they studied. Namely external control, integration mechanisms and tools, size and composition, formalization and accountability, and network inner stability. External control refers to the dependencies of the network, such as its constituency and the context in which the network is active. An example is if the state has certain fiscal policies or regulation by which the network is limited in taking action it directly affects the extent to which a network can be effective (Turrini et al., 2010).
External control as defined by Turrini et al. (2010) can be considered as an external task, protecting the network from external shocks, underlining the importance of protection as an external task. Second, integration is one of the more important factors that have been found to affect the effectiveness of a network. Integration is the interconnectedness and cohesion of the members of the network (Turrini et al., 2010). If integration is coordinated through a central agency it is more effective than when organizations organize multi-lateral agreements (Turrini et al., 2010, p. 541). Integration that actually supports effectiveness needs certain mechanisms of coordination which are “common information and communication systems, joint staff activities such as marketing, funding and planning, and integration at the service level such as the creation of one-stop, integrated service capacity for clients” (Turrini et al., 2010, p. 541). These mechanisms of coordination should be considered when looking for the internal tasks of a network during the analysis stage of this thesis. Formalization and accountability refer to the functioning of the network as a whole, such as who is accountable to whom and the transparency of the network. The effects of these factors are dependent on the level of integration of the network (Turrini et al., 2010). However, accountability of the network towards external stakeholders forces the network to be transparent about their results and thus forces the network to evaluate the progress of projects and results. The last mechanism Turrini et al. (2010) found is
the creation of synergy between members through sharing knowledge, perspectives, and skills which are important for increasing the capacity of the network to reach network goals. Formalization is, therefore, an external task of a network as it focusses on the perception of outside stakeholders.
Overall, interdependence, integration, and formalization add to creating a level playing field for the members of the network. A level playing field is when all members of the network stand on equal ground. For ERNs this means that each MS has harmonized its legislation with all other MS where every member has similar ways of implementing and enforcing legislation. The importance of this is that when a MS does not see the benefits of participating in an ERN, for the above-mentioned reasons. The MS will hinder the ERNs ability to support the creation of a level playing field. Because, when one member does not participate in complete harmonization of legislation in the EU cannot be attained. Therefore, when a MS has a need for these network-level competencies the chance is higher for the network to reach its goal which is to support the harmonization of EU legislation, thus the hypothesis is:
H6: A high need for network-level competencies positively affects network effectiveness.
Management
The need for network-level competencies is closely related with the management of a network. However, there are some tensions that need to be considered by the management of a network to execute good governance after the contingencies are established. Provan and Kenis (2008) discuss three tensions in their article which need to be taken into account when the type of governance structure has been chosen and how the balance of the four factors can be maintained. Whenever there is a mismatch between the contingencies and the governance structure these tensions become hard to manage. In this sub-section, these tensions will be discussed and suggestions on how to manage them. For the sake of this research, these tensions need to be considered as a tool to find indications of problems leading to effectiveness in the day to day work of the network. The four (or five including political salience) contingencies discussed above might be vague concepts to discover in the interviews with professionals working with IMPEL and in the documents of IMPEL.
The first tension can be related to goal consensus and is the battle between having an efficient administration and inclusive decision-making. The point here is that inclusive decision making will help with maintaining and gaining trust which is important for every network (Provan & Kenis, 2008) This tension is similar or the same as the unity versus diversity tension
