MMartina Fuchs /s1902407)
Master Thesis June 11, 2o18
Determinants of Effectiveness in Networked Modes of
EU External Governance
Do Economic Conditions Matter?
A qualitative comparative case study of the civil aviation authorities of Moldova and Ukraine within the scope of their cooperation with EASA
Abstract
European regulatory networks (ERNs) have already received much scholarly attention. Owing to a growing body of literature that is making inquiries into the conditions of ERN’s effectiveness, for instance, their role in the implementation of EU policies is now widely acknowledged. Recently, ERNs have been developing an external dimension: Functional considerations are prompting EU agencies to increasingly open their doors to the inclusion of non-EU member state regulators to address issues of sectoral interdependence. However, in contrast to networks of EU agencies and member states’ national regulatory agencies (NRAs), at this point we still know only little about networked modes of EU external governance. What effects do they yield? What are the conditions of their effectiveness? This thesis advances the knowledge of ERN’s external dimension. The study argues that previaling economic conditions in non-EU member states affect the effectiveness of networked modes of EU external governance and illustrates its claim with a small-N comparative case study that examines the cooperation of the European Aviation Safety Agency (EASA) with aviation experts of Moldovan and Ukrainian civil aviation authorities. Findings indicate that networked cooperation of EU regulators and non-EU member states’ regulatory agencies is less successful when ENCs' economic conditions are adverse. In turn, the extent of regulatory convergence increases in correlation with better economic conditions.
Key words: EU external governance, European Regulatory Networks, European
Neighbourhood Policy, network governance, network effectiveness, regulatory convergence
Master Thesis August 10, 2018
Leiden University
Institute of Public Administration
Faculty of Governance and Global Affairs Supervisor. Dr. Dovilė Rimkutė
Martina Fuchs (s1902407) Contents Acronyms ... i Introduction ... 1 Literature Review ... 8 Theoretical framework ... 19 Methodology ... 23 Analysis ... 39 Discussion ... 62
Limitations and Future Research Avenues ... 68
Conclusion ... 71
[i]
Acronyms
CEPOL European Union Agency for Law Enforcement Training
CSU Christlich Soziale Union (Germany’s Christian Social Union Party)
DREF Disaster Relief Emergency Fund
EaP Eastern Partnership
EASA European Aviation Safety Agency
EPAS European Plan for Aviation Safety
EMCDDA European Union Agency for Law Enforcement Training
ENC European Neighbourhood Countries
ENP European Neighbourhood Policy
ERN European regulatory network
EU European Union
FDI Foreign Direct Investment
FRONTEX European Border and Coast Guard Agency
GDP Gross Domestic Product
IMF International Monetary Fund
IRA Independent Regulatory Agency
IFRC International Federation of Red Cross and Red Crescent Societies
NPM New Public Management
NRA National regulatory agency
OECD Organization for Economic Co-operation and Development SAAU State Aviation Administration of Ukraine
SBA Stand-by Arrangement
TAIEX Technical Assistance and Information Exchange instrument
TGN Transgovernmental network
[1]
Introduction
European regulatory networks (ERNs) have received scholarly attention across different strands of literature. For example, in Europeanization research, scholars have advanced theories to explain network emergence (Eberlein & Grande, 2005; Coen & Thatcher, 2008; Eberlein & Newman, 2008), whereas Public Administration research has focused on the effects of network membership on participating regulatory bodies (Eberlein & Newman, 2008; Yesilkagit, 2011; Danielsen & Yesilkagit, 2013; Bach, Ruffing & Yesilkagit, 2015; Mathieu, 2016). Recently, EU external governance research has also begun to examine networked forms of collaboration between European Union (EU) agencies and regulatory authorities of non-member states. In this emerging field, transgovernmental regulatory networks (TGNs) that encompass EU regulatory agencies and their counterparts from the European Neighbourhood Countries (ENCs) are figuring prominently. Studies show that to address complex regulatory problems arising from transboundary interdependence, EU agencies increasingly extend their organizational boundaries and integrate regulators from the sixteen countries, whose external relations with the EU are managed under the roof of the European Neighbourhood Policy (ENP) (Lavenex, 2014; Freyburg, 2015; Lavenex, 2015; Rimkutė & Shyrokykh, 2017; Rimkutė & Shyrokykh, forthcoming).
Akin to the work on regulatory networks of EU agencies and member states’ regulatory bodies (NRAs), theory building efforts to account for network emergence figure prominently. Similarly to the two-level regulatory policy making within the EU (Eberlein & Grande, 2005; Eberlein & Neman, 2008; Coen & Thatcher, 2009), the extension of the EU’s regulatory scope to third countries rests on more than a single institutional solution. For example, the Europeanization of third countries can follow a territorial logic along the lines of traditional foreign policy frameworks, effectuated through official state actors that operate in a hierarchical governance system, rely on conditionality, and lever the weight of powerful external incentives to ensure non-member states’ regulatory compliance with EU-preferences (Lavenex, 2014). At the same time, aside from official foreign policy making in the EU, on the level of EU sectoral politics and in the absence of conditionality, functional considerations based on the need to address issues of sectoral interdependence with third countries drive EU regulators to integrate their foreign counterparts into transboundary technocratic regulatory networks (Lavenex, 2015; Rimkutė & Shyrokykh, 2017).
However, while theory building efforts to account for the emergence of ERNs help to situate the phenomenon of ERNs and TGNs per se and mark a point of departure for this thesis,
[2]
the conditions under which these networks emerge are not the focal point of this research. Instead, guided by the research agenda in EU external governance scholarship, I seek to contribute to the expansion of the knowledge base on the very conditions under which regulatory networks, that engage EU and ENC regulators, yield the desired results and effectuate non-EU member states’ regulatory convergence with the EU’s acquis. In EU external governance scholarship, research questions pertaining to the EU’s success at projecting its own institutions beyond the circle of member states take a central position (Lavenex & Schimmelfennig, 2009). However, while at this point we already have at our disposal an extensive body of literature on the emergence of ERNs that engage third country regulators alongside EU regulatory bodies, we know little about the conditions under which the collaboration of EU and third country regulators effectuates the transfer of EU rules beyond EU borders.
With this thesis, I aim at addressing this particular gap in the literature, and my selection of this study’s independent variable rests on conflicting claims in the literature on regulatory networks and the corroboration of one of the opposing hypotheses by tenets made in the research on public agencies. Specifically, I seek to examine to what extent prevailing economic conditions on the ENCs impact on the effectiveness of EU external governance by regulatory networks. The remainder of this section, then, proceeds as follows: firstly, I align tenets of the network literature on the relationship between regulatory networks’ supposed insulation from political pressures and network effectiveness with the agencification literature’s claim that public agencies, the network participants, remain exposed to political control by their domestic principals. Secondly, I outline the relationship between conditions in the economic environment and the implementation of control mechanisms that curb the autonomy of public agencies before presenting my research question. The section then concludes with a presentation of the research methods I employ, and a description of the study’s academic and societal value.
In a seminal study on the limitations of transgovernmental regulatory networks (TGNs), Verdier (2009), taking issue with what he considers to be an all too common argument in the relevant literature, criticizes the tenet that regulatory networks ‘work’ because they allow their members to operate in insulation from domestic political pressures (p. 115). In contrast to Verdier, I would advise caution and instead argue that even though scholars of TGNs describe how these networks have become governance institutions of the international domain and work separately from unitary states, the literature does not generally refute claims of continued
[3]
influence on regulatory networks by authorities on the national level. For instance, Slaughter (2004) holds that TGNs can take on governance tasks in the international domain when the creation of internationally valid rules is imperative but centralization of authority to establish and enforce such rules in the hands of a single government is undesirable. Indeed she argues that the change towards international governance by TGNs implies a rising legitimacy of non-state, non-elected actors that participate in regulatory networks and go about their business “apart from”, “above”, and “detached” from unitary states (p. 5; p. 7; p. 9). However, Slaughter (2004), for instance, is also quick to assert that even though TGNs take on governance duties that individual government organizations could not adequately fulfill, the ties between TGNs and national governments are never fully severed as individual network members continue to be accountable to their domestic principals (p. 18-19). Therefore, key lines of argument in the literature on TGNs do correspond with central tenets in agencification research. Public agencies, though staffed with technocratic experts rather than elected government officials, are by no means fully insulated from the controlling reach of their domestic principals (Verdier, 2009; Verhoest et al., 2010).
Conceptually, agency autonomy captures the extent to which public agencies are insulated from the preferences of their domestic principals, and it denotes the degree of discretion that public agencies possess to make all relevant organizational decisions (Groenleer, 2009). Even though scholars of agencification have highlighted the difficulties of arriving at a universal definition of public agencies (Pollitt et al., 2004), the autonomy of public agencies, and of regulatory agencies among them, stands out as a primary defining feature that sets these organizations apart from other institutions of traditional bureaucracy. For instance, the strands of agencification theory that respond to the international proliferation of the agency form and aim at accounting for their rise in popularity, rely on agency autonomy as a central explanatory factor for the trend to delegate government functions to public agencies. As an example, rational choice-based theories of agencification posit that the popularity of this particular organizational form derives from their ability to address the growing demand for high-quality public services faster, better, and at lower cost than traditional bureaucracy (Pollitt et al., 2004; Christensen & Lægreid, 2007). In this view, autonomous agencies possess discretion to make relevant organizational decisions freely rather than in consideration of their principals’ preference profiles (Pollitt et al., 2004; Christensen & Lægreid, 2007). Consequently, in contrast to the divisions of traditional bureaucracy, agencies are more successful at finetuning and setting their goals according to clientele and policy needs, and they are better at accomplishing their goals
[4]
because their disentanglement from parent ministers enables agency managers to improve their organizations’ efficiency (Pollitt et al., 2004; Christensen & Lægreid, 2007).
However, the relationship between the autonomy and efficiency of public organizations, which is advanced in conceptual definitions and figures prominently in agencification theory, might advance a misleading picture of the actual extent and consistency of the insulation from external pressures that agencies enjoy. In fact, it is a misconception to envision public agencies as organizations that, as a general rule, display invariably high degrees of bureaucratic autonomy because their tasks demand it. As a matter of fact, the extent of autonomy that agencies are vested with is highly variable across individual cases and even within single cases, it may change over the course of time (Groenleer, 2009).
In the post-creation period, the implementation of control and oversight mechanisms by agencies’ domestic principals can set the pace for the development of agency autonomy (Gilardi, 2002; Pollitt et al., 2004; Groenleer, 2009; Wonka & Rittberger, 2009). Agencies’ principals can curtail agency autonomy by obligating agencies to adhere to a range of reporting duties that require the latter to disclose and justify their operational decisions and enable the former to control and redefine the extent of agencies’ decision-making discretion in multiple areas (Gilardi, 2002; Wonka & Rittberger, 2009). Concretely, following imposed reporting duties, public agencies may be bound to transparency vis-à-vis their principals in matters pertaining to the recruitment and hiring of staff, to policy-related decisions, and to the design and use of their budget (Gilardi, 2002; Wonka & Rittberger, 2009).
As regulatory agencies engage in regulatory networks, the extent of network participants’ insulation from domestic constraints thus depends on the presence of motifs for their domestic principals to apply control and oversight mechanisms to curtail agencies’ decision-making discretion. Therefore, there is reason to assume that the presence of conditions, under which the domestic principals of network participants opt to curtail the decision-making discretion of their agencies, will affect the degree to which these networks succeed at obtaining positive network-level outcomes. Existing research has shown that certain developments and changes in a country’s economic climate provide principals of public agencies with motifs to decrease the latter’s autonomy to manage their own affairs. For instance, the presence of economic pressures act as an incentive for risk-averse politicians to resist isomorphic pressures to converge with the NPM ideal-type of autonomous agencies (Pollitt & Bouckaert, 2004). Specifically, in times of economic and budgetary-fiscal crises, risk-averse politicians tend to
[5]
exert stronger control over public agencies by curtailing their discretion to make their own decisions (Verhoest et al., 2010).
Motivated by congruent claims in the literatures on regulatory networks and agencification on the persistence of the controlling reach of network participants’ domestic principals and informed by research on the very factors that induce principals to exert their control, in this study, I seek to further examine the relationship between the conditions in the economic environment of network participants and the effectiveness of regulatory networks on the other. Concretely, I intend to study if a deteriorating economic climate affects the extent to which EU agency-mandated regulatory networks that engage ENC regulators can accomplish their goals of projecting EU rules beyond EU borders. To this end, then, I ask the following research question:
RQ: To what extent do the economic conditions in ENP countries, under which ENC and EU regulators meet, affect the effectiveness of networked forms of EU external governance?
To determine the plausibility of my argument that the economic conditions in ENCs’ impact on the effectiveness of networked forms of EU external governance where EU agencies engage third country regulators in regulatory networks, I conduct a qualitative, small-N comparative case study. Specifically, I contrast the networked collaboration of EASA regulators and Moldovan civil aviation agency with the networked cooperation of EASA and Ukraine’s State Aviation Administration. I rely on extensive document analysis to measure economic conditions in Moldova and Ukraine and the effectiveness of the networked cooperation between EASA and Moldovan and Ukrainian civil aviation agencies in effectuating the projection of the EU’s aviation acquis. The choice of methods is driven largely by the novelty of the theoretical argument that this research is taking on. Applying a covariational approach at establishing causality, this thesis leverages a case selection strategy that, guided by principles of the most similar systems design, rests on the cases’ display of similar values on previously identified control variables and warrants high internal validity of findings. In addition, a qualitative, small-N comparative study was deemed most suitable as its focus on a small set of cases allows for the collection of thick, qualitative data and intense reflection on selected indicators. In consequence, I am able to leverage low case count to increase concept validity of measurement. Finally, the chosen qualitative, small-N comparative approach facilitates the process of data triangulation, thereby increasing the reliability of the study’s
[6]
findings. The examination of only two cases allowed for a consultation of a variety of sources to measure the central concepts, thereby correcting for the effects of potential random measurement error and the introduction of systematic bias to data analysis.
The contribution I am making with this study is twofold. As described above, the study addresses a relevant gap in the research on the factors that impact on the effectiveness of regulatory networks that engage non-member states’ regulatory agencies alongside EU decision-making bodies. Furthermore, in Public Administration research, specifically, in the agencification literature, there is an overly strong focus on EU agencies (Groenleer, 2009; Groenleer, 2014), agencies of EU member states (e.g. Yesilkagit & van Thiel, 2011; Maggetti & Gilardi, 2014; Bach, Ruffing & Yesilkagit, 2015; Mathieu, 2016) and, as Yesilkagit and van Thiel (2011) point out, on North American agencies. Therefore, this thesis addresses analytical limitations of previous research by extending the focus of the agencification literature through its focus on regulatory agencies of ENCs. Moreover, this project also relates to the network literature. Here, recent stock-taking of the achievements and the distribution of findings across the network literature further confirm the value of the research question pursued with this project. In an assessment of achievements since the first publication of his seminal article “Treating Networks Seriously: Practical and Research-Based Agendas in Public Administration” (1997), O’Toole (2015) emphasizes that more research is required before we can fully understand how TGNs can accomplish their goals (p. 368). This study is taking a step in this direction.
Finally, aimed at the identification of factors that determine the effectiveness of networked modes of EU external governance, this study also has social relevance and implications for Public Administration professionals. The EU, too, is exposed and vulnerable to negative externalities that result from policy decisions of other actors in its neighborhood and beyond. For example, despite the Union’s own efforts at effectively managing and reducing safety risks in civil aviation within its boundaries, for instance through the European Plan for Aviation Safety (EPAS), aircrafts operated by Lufthansa, KLM, and Air France (just to name a few) are not the only carriers descending on EU airports. However, regulatory standards imposed on aviation industries of foreign countries may not match the EU’s sectoral preferences. Nevertheless, unilateral policy responses to safety risks, for instance in the shape of EPAS, cannot fully remedy the problem: When it comes to non-member states, the EU has no leverage to enforce compliance with its standards. Instances of EU external governance exemplify what theories of intergovernmentalism, neofunctionalism and supranationalism hold:
[7]
“governments, when becoming aware that unilateral policy options do not produce desired results, engage in multilateral co-operation to obtain economic and security benefits they could not achieve otherwise” (Schimmelfennig et al., 2015, p. 771). Existing research shows (Lavenex, 2014; Lavenex, 2016; Rimkutė & Shyrokykh, 2017) that multilateral co-operation is increasingly institutionalized in networks. Yet, in the absence of conditionality, there is real need for a thorough understanding of the conditions under which these networks operate effectively: the results they produce – in the aviation sector and elsewhere – are supposed to and can make valid contributions to international safety and prosperity.
[8]
Literature Review
In this section, I review publications in three strands of literature. Firstly, I discuss previous research on EU external governance effectiveness. Secondly, I review the literature on agency autonomy and, thirdly, on determinants of network effectiveness. The overall objective of this section is twofold. For one, with these reviews, I intend to provide the definitions of the basic concepts that guide this thesis. In addition, I am conducting extensive literature reviews to inform my case selection strategy, which I will discuss in more detail in the methodology section. The literature reviews serve to generate extensive knowledge about factors that have previously been identified as causally relevant in alternative explanations of the effectiveness of networked modes of EU external governance. Knowledge of such factors is crucial and must guide case selection to ensure that the study can make valid causal inferences about a deterministic, invariant effect of agency autonomy on the effectiveness of networked modes of EU external governance (Blatter & Halverland, 2012; Toshkov, 2016).
External governance effectiveness
With the evolvement of scholarship in this field, the conceptual definition of EU external governance has broadened over time. In accordance with the definition of EU external governance that guides early publications in this research field, the concept captures the EU’s endeavors to project its own regulatory scope beyond its formal borders and the subsequent transposition of EU rules into domestic legislation by non-EU member states (Schimmelfennig & Sedelmeier, 2004; Lavenex, 2004). However, with the advancement of the knowledge base in EU external governance scholarship, the initial, narrow conceptual definition extends. Recently, scholars have argued that the external dimension of EU internal governance can follow functionalist needs, comprises a differentiated movement of the EU’s legal and institutional boundaries, can take participatory forms, and also captures the joint production of rules rather than the transfer of a given acquis (Lavenex & Schimmelfennig, 2009; Lavenex, Lehmkuhl & Wichmann, 2009; Lavenex, 2015).
A review of the literature on EU external governance clearly shows that even though a certain research focus dominates the field and has hitherto guided a variety of academic studies, research on EU external governance still faces a range of serious challenges. In the literature on EU external governance, research questions targeting the effectiveness of instances of EU external governance occupy a central position on the research agenda (Lavenex & Schimmelfennig, 2009, p. 800). The concept of EU external governance effectiveness denotes the successful projection of EU rules to third countries beyond the circle of member states
[9]
(Lavenex & Schimmelfennig, 2009, p. 800). The literature that addresses the EU’s external relations with the ENCs and examines the conditions of external governance effectiveness, however, demarcates a fragmented research field that deals with a number of serious challenges. Since its early days, research on EU external governance effectiveness has departed from a monolithic focus on conditionality as its main independent variable and now includes a range of additional variables in explanations of governance effectiveness. Nevertheless, the field is dominated by qualitative, small-N studies that advance explanatory theories and provide evidence for plausibility by examining small sets of cases (e.g. Barbé, Costa, Herranz & Natorski, 2009; Dimitrova & Dragneva, 2009; Freyburg, Lavenex, Schimmelfennig, Skripka & Wetzel, 2009; Sasse, 2008)1. In consequence, the advancement of theoretical ideas is not matched by large-N research efforts aimed at theory-testing. Thus, in this field, the generalizability of findings remains limited.
Early publications on EU external governance in the context of the ENP posit that the policy is unlikely to result in the external projection of EU rules. Researchers point out that the ENP has been modeled on the EU’s enlargement policy and relies on conditionality as a mechanism of rule transfer. They stress, however, that compared to the EU’s enlargement politics, the ENP suffers from crucial institutional weaknesses. In contrast to the application of conditionality in the context of EU enlargement, the ENP does not tie the adoption of EU rules by third countries to similarly sizable external incentives (i.e. accession to full EU membership): even though the EU and some of the ENCs might differ on this issue, officially, accession to EU membership or its preparation has not been an objective of the EU’s external relations with non-member states in its eastern and southern neighborhood (Magen, 2006, p. 427, in Casier, 2011, p. 38; Sasse, 2008; Schimmelfennig, 2012; Schimmelfennig & Scholtz, 2008). However, following the acknowledgement that the conditionality model (Schimmelfennig & Sedelmeier, 2004) cannot account for observed variation in regulatory convergence across the neighborhood countries (Casier, 2011), a range of alternative hypotheses have been advanced to explain the effective transfer of EU rules to third countries in the neighborhood.
1 Notable exceptions are the following large-N studies: Knill, C., & Tosun, J. (2009). Hierarchy, networks, or
markets: How does the EU shape environmental policy adoptions within and beyond its borders? Journal of
European Public Policy, 16(6), 873-894. doi: 10.1080/13501760903088090; and Freyburg, T. (2015).
Transgovernmental network as an apprenticeship in democracy? Socialization into democratic governance through cross-national activities. International Studies Quarterly, 59(1), 59-72. doi: 10.1111/isqu.12141
[10]
Akin to the conditionality hypothesis, the institutionalist hypothesis and the power-based hypothesis (Lavenex & Schimmelfennig, 2009) focus on the impact of alternative, given EU-level factors on the effectiveness of the Union’s external governance. The institutionalist explanation stipulates that the quality of existing EU institutions determine the extent of their effective transfer, positing that highly legalized and legitimized EU rules are more likely to be transferred effectively (Lavenex & Schimmelfennig, 2009, p. 802).
The institutionalist and conditionality hypotheses compete with a range of domestic-structure hypotheses that look at non-member state-, country-level factors to account for third countries convergence with EU rules. As an example, the misfit hypothesis argues that a lack of fit or congruence of the EU’s and non-member states’ domestic institutions is a necessary condition for convergence with EU rules to occur (Börzel & Risse, 2000).
The power-based hypothesis is supported by research conducted by Ademmer and Börzel (2013). A comparative study of Georgia and Armenia shows that policy-specific conditionality acts as a driver of non-member states’ convergence with EU institutions. However, this study also shows that the relationship between policy-specific conditionality and convergence is moderated by preferential fit of incumbent elites’ political agendas with EU policies (p. 581), which results in stronger compliance.
Other researchers have argued that the effectiveness of EU external governance is mitigated, if not inhibited, by the influence of alternative governance providers that compete with EU influence. Dimitrova and Dragneva (2009) research EU external governance in the case of Ukraine and find that in three different policy sectors, albeit to varying degrees, the effectiveness of EU external governance is constrained as rule adoption in Ukraine is influenced by interdependencies with Russia. For instance, they show that in energy policy, even though competing CIS institutions are soft and thus encouraging Ukraine’s departure towards EU institutional arrangements, strong infrastructural interdependence with Russia enhances Russia’s influence over Ukraine and limits the scope of EU external governance in this case. Ademmer and Börzel (2013) demonstrate that Russia used incentives based on gas price and expressed commitment to co-operation with Armenia to lever its influence, thereby constraining the effectiveness of EU external governance in Armenia in spite of a preferential fit between the EU’s demands for energy policy and the political agenda of the Armenian government.
[11]
The institutionalist, power-based, and domestic structure hypotheses each reflect a different theoretical approach to external governance effectiveness. However, researchers have previously argued that any relationship derived from these hypotheses will be moderated by the mode of external governance employed. Lavenex and Schimmelfennig (2009) posit, that hierarchical governance will be the governance mode most likely to result in convergence with EU norms. However, the EU exerts external governance not exclusively in a hierarchical manner (Lavenex, 2008; Lavenex et al., 2009). On the sectoral level, EU external governance engages non-member states in horizontal and participatory ways as third country representatives gain access to EU decision-making bodies such as EU agencies (Lavenex, 2015).
The research on the relationship between networked modes of EU external governance and external governance effectiveness is still limited. Yet, in spite of previous assumptions, it has been shown that networked modes of EU external governance have impacted positively on third countries’ alignment with EU rules. In the ENP context, Freyburg (2015) studies Moroccan state officials and asks if their involvement in Transgovernmental Networks, specifically in the EU’s Twinning Program, affects democratic socialization (p. 59). The study draws on the literature on international regulatory networks, as well as on international socialization and democratic diffusion studies, two strands of the International Relations literature (p. 60). According to Freyburg, network involvement and democratization are causally linked because networks provide an opportunity structure for deep, repeated social interaction of network participants, which spurs social influencing processes, thereby facilitating learning, emulation, and, consequently, attitude change in favor of democratization. In one of the very few quantitative studies in the field, Freyburg uses cross-sectional survey data and tests it by multiple regression analysis to examine if integration in networked governance systems results in democratization. Based on the argument in the international socialization literature that contact represents a necessary but not sufficient condition for socialization, whilst the quality of contact constitutes an important determinant for results (p. 62), Freyburg controls for the influence of prior exposure through study abroad programs or use of Western media. The study then finds that, in the case of Moroccan state officials, Transgovernmental Networks do indeed have a positive effect on their attitudes towards democratic governance (p. 69).
Turkina and Postnikov (2014) conduct a comparative study and look at Ukraine, Russia, and Moldova to investigate if cross-border inter-firm networks of EU and non-member state
[12]
business organizations further regulatory convergence with EU norms. They find that whilst these networks are indeed conductive to regulatory convergence, network involvement alone cannot account for a lack of uniformity in convergence-related outcomes. Turkina and Postnikov (2014) show that the relationship between participation in cross-border inter-firm networks is conditioned by network density, the extent of local authorities’ participation in the networks, and the extent of institutional distance between the EU and non-member states under study.
As indicated earlier, research on EU external governance effectiveness with respect to the ENP-countries faces a number of challenges and has at time seemed to stall. These circumstances may relate to the fact that a considerable part of previously conducted research either has a strongly or purely theoretical approach (e.g. Lavenex & Schimmelfennig, 2009) or presents results based on small-N studies. At this point, quantitative large-N studies might aid assessments of the value of single hypotheses by testing them over a broader range of cases. In addition, researchers of EU external governance have been criticized for negligence as their work, in some instances, omits to discuss motivations for research design, case selection, and the inclusion of control variables (Ladi, 2012).
Agency autonomy
Autonomy is a necessary condition that public bodies need exhibit to be classified as an agency (Yesilkagit, 2004, p. 120; Groenleer, 2009, p. 18). The concept refers to a public agency’s discretion to manage its affairs independently and in line with the priorities they set for themselves (Verhoest et al., 2010, p. 18-19). However, fully autonomous agencies do not exist. Agency autonomy is a complex concept: it is multi-dimensional and continuous (Verhoest et al., 2010; Groenleer, 2009; Groenleer 2014).
Firstly, the decision-making competencies of public agencies cover multiple dimensions. The literature commonly distinguishes between legal- (Groenleer, 2009), personnel- (or HRM-), financial-, and policy autonomy (Groenleer, 2009; Verhoest et al. 2010). Legal autonomy captures agencies’ delegated ability to act as an autonomous legal personality, to decide independently on entering into agreements with other organizations, for instance to procure and provide goods or services, or to establish and cultivate diplomatic relations (Gilardi, 2002, p. 880; Verhoest et al., 2004, p. 104; Groenleer, 2009, p. 32; Busuioc, Curtin & Groenleer, 2011, p. 851; Groenleer 2014, p. 260).
[13]
Public agencies that possess HRM autonomy have the discretion to recruit, train, promote, pay and dismiss their staff (Groenleer, 2014, p. 260; Verhoest et al., 2020, p. 19; Busuioc, Curtin & Groenleer, 2011, p. 851). This conceptual dimension also captures regulations concerning the appointment and term of office of senior agency executives. For instance, agencies with high degrees of HRM autonomy can freely decide on how to develop and manage their body of staff in four relevant areas. Firstly, such agencies can freely decide when and whom to newly recruit to join the organization, and, secondly, when and whom to dismiss from their services (Groenleer, 2009, p. 32). Thirdly, agencies with high HRM autonomy have the capacity to decide independently on how to socialize and develop their body of staff through additional training (Groenleer, 2009, p. 32). Finally, high HRM autonomy enables agencies to decide unrestrainedly on height and adaptation of the salaries it offers organizational members (Groenleer, 2009, p. 32).
Financial autonomy refers to agencies’ discretion to raise their own financial resources, and to decide how and where to use them in their budgets (Gilardi, 2002, p. 883; Groenleer, 2009, p. 32; Verhoest et al., 2010, p. 21; Busuioc, Curtin & Groenleer, 2011, p. 851). Comparatively, agencies, whose financial resources are not allocated annually by their parent ministries, that produce their own income at least partially and possess authority over budget planning and allocating within the organization command higher degrees of financial autonomy (Groenleer, 2009, p. 32). In contrast, agencies that are financially dependent on their principals and accountable in matters of budget design have lower levels of financial autonomy (Groenleer, 2009, p. 32).
Policy autonomy is understood as agencies’ leeway to decide on how to prioritize their tasks and goals, the working methods and policy instruments they use, on their target actors, and to draw their own conclusions or formulate their opinions (Groenleer, 2009, p. 33; Verhoest et al., 2010, p. 23; Busuioc, Curtin & Groenleer, 2011, p. 851;
Agency autonomy is also a continuous concept and the literature differentiates between an agency’s formal (de iure) autonomy on the one hand and its actual (de facto) autonomy on the other. Agencies’ formal autonomy refers to the decision-making discretion principals grant public agencies by design and at the time of their creation (Verhoest et al., 2004, p. 104; Groenleer, 2009, p. 33; Groenleer, 2014, p. 259; Verhoest et al., 2010, p. 18-19; Busuioc at al., 2011, p. 852). Actual autonomy captures the continuity aspect of agencies’ discretion, and it indicates to the existence of certain factors that, in the post-creation period, can impact on and
[14]
alter the extent of autonomy that agencies have been vested with at the time of their creation (Groenleer, 2014, p. 258). In other words, the degree of decision-making discretion delegated to public agencies at the time of their creation is not necessarily static, and whilst it may not change over time as a general rule, it most certainly can.
Supranational determinants of agency autonomy
The notion of agency autonomy as a continuous concept is reinforced by empirical research. Here, agency autonomy is often treated as the dependent variable. One important body of literature looks at supranational determinants of agency autonomy and studies the effects of the participation of EU member states’ national regulatory agencies in ERNs on the autonomy of national regulatory agencies (NRAs). Maggetti (2007) conducts a panel-analysis of 118 independent regulatory agencies (IRAs) and triangulates quantitative data with 20 semi-structured interviews to examine if IRAs participation in ERNs positively affects organizational size (H1) and delegation of more regulatory powers to the former (H2). The study finds that the position of IRAs in opposition to their domestic principals is bolstered owing to network membership.
In a single case study of the Dutch data protection authority, Yesilkagit (2011) finds that the extension the agency’s autonomy vis-à-vis its domestic principals is indeed a result of the agency’s participation in an ERN. However, he also shows that in the case of the Dutch data protection authority, network membership interacted with the transposition of EU legislation into Dutch domestic law to produce the outcome.
Bach and Ruffing (2013) study German federal agencies to establish whether their involvement in ERNs effectuates an increase in their policy autonomy. This study occupies a unique position as the authors adopt a differentiated perspective on ERNs. Distinguishing between a sectoral and an intergovernmental logic of network organization, the authors employ a network typology and study the relationship between network engagement and agency autonomy for ERN-types that differ with respect to their lead organizations. The study’s results indicate that participation in agency-led networks (as opposed to ministry-led networks) is linked more strongly to the enhancement of German NRAs’ policy autonomy. Finally, the study also suggests that agencies’ policy autonomy increases with the depth or intensity of their engagement in ERNs.
Whilst studies on the relationship between NRAs’ autonomy and their participation in ERNs demarcate an important field in agencification research, some of the key findings are established
[15]
by single-case studies. To demonstrate broader generalizability of these results, the field would benefit from quantitative large-N studies that address the same research questions. Furthermore, existing research displays a strong focus on NRAs of EU member states, EU agencies, and North American agencies (Yesilkagit, 2011, p. 964). Considering the rise of international regulatory networks (Slaughter, 2004) and the increasing engagement of non-member state NRAs in EU decision-making bodies, the agencification literature might benefit from a broadened focus on public agencies beyond North American and European borders.
Alternative determinants of agency autonomy
Yesilkagit and van Thiel (2014) find that an agency’s task has a differentiated effect on agency autonomy. The study finds a strong relationship between task and formal agency autonomy for agencies tasked with regulation, certification, registration, research and maintenance. However, agencies tasked with certification, regulation and registration display lower levels of formal autonomy. The study also suggests that whilst agency task is indeed related to formal agency autonomy, the task variable is not the only determinant of agency autonomy: agencies’ formal autonomy (or legal type) has, on average, an even stronger effect on agencies’ financial-, HRM, and policy autonomy (p. 334). Finally, agency budget size is related to formal agency autonomy: agencies with large budgets tend to score lower on formal agency autonomy. Determinants of network effectiveness
Researchers point out that networks not only come in various shapes and sizes but that these days, the phenomenon has become quite ubiquitous (Slaughter 2004; Slaughter & Zaring, 2006; Maggetti, 2007). The omnipresence and heterogeneity of the network phenomenon are reflected in the large number of conceptual definitions that aim to capture the essence of this particular type of collective action effort. In this section, I review relevant conceptual definitions to situate the networks under study in this thesis. Furthermore, I review existing research on the determinants of network effectiveness to inform my case selection strategy by identifying relevant control variables.
Broadly, the network concept refers to group arrangements that are predominantly non-hierarchical and encompass a multitude of social actors that are connected with each other by sets of network ties (O’Toole & Meier, 2004, p. 682). In this thesis, however, I study a specific type of regulatory network that is mandated by EU agencies (like EASA), and that brings together EU regulators (for instance, EASAs civil aviation experts), with their foreign counterparts (for example, civil aviation experts representing civil aviation authorities of non-member states in ENP countries). These networks are more than structural symbols of mutual
[16]
dependencies that exist between network actors: they aggregate the regulatory capacity of the EU’s and third countries’ regulatory authorities and thereby come to function as supranational institutions of international regulatory governance (O’Toole & Meier, 2004, p. 682; Maggetti & Gilardi, 2014, p. 1295). They “share a number of common features, including usually agency / ministry membership, action through rule making, a taste for subsidiarity, enforcement through peer review, small secretariats and informal process” (Newman & Zaring, 2013). Encompassing representatives of EU- and national regulatory agencies, the networks under study here are also interorganizational, and they are formed with specific objectives in mind. Goal-directed, interorganizational networks encompass at least three organizations that collaborate as they seek to accomplish network-level outcomes, they can emergence serendipitously or may be created consciously, and they are often chosen where solving complex, collective action problems presupposes an institutionalized coordination of several organizations (Provan & Kenis, 2008, p. 3; cf. also Provan, Fish & Sydow, 2007, p. 482).
The concept of network effectiveness refers to the successful accomplishment of network-level goals and the production of outcomes that no organization could achieve when acting alone (Provan & Kenis, 2007, p. 230; Turrini et al., 2010, p. 529). The literature on determinants of network effectiveness has identified a range of network-level variables that interfere in some way with the accomplishment of such network-level goals. These variables may be grouped according to the theoretical lenses applied by different strands of literature, and they fall into different categories of network structural variables, variables related to network participants and their resources, and network governance variables.
Research on policy networks as well as the literature on organizational sociology focus on the interaction and combined effects of the network structure and network resources to account for the attainment of positive network-level outcomes. Specifically, there is a consensus amongst scholars across both research fields that information as a network resource affects network effectiveness. Research on policy networks indicates that information in the form of network participants’ expertise translates into the authority to shape policy initiatives (Newman & Zaring, 2013). Similarly, international relations research suggests that information, when pooled, channeled and disseminated through a network structure, serves to address coordination problems (Newman & Zaring, 2013). Slaughter (2004) argues that international regulatory networks succeed at advancing transboundary harmonization of regulatory standards not only because they serve to expand the knowledge bases of their members as the resource is shared through their network ties. Instead, according to Slaughter (2004), networks also filter
[17]
information for relevance and truth. While the consensus amongst researchers across these bodies of literature seems to corroborate the hypothesis that information plays a central role for network effectiveness, these theory-building efforts also share a common weakness. Even though the importance of information is undisputed, the exact causal mechanisms that link pooled information resources with positive network-level outcomes remain unspecified.
Network analytical approaches like social network analysis (cf., for instance, Wassermann & Faust, 1994) adopt a different perspective, arguing that network effects had best been explained by network structure and its impact on the cooperation of network participants. In one of the earliest ever empirical studies of conditions of network effectiveness, Provan and Milward (1995) investigate the impact of network structure and context on the effectiveness of mental health networks. A “case survey approach” (p. 4) with data collected on different units of analysis, the study provides a high-resolution picture and in-depth analysis of a single case. Study findings suggest that networks effectiveness is conditioned by network centrality, stability, availability of sufficient resources, and their exposure to external control that is direct and not fragmented.
In a conceptual study, Provan and Kenis (2008) posit that four network-level variables, namely distribution of trust, number of network participants, network goal consensus, and participants’ network-level competencies affect network effectiveness. If trust is pervasive, if network members furthermore agree on network objectives, and if participants are competent to address demands of the networked co-operation, networks are more likely to achieve their objectives. Extended network size, on the other hand, inhibits network effectiveness. The study also differentiates between three different modes of network governance that act as moderators of the relationships between above mentioned network-level variables and network effectiveness: participant governance, governance through a lead organization, and the network administrative organization (NAO).
Raab, Mannak and Cambré (2015) look at 39 Dutch mandated crime-prevention networks and conduct a QCA (Qualitative Comparative Analysis) to study the relationship of network structural characteristics and network governance mode on network effectiveness. Network structural characteristics are operationalized into network integration, resource munificence and stability. The study indicates that all three variables affect network effectiveness. Networked forms of collaboration that have lasted three years or more, that are
[18]
centrally integrated and command substantial resources tend to be more effective in accomplishing their objectives.
The literature on network effectiveness faces issues similar to those encountered by the literature on EU external governance effectiveness. There are not very many empirical studies that have taken the theoretical approaches developed by the policy network literature and international relations research to the test. Empirical research, however, is needed to sharpen assumptions on the role of information for effectiveness, for instance through an identification and disentanglement of the causal relation between information and network effectiveness in specific cases.
[19]
Theoretical framework
One of the most central variables in resource-based explanations of network effectiveness is information. Where the solution of difficult collective action problems depends on the availability of a versatile knowledge base, fed by heterogeneous experiences and information of a multitude of stakeholders, networks can outperform hierarchies and market modes of coordination (Powell, 1990). The information advantage of networks (Eberlein & Grande, 2008), however, does not simply result from the mere congregation of a large number of network actors and the aggregation of their non-material information resources. Rather, in networks, individual resource contributions of single members only reach their full potential, for instance to set in motion socialization processes that, in an external governance context, could lead to the projection of EU rules to third countries, as the network-specific structure is utilized for resource dissemination. As individual actors engage in networked forms of cooperation, they establish dense webs of network ties that enable them to communicate with the network as a whole rather than merely with a limited set of contacts within the reach of dyadic relationships (Wassermann & Faust, 1994). Therefore, networks provide a structure that spurs the transfer of institutions from one community to another because network ties exist and can facilitate a far-reaching dissemination of information and knowledge, which constitutes a prerequisite for socialization processes (Checkel, 2005).
Nevertheless, even within network structures, information, or any resource for that matter, does not simply ‘flow’ from one node to another in an automatized, self-sustained manner. Much rather, for socialization processes and rule transfer to occur within networks, certain conditions have to be fulfilled. A first prerequisite for network effectiveness relates to the knowledge resources that individual network participants ‘upload’ into the network. For instance, regulatory networks address technocratic regulatory experts, who have at their disposal expert knowledge required to respond to and find solutions to cross-boundary regulatory problems. Secondly, for networks to drive on the transfer of rules through socialization processes, certain requirements with respect to the quality of the contact between network members have to be met. Socialization processes are contingent on the exchange of information resources through sustained and deep contact between individual network members (Checkel, 2005; Slaughter, 2004). Therefore, the potential of networks to function as socialization agents depends on the behavior of single organizational participants. Individual network participants must engage actively in the network, regulatory professionals exchanging expert information and knowledge through regular and ongoing interaction.
[20]
In the context of a regulatory network, regulatory agencies that display high degrees of autonomy should be able to leverage their network ties optimally, and to respond flexibly and adequately to the demands arising from networked forms of cooperation. The concept of agency autonomy refers to the capacity of a given public agency to make any of its operational decisions freely and unrestrained by the predilections of their domestic principals (e.g. Groenleer, 2009). As such, agency autonomy, a multidimensional concept, denotes, for instance, an agency’s decision-making discretion in all staff-related matters (e.g. Groenleer, 2009; Verhoest et al., 2010). In consequence, agencies vested with high degrees of HRM autonomy have discretion to align their staffing policy with their operational needs (e.g. Groenleer, 2009; Verhoest et al., 2010). Consequently, in the context of a regulatory network, regulatory agencies that possess a high degree of HRM autonomy should be able to freely make staff-related decisions in support of their engagement in regulatory networks. For example, they should be able to freely decide to hire additional staff or to recruit staff with appropriate profiles to maximize the network structure they engage in. In contrast, agencies whose HRM autonomy is low may have to align with the preferences of their principals who might take issue with the financial resources freed up to support a specific staffing policy – potentially to the detriment of the agency’s technocratic orientation.
Aside from HRM autonomy, agency autonomy also refers to an agency’s discretion in managing its financial affairs independently. Vested with a high extent of financial autonomy, agencies can, for instance, decide freely on how to allocate the financial resources at their disposal within the organization. Therefore, highly autonomous agencies should have leeway to free up and allocate adequate financial resources to leverage the existing network structure. For example, agencies with high financial autonomy should be able to allocate funding to finance their delegates’ attendance of network meetings. If agencies do not possess a high extent of financial autonomy, they may not be able to allocate funds freely or to shift financial resources from one budget post to another (Wonka & Rittberger, 2009). Consequently, lacking funds for network events, such agencies may not be able to engage actively in the network.
Not every regulatory agency that engages in forms of cooperation with other agencies will also be vested with high degrees of decision-making discretion networked (e.g. Groenleer, 2009; Verhoest et al., 2010). Consequently, not every agency will have the capacity to control and invest its resources in a way that warrants active and effective network participation. Among the reasons for the variation in autonomy across agencies and, subsequently, their ability to leverage their network ties in a way that is conductive to socialization, is the ongoing
[21]
capacity of their domestic principals to shape agency autonomy through the implementation of control and oversight mechanisms (Wonka & Rittberger, 2009). As previously discussed in the introduction, the domestic principals of public agencies may opt to oblige agencies to fulfill extensive reporting duties (Wonka & Rittberger, 2009) including, but not limited to, agencies’ approaches to human resources and financial management. Nevertheless, the restriction of agencies’ decision-making discretion through interference by domestic principals does not occur arbitrarily. Instead, research has identified certain conditions under which the principals of public agencies are more likely to reinforce their control by decreasing the autonomy of public agencies.
Variation in the extent of agencies’ discretion to decide freely on the acquisition and investment of resources to sustain the flow of information through active engagement in a networked context may result from variation in the economic conditions that prevail in their countries of origin. For instance, in the face of a deterioration of economic conditions, governments display a tendency to react with recentralization through the reinforcement of hierarchical control by shifting back power to central authorities. Recent case studies describe how European countries have responded to ongoing economic crises with a reversal of the long-lasting trend of decentralization through the implementation of austerity measure (Bolgherini, 2014; Muro, 2015). Recentralization research is an emerging field (Muro, 2015). However, even though most existing case studies focus on the effects of recentralization on the autonomy of sub-state governments, changes in the economic environment in the form of economic crisis also affect the development of public agencies’ autonomy (Verhoest et al., 2010). As an example, as countries navigate through periods of economic crisis, attempts to stabilize the fiscal deficit may bring about public sector reforms through which risk-averse politicians curtail the financial autonomy of public agencies (Verhoest et al., 2010).
Therefore, as adverse economic conditions induce agencies’ principals to reverse decentralization tendencies in favor of tighter control mechanisms, regulatory agencies are less likely to be able to engage in regulatory networks in a way that enables them to leverage the advantages of the network structure for the exchange and processing of new information. For instance, if the principals of regulatory agencies react to deteriorating economic conditions by curtailing agencies financial autonomy, regulatory agencies engaged in regulatory networks might no longer be able to align staffing decisions with the demands that arise from their cooperation with other agencies. If there is a need to invest financial resources in the recruitment of additional staff to facilitate networked cooperation with fellow network participants, newly
[22]
imposed budgetary restrictions in response to economic pressures might prevent regulatory agencies from responding adequately. In consequence, adverse economic conditions in network participants’ countries of origin can impact negatively on the effectiveness of networked modes of EU external governance as they may decrease the extent to which the network structure can function as an opportunity platform for the exchange of expert knowledge.
Furthermore, as agencies’ domestic principals react to negative changes in their countries’ economic climate through a reversal of a previous devolution of financial autonomy to public agencies, regulatory agencies that participate in transgovernmental regulatory networks might face newly imposed budgetary restraints that detain them from leveraging the network structure for intense and direct contact with their foreign counterparts. Concretely, under adverse economic conditions, regulatory agencies might have to hold off on financial investments in the attendance of their delegates in events or meetings within the scope of the networked cooperation they engage in.
Based on these reflections, I expect that the economic conditions in the national context of regulatory agencies that engage in regulatory networks might impact on the effectiveness of networks in fostering the transfer of EU rules to ENCs. Specifically, I assume that negative changes ENCs’ economic environment and adverse economic conditions will negatively affect the extent to which the participation of ENCs’ regulatory agencies results in regulatory convergence with EU sectoral rules. Derived from this line of argument, I advance the following hypothesis:
• Hypothesis: Where EU agencies cooperate with regulatory agencies of the ENCs in regulatory networks, adverse economic conditions in the respective ENCs will negatively affect the effectiveness of networked modes of EU external governance
[23]
Methodology
In this section, I explain my choice to address my research question with a qualitative, small-N comparative case study (Toshkov, 2016) that takes a co-variational approach (Blatter & Haverland, 2014; Gerring, 2017) and employs a most similar systems design (MSSD) (Toshkov, 2016; Blatter & Haverland, 2014). The section is structured as follows: Firstly, I discuss how my research design choice relates to this study’s research objective and data availability. Secondly, tying this section back to the literature review(s) above, I explain how my case selection strategy fulfills the requirements of establishing causality with a co-variational approach and thus warrants valid causal inferences. Thirdly, I discuss the issues that this study faces with respect to the external validity of its findings and I evaluate the weight of these issues in light of the study’s research objective. Fourthly, I discuss the benefits of the chosen approach for concept validity of measurement before turning to the effects of data triangulation on random measurement errors and systematic bias. Finally, I turn to concept operationalization and the measurement process.
Research design: The case for a qualitative, small-N comparative case study with a co-variational approach and a most similar systems design (MSSD)
To establish the plausibility of a relationship between the economic conditions in ENCs and the effectiveness of networked modes of EU external governance, I conduct a qualitative, small-N, cross-sectional comparative case study of the networked collaboration of civil aviation authorities of Ukraine and Moldova with the European Aviation Safety Agency (hereinafter: EASA). The study takes a co-variational approach and employs a most similar systems design to establish causality: the selected cases display variation on the independent variable (economic conditions). At the same time, other factors that have previously been identified as causally relevant for the outcome variable (effectiveness of networked modes of EU external governance) are treated as control variables. Following the blocking conditioning strategy (Toshkov, 2016), these factors are then kept constant across the selected cases.
This approach to answering the guiding research question is informed, firstly, by the study’s research objective and its relationship to theory. With this study, I aim at establishing whether a relationship exists between the economic conditions in ENCs and the effectiveness of EU external governance, and whether the effect is positive or negative. In other words, at this point I am more concerned with the sharpening of a novel, broader theoretical idea than with a detailed, in-depth study of the specific causal effects that link my variables. The more
[24]
suitable tool for the latter research goal are large-N observational analyses (Gerring, 2009; Neumann, 2014; Toshkov, 2016). Large-N studies are able to strengthen theories by confirming the existence of strong causal mechanisms by observing regularities and patterns across a large set of cases. To do so, however, large-N studies require well-established, well-specified theories and concise hypotheses to test (Afonso & van der Zwan, 2017). In contrast, my study embarks on largely unmapped territory: we already know much about the effect of agencies’ involvement in regulatory networks on the development of their autonomy, determinants of EU external governance effectiveness have been widely researched, and recent research has shown, that EU agencies increasingly integrate foreign regulators to address issues of functional interdependence. However, the relationship between prevailing economic conditions in ENCs and the effectiveness of networked forms of EU external governance effectiveness remains largely unexplored, and even though the basic claim is theoretically not unfounded, it is new and unspecified. As such, this study has to handle the lack of a well-established body of literature. Therefore, the conditions for large-N research are not present: it does makes little sense to inquire for the ‘how’ as long as we do not know if there actually is a relationship.
In light of this study’s research objective, the approach has a clear advantage over large-N approaches in terms of its ability to establish internal validity, that is “the validity of inferences about X’s relationship to Y for the studied example” (Gerring, 2017, p. 195). To lay the groundwork of a novel theoretical idea, this study needs to produce results that show that specific economic conditions covary alongside EU external governance effectiveness. In qualitative, small-N comparative research, internal validity is enhanced because of its approach to causal comparability and data collection. I will continue to show the merits of this study for internally valid results in more detail below and discuss the blocking conditioning case selection strategy employed for this co-variational approach. At this point, I will briefly turn to the relationship of the selected research design and data availability.
Data availability was an issue for this research. With this study, I seek to establish a novel theoretical idea about the relationship between economic conditions in ENCs and networked forms of EU external governance effectiveness. However, heightened interest in networked forms of EU external governance has emerged only recently (Lavenex, 2014; Lavenex, 2015; Rimkutė & Shyrokykh, 2017; Rimkutė & Shyrokykh, forthcoming), and in contrast to, for instance, EU-third country cooperation in the telecommunications sector (e.g. Langbein & Wolczuk, 2012), cooperation in the area of aviation safety has not yet been widely researched. As this study is situated in an emergent field, it was not surprising to find that data
[25]
sets on ENCs’ convergence with the EU’s aviation acquis do not yet exist. This is not to say that assembling such data would have been impossible. However, given the time and other resource constraints associated with a master thesis, a quantitative approach in this context was less feasible. Instead, a qualitative small-N approach allows for leveraging the small scope of the study to enhance internal validity by collecting thick, qualitative data and generating thorough, in-depth knowledge of the cases under study. In contrast to large-N quantitative studies and due to small number of selected cases, this study is much less vulnerable to the risk of conceptual stretching. The small number of cases and the intuitive approach to data collection enable the development of indicators that are sensitive to the cases’ context and cross-cultural differences (Blatter & Haverland, 2012) and facilitate establishing equivalence (Landmann, 2013).
Case selection strategy
Launched in 2004, the ENP is a policy framework that manages the EU’s external relations with the following 16 countries in its eastern and southern neighborhood: Algeria, Armenia, Azerbaijan, Belarus, Egypt, Georgia, Israel, Jordan, Lebanon, Libya, Moldova, Morocco, Syria, Palestine, Tunisia, and Ukraine (Commission, 2003). To examine the relationship between prevailing economic conditions in ENCs and the effectiveness of networked modes of EU external governance, following MSSD principles, I select Ukraine and Moldova as cases from the general population of 16 states. In addition, I focus on one specific network, namely the collaboration of civil aviation experts of EASA and aviation professionals from the Civil Aviation Authority of Moldova (hereinafter: CAA) and the State Aviation Administration of Ukraine (hereinafter: SAAU).
Selecting the cases of this study, I follow the principles of the MSSD research design to satisfy the requirements for establishing causality with a co-variational approach. With a focus on the networked collaboration of CAA Moldova and SAAU Ukraine with EASA, I allow for variation on my independent variable: As will be shown in the data analysis section, economic conditions in Ukraine are much more adverse than in Moldova. Furthermore, I selected Moldova and Ukraine due to their displayed similarity on relevant control variables, which I have previously identified in the literature review. Seeking to draw causal inferences from an observation of co-variance of certain economic conditions and instances of effective networked modes of EU external governance, an MSSD-based case selection strategy enables me to satisfy the two conditions for making valid causal inferences, which result from the deterministic, counterfactual logic of causation that underlie the co-variational approach (Blatter &
