Governing by network: Counter UAS

Governing by network: Counter UAS

Author: R. Keereweer BSc

Supervised by: Dr. J. Reijling

Leiden University

Faculty of Governance and Global Affairs

MSc Crisis & Security Management

Governing by network: Counter UAS

How organizations network themselves anticipating new UAS

threats and challenges

Master Thesis Crisis & Security Management

Preface

Before you lies the thesis “Governing by Network: Counter UAS”, the basis of which is a research on how organizations network themselves anticipating new threats and challenges coming from Unmanned Aircraft Systems, perhaps better known as a drone. It has been written to fulfil the graduation requirements of the Crisis, Security and Management Master program at the University of Leiden. I was occupied in researching and writing this thesis from September 2018 to January 2019.

During the research I did not know for some time which approach I was going to use. First, I wanted to see how policymakers are dealing with technology and how legislators are anticipating possible future inventions. Since scientific studies into the future are difficult to implement, my course changed to the way the current legislation was originated. However, this perspective had little to do with my master's study, and frankly, it was also hard to go through all those legal texts. In the meantime, I had contact with someone who was part of a security network where organizations came together to discuss countermeasures for 'drones'. The question of whether this network might be interesting to do research in, received a positive response. My research question was formulated together with my supervisor, Dr. J. Reijling.

I would like to thank my supervisor for his guidance and support during this process. I also wish to thank all of the respondents, without whose cooperation I would not have been able to conduct this analysis. Everyone took extensively time to talk about their view on the network. Thank you very much! My gratitude goes especially out to the NCTV, who invited me to present the results to the C-UAS network.

You never write a thesis alone, and therefore my girlfriend deserves also a particular note of thanks: your motivation and kind words have, as always, often done me well. I hope you enjoy the reading.

Rick Keereweer

Table of Contents

Preface ... 4

Abbreviations ... 6

Abstract ... 7

Terms & Definitions ... 8

1. Introduction ... 9 1.1 Research ... 10 1.2 Relevance ... 10 1.3 Reading Guide ... 11 2. Theoretical Framework 2.1 Network Collaboration ... 13

2.2 The effectiveness of network collaboration ... 18

2.3 Criteria assessment network effectiveness ... 19

2.4 Analytical Framework ... 25

3. Research Design & Methodology 3.1 Research Design ... 28

3.2 Operationalization ... 29

3.3 Methods of Data Collection ... 30

3.4 Participants ... 31

3.5 Data Analysis ... 33

3.6 Confidentiality ... 33

3.7 Validity & Reliability ... 34

4. Findings 4.1 Setting the Scene ... 35

4.2 Measuring the Effectiveness ... 44

5. Conclusion ... 56

6. Discussion & Recommendations ... 62

7. Bibliography ... 69

Abbreviations

UAS Unmanned Aircraft System

UAV Unmanned Aerial Vehicle

NCTV The National Coordinator for Security and Counterterrorism

DGLC Defense Ground based Air Defense Command

DSI Special Intervention Unit

DJI Custodial Institutions Agency

KMar The Royal Netherlands Marechaussee

LVNL Dutch Air Traffic Control

AIVD General Intelligence Security Service

MIVD Military Intelligence Security Service

DKDB The Royal and Diplomatic Security Department

ILT Human Environment and Transport Inspectorate

ABL Unit of the Civil Aviation Authority the Netherlands

TNO The Netherlands Organisation for applied scientific research

Abstract

This research is concerned with examining the effectiveness of the Counter Unmanned Aircraft Systems (C-UAS) network in the Netherlands. The theory of Provan and Kenis (2007) and other related scholars are used as theoretical lens through which the research question will be answered. Based on the understanding of theories in relation to certain concepts, this research might arrive at remedies for potential structural defaults between the current network management model and the ideal-typical network model. These concepts are; the way the network is created, the presence of network agreements, the joining and number of participants in the network, the degree of interaction between the participants, the level of trust, goal consensus and the degree of involvement of the organizations within the network. These concepts form the process effectiveness. The concepts about the amount of practical products produced by the network and the absence of duplicate tasks are examined to conclude the result effectiveness.

The C-UAS network tries to achieve to have well-functioning operational countermeasures available in the Netherlands. However, organizations may prefer to pursue their own goals over the main goal. Decisions often stay at the administrative level and therefore the network has a hard time in producing any practical results. Central within the network is the lack of clarity about responsibility. This leads to the situation where another network model could be more efficient.

Keywords:

Drones, UAS, C-UAS, Countermeasures, Effectiveness, Public Management Networks, Inter-organizational Networks, Trust, Network agreements, Number of participants, Goal Consensus, Involvement, Output, Result effectiveness, Process Effectiveness, Collaborative Approach Counter Drones

Terms and Definitions

Every UAV is a drone, but not every drone is a UAV

There are a lot of different names for what is popularly known as a drone. This study will use the term UAS, because a UAS (Unmanned Aircraft System) is the system that includes the Unmanned Aerial Vehicle (or drone), but also the remote operator and the system with all communication links between the operator and the aerial vehicle (Osborn, 2009). Basically, a UAV is part of a UAS, and only refers to the part that is in the sky. The difference between a drone and a UAV is that many professionals believe UAVs need to have autonomous flight capabilities and drones don’t. A drone could also be an unmanned underwater vehicle for example (Rao, Goutham Gopi and Maione, 2016 and Herrick, 2017).

1.

Introduction

The use of an Unmanned Aircraft System (UAS), often referred to as drone, has increased dramatically in recent years. With the value of UAS-activity rising from $40 million in 2012 to about $1 billion in 2017, the market in the United States is illustrative for the increasing use of UASs (Finnegan, 2017). Since the early 1900s, UASs were originally being used for military operations, but in the last decade, UASs have become more sophisticated, easier to use, cheaper to buy and lighter to handle and it is therefore expected that the use of UASs by private individuals, businesses and public authorities will increase tremendously in the upcoming years (Cohn, Green, Langstaff and Roller, 2017). These developments also imply new risks that have to be mitigated and raise questionsas to what is technologically feasible, legally permissible and socially desirable. The rising capabilities and growing accessibility of UASs widens the risks of misusing UASs and for example the success of terrorist’s attacks. In August 2018, while the President of Venezuela was giving a speech, a device - what looked like a UAV - detonated nearby his stage and injured at least seven soldiers. Although the question is whether it really was a UAV that caused the explosion, it showed the potential danger.

In 2005, Eugene Miasnikov simulated already in his report Threat of Terrorism Using

Unmanned Aerial Vehicles a scenario whereby “900 grams of weapons-grade anthrax

was dropped from a drone at a height of 100 meters just upwind of a large city of 1.5 million people”. If this would occur, all inhabitants would become infected and “even with the most aggressive medical measures that can realistically be taken during an epidemic, this model estimates that approximately 123.000 people would die” (Miasnikov, 2005:7). Just only the hope that such an unthinkable and bizarre act would never occur in cities like Amsterdam or Rotterdam is of course not deterrent to terrorists. It is important that we should not allow ourselves to be distracted from the larger goal of protecting public safety simply because the thought of such act is ‘unthinkable’. If such acts are thinkable for terrorists, then they must be on top of mind for those who are responsible for public safety and security (Bartsch, Coyne and Gray, 2017). To counter these kinds of misuses of drones, it is important that organizations work together and therefore this research will focus on the safety network that has been formed or should be formed in order to deal with these new

threats. Even though the existence of safety and security networks is already studied by many scholars (Brodeur and Dupont 2008; Dupont 2004; Gill 2006; Palmer and Whelan 2006; 2014; Whelan 2012; 2014. In Whelan, 2015:537), much is still unknown about the involvement of different departments and agencies in national security and how they work together.

1.1 Research

This research will therefore focus on how different (security) organizations in the Netherlands have organized themselves in a Counter Unmanned Aircraft System network, in order to address this potential threat and come up with concerted national countermeasures. It is interesting to see how different departments and different agencies have to work together in order to maintain citizen-safety. In this broad approach we might say something about the efficiency of the cooperation within the network. Therefore, the main question in this research is:

To what extent is the present counter UAS network in the Netherlands tailored towards citizens-safety and how can potential discrepancies be addressed?

The theory of Provan and Kenis (2007) and other related scholars will be used as the theoretical lens through which the research question will be answered. This theory explains how certain concepts about culture, structure, output and management could influence the effectiveness of a network. These concepts are the subject of four sub-questions, which will be discussed in chapter two. By means of the answers to these questions the main question will be answered.

Central to this research will be to determine how all participants from different departments, in a (technical) difficult dossier like (counter) UAS, are working together. What are the structural characteristics of the network? Could we describe this as an effective approach? Based on this understanding we might be able to formulate possible remedies for potential structural defaults.

1.2 Relevance

As described before, the concerns for misusing UASs for terrorist attacks are present, even in The Netherlands. In April 2017 for example, the National Coordinator for

Counterterrorism and Security (NCTV) wrote in Threat Assessment Terrorism No. 44 that they are worried that jihadists will use UASs with explosives to carry out an attack. “The innovative capacity of IS is a cause for concern,” wrote NCTV chairman Dick Schoof. He refers to the fact that groups in Syria and Iraq are already using armed UASs on the battlefield. Because this method is being glorified on media channels of IS, it can also be inspiring for jihadists in the West (NCTV, 2017). It is not only from a counter-terrorism perspective that this research is relevant, but also to safety in general. As an example of this, we do not have to go far back in time. Quite recently, in December 2018, air traffic at London's Gatwick Airport has been disrupted for days due to the presence of UASs in the airspace above one of the busiest airports in the UK. The authorities believe that this was done deliberately (BBC, 2018a). Therefore, it is important to have a clear picture about how organizations in the Netherlands are coping with these relatively new threats and challenges and examine to what extent ‘we’ are prepared for such an incident at one of our airports.

Secondly, the fact UASs are relatively new in the society and not familiar for everybody, (local) authorities do not always know how to deal with this new phenomenon. By creating more awareness on this topic, the importance of counter UAS measures is easier to accept and this could benefit the progress. This can be seen as the societal relevance.

As already said, technology has progressed tremendously in recent years and some inventions are being made of which it was said years ago that this would never happen. Actually, nowadays everything is possible and that makes it difficult for policymakers and law enforcers to respond to. The question of how these dilemmas relate exactly to each other and how security services can work together in the most efficient way makes this research scientifically relevant. Not alone for UASs, but also for challenges that come with new technologies that are inter-departmental.

1.3 Reading guide

This chapter forms the introduction to the research. In the next chapter the theory of Provan and Kenis (2007) and others, is discussed to help to understand how a network could be managed and how a network should perform in order to be effective. On the

basis of four sub-questions, the analytical framework shows how the theory is used in order to answer the main research question. The third chapter explains the methods in which the data is collected, how the data is analyzed and how the results are used to arrive at the point where we can assess the effectiveness of the ‘C-UAS-network’. These finding are discussed in chapter four. Both the possibilities and dangers of the (use of) UASs are elaborated on in chapter four. In addition to this, consideration is given to the fact that there is no easy solution for combating these specific threats because of the different types of UASs. In this chapter is also consideration given to the various intentions of which a pilot can use a UAS, because different use may require different action.

In the second part of the findings, the C-UAS network has been evaluated in detail in order to see how the Dutch security organizations are trying to have well-functioning operational countermeasures. On the basis of various criteria, it is examined whether this is done effectively or not.

On the basis of the most important results from the previous chapters, chapter five gives a conclusion about the effectiveness of the functioning of the network. In the last chapter the focus is on the discussion and possible recommendations.

2. Theoretical framework

This chapter discusses the theoretical background of the core concepts within this research: the definition of networks, the need for networks and the effectiveness of networks.

2.1 Networks

2.1.1 Definition of networks

Provan and Kenis (2007:231) define networks as "three or more autonomous organizations that work together to achieve not only their own goal, but also a collective goal". This definition will be used in this research and is important for understanding and explaining the operation of the network.

2.1.2 Emergence of networks

In today’s society, public organizations must deal with complex social problems and these problems go beyond the boundaries of one organization and “are characterized by a high degree of wickedness” (Radford, 1977; Mason and Mitroff, 1981. In Koppenjan and Klijn (2004:1). According to these scholars, there are three uncertainties surrounding complex social problems in a network setting: substantive

uncertainty, strategic uncertainty and institutional uncertainty. Substantive uncertainty “concerns the availability of information” (2004:6) and the way

organizations could interpret this available information differently. Strategic

uncertainty encompasses the uncertainty resulting from different perceptions, goals

and interests of the organizations involved in a complex problem. The unique perceptions of organizations on these problems can lead to different strategies. Because of this mechanism, it is uncertain what kind of strategy an organization will choose. Institutional uncertainty is the uncertainty that arises from the different backgrounds of the organizations. Complex and wicked problems are often inter-departmental and every organization will work from their own point of view. These differences in background could cause clashes between deviant backgrounds. When a new wicked problem needs to be tackled, there will be a high degree of uncertainty on how the interaction will develop (Koppenjan and Klijn, 2004:7).

Public organizations can no longer deal with these complex problems on their own, and they have to collaborate with other actors and are forced to operate in a network (Geuijen, 2011). Goldsmith & Eggers (2004) describe homeland security in America as an example for such a network. If organizations involved in homeland security, like FBI or CIA, work on their own, they will probably not act effectively. These organizations need the help of law enforcement networks that crosses agencies and levels of government to actually provide ‘homeland security’ (Whelan, 2015). That is why networks have become a dominant form of management and organization, in contrast to traditional forms such as hierarchy and markets (Geuijen, 2011; Powell, 1990).

2.1.3 The network perspective

The three uncertainties that Koppenjan and Klijn (2004) talk about make it very difficult for organizations to act in the event of complex problems, which come with a great amount of pressure to do something about it (Koppenjan and Klijn, 2004). The interdependencies between organizations make it impossible for a public organization to tackle a problem alone. This leads to cooperation with other actors who have an interest in the problem. In order to deal with these uncertainties, Koppenjan and Klijn (2004) have developed the network approach, in which the mutual dependency of actors to achieve their goals on wicked problems is central. The main question in this approach is “how joint action can be achieved” (Koppenjan and Klijn, 2004:10).

2.1.4 Characteristics of networks

Networks have different characteristics. First the general characteristics are mentioned and later more characteristics are described to assess the effectiveness of the functioning of a network.

Networks are basically non-hierarchical and no organization is therefore in a dominant position to coordinate the network cooperation. The organizations are equal to each other (Provan and Kenis, 2007). The difference with individual organizations is that networks are managed without hierarchy or ownership. As a result, participants of a network automatically have a limited formal responsibility with regard to the goals of the network collaboration. Adapting to the rules and procedures of network collaboration is therefore voluntary (Provan and Kenis, 2007). Decisions are usually

made on the basis of consensus (Agranoff, 2006). In a network, there is a common goal according to Provan and Kenis (2007), such as solving a social problem or developing a new product (Geuijen, 2011). Otherwise there is no need to work together (Provan and Kenis, 2007).

2.1.5 Network models

Based on the degree of interaction and coordination between the network partners and if a network is managed externally or internally, Provan and Kenis (2007) distinguish three types of networks: the self-governed network (also called the shared-governance

network), the lead organization network and the network administrative organization (NAO). The most common form is the self-managing network where the participants

of the network itself govern the network. All participants are on an equal level regarding the control of the network; there is a form of symmetry. Decisions are taken collectively; the participants within the network are dependent on each other’s involvement and commitment and are responsible for both internal and external relations.

Due to the inefficiency of the decision-making process that shared governance could entail, there is often a call for more centralization. In the most far-reaching form, this means that there is a lead organization, i.e. one participant within the network that coordinates important decisions and all network activities. The result of this lead

organization network is a network with a high degree of centralization and an

asymmetrical distribution of power. The lead organization is responsible for the administration and facilitates the other participants of the network so that the intended goals can be achieved. It is very likely that these intended goals will have a major overlap with the goals of the lead organization (Provan and Kenis, 2007).

The final form is the network administrative organization (NAO) in which an independent external network advisor plays a key role in coordinating and maintaining network collaboration. This organization has the sole purpose of managing the network cooperation. A NAO can be either an individual or an organization. In this type of network all important activities and decisions are coordinated by the NAO and is therefore a centralized model (Provan and Kenis,

2007). In the next sections, the effectiveness of the different models will be discussed. The models are shown in the figure 1.

2.1.6 Network competencies

Within these three network models, all organizations work together and pursue some common goal in order to achieve this. Because when they do not, they will not achieve the same result. When the task of the network makes the organizations dependent on each other, it requires specific task-specific competencies. The extent to which organizations need network competences to enable a network to function effectively depends on the tasks. The higher this need is, the more effective a central form of governance will be (Provan and Kenis, 2007).

Fig. 2. Network models (Provan and Kenis, 2007)

This theory will be used to form an answer to the first sub-question:

1. What type of network has been created by organizations’ in the Netherlands involved in counter-UAS measures, according to the characteristics of the network models of Provan and Kenis (2007)?

2.1.7 Outcome type of network on effectiveness

Provan and Kenis (2007) assume that whether or not there is consensus on goals and necessary network competences may be present or not, the trust between network partners and the extent of network collaboration are important for the effectiveness of the management structure. Provan and Kenis do not claim that these are the only four criteria to assess the effectiveness of network collaboration. However, they argue that these four criteria are very important for the assessment of the effectiveness of the steering form of network collaboration (Provan and Kenis, 2007). The following table lists the four criteria and how they relate to the various forms of governance. Other criteria that also have an influence are discussed hereinafter.

Fig. 3.Provan & Kenis (2007: 237)

It is interesting to see if the predictors indicate that another network model could be more effective. In the last part of the findings, the last sub-question is about these predictors to see if the C-UAS network model deviates from the ideal-typical characteristics. The reason to address this question at the end of the research is because first the theories of the different predictors have to be discussed.

4. Which theoretical and practical explanations can be found for the possible deviations from the ideal-typical network models outlined by Provan and Kenis (2007)?

2.2 The effectiveness of network collaboration

With the knowledge on the different network models, the effectiveness of such a network will be assessed. First, it is explained what definition of effectiveness (in relation to networks) is used and how the output is measured. Afterwards, the characteristics of network collaboration will be examined to assess the effectiveness.

2.2.1 Definition of effectiveness

Provan and Kenis (2007:230) define the effectiveness of a network as "achieving positive outcomes that could not normally be achieved by individual organizations or participants in a network that operate separately from each other". This definition will also be used in this research, especially since this definition says something about the cooperation within a network and also about the outcomes. However, because Provan and Kenis (2007) concentrate mainly on the effectiveness of the structure of a network, articles of other researchers will also be used to examine the effectiveness of the collaboration.

2.2.2 Assessment of the output

According to Provan and Kenis (2007) effectiveness of a network depends on a positive outcome. However, it is very difficult to demonstrate the relationship that effective network collaboration provides to the outcome for various reasons. First, the cause and effect relationships are not clear and knowable, as is often the case in the public sector (Noordegraaf, 2008). Secondly, Provan and Milward (2001) claim that a network has different types of stakeholders with different ideas about what goals the network partnership strives for and when it is successful. This also has to do with the wicked nature of problems in governance networks: what is a good result for one participant does not have to be a good result for the other. This is a battle of values and determining the outcome is therefore a normative choice (Noordegraaf, 2008). Provan and Kenis (2007) point out the differences between effectiveness (as a measure of the quality of outputs) and efficiency (as a measure of the quantity of outputs). As it is difficult to determine when a better outcome is achieved through effective network collaboration, Provan and Kenis (2007) state that measuring the

output of network collaboration is the first step in determining how the collaboration

is effective. Since it is also often difficult to measure the output in the public sector (Noordegraaf, 2008), a researcher must measure the process with the assumption that

a better process can also lead to a better output (Provan and Kenis, 2007). In this study, therefore, the first step will be to measure the process effectiveness and second to assess the output of network collaboration.

2.3 Criteria assessment network effectiveness

This part discusses the criteria of Provan & Milward (2001) and Provan & Kenis (2007) that are used in this study to measure the effectiveness of network collaboration. The focus will be on the structure, culture, output and management. The focus on structure and culture deals with the process effectiveness of network collaboration. The ‘output’ focus is about the effectiveness of the practical results of the network collaboration and will be referred to as result effectiveness. The last focus on management is about the choice for a certain network model.

In this part and the next the process effectiveness will be discussed to form a basis for answering the third sub-question:

2.3.1 Structural characteristics

This section deals with structural characteristics of network collaboration, such as the way in which the network cooperation originated, the network agreements that have been made, the entry of new partners and the number of participants.

2.3.1a The way the network is created

When a network is bottom-up originated or organized, the network participants themselves recognize the nature and seriousness of the problem and determine the approach themselves. According to Terpstra (2004), this method can ensure that the participants see the network cooperation as legitimate.

2.3.1b Existing network agreements

Most safety networks are based both on formal and informal ways of cooperation. This research focuses on such a safety network (Terpstra, 2004). The agreements between the partners are often laid down in a document.

2. To what extent does this network meet the structural and cultural characteristics of several scholars with respect to its process effectiveness?

(Local) safety networks are mainly based on mutual trust, informal agreements and knowing each other personally (Terpstra, 2004).

The informal basis of this network can make the collaboration within the network vulnerable. A risk of the extent of this informality is that it becomes unclear what the role and tasks of the various partners are (Terpstra, 2004). Informal agreements offer flexibility on the one hand, which can contribute positively to this criterion for the effectiveness of network collaboration. But on the other hand, it can also lead to ambiguities, which can negatively affect the effectiveness of network collaboration. It is case-dependent to find the right balance between formal and informal agreements, since they both know the advantages and disadvantages that can determine the effectiveness of network collaboration (Terpstra, 2004).

2.3.1c Joining and number of participants

When a network recently has been formed, it is necessary to attract and retain enough members to ‘survive’ (Provan and Milward, 2001). When a network has established its existence, this becomes less important. A network can expand with organizations on the sides that are informally connected to the network. An effective network has a core group of organizations responsible for the most important services (Provan and Milward, 2001). But the entry of new partners can also cause a large number of participants, which makes it more difficult to coordinate the network cooperation (Provan and Kenis, 2007). Due to the cross-sector nature of safety problems, organizations are often involved with several departments or services in tackling security threats.

Provan and Kenis (2007) raise the fundamental persistent challenge of network control that the behavior of different network members must be coordinated. The more actors there are in network collaboration, the more relationships need to be coordinated. According to Provan and Kenis (2007), a structural solution to this problem is to transform the network collaboration into a leader organization network or a NAO. It will then become easier to facilitate larger groups, because it is not necessary to involve all organizations in the decision-making process of network collaboration. Direct interaction between all members of the network collaboration becomes less necessary.

2.3.2 Cultural characteristics

This section contains various criteria about the culture within the network to assess the process effectiveness of network collaboration in a different way. The criteria discussed in this part are: the frequency of interaction, trust, goal consensus and the

involvement of network participants in the goals of network collaboration. 2.3.2a Frequency of interactions

Klijn, Edelenbos and Steijn (2010) call the frequency of interactions a positive process outcome of network collaboration. According to these authors, a higher frequency of contact can mean that the respondents know each other better, trust each other more and are more involved in network collaboration. The concepts of trust and involvement will be discussed further in the following paragraphs.

2.3.2b Trust

Many authors in the literature mention trust as an important criterion for working well and effectively together in a network (Provan and Kenis, 2007, Koppenjan and Klijn, 2004, Nielsen, 2007). By trust Provan and Kenis (2007) mean “the willingness to accept vulnerability, based on positive expectations about the intentions and behavior of another” (McEvily, Perrone and Zaheer, 2003, In: Provan and Kenis, 2007:237). According to Klijn, Edelenbos and Steijn (2010), trust also means that the actors keep each other's interests into account. According to these authors, trust lowers transaction costs because, for example, less expensive contracts and laws are required. It also ensures that an organization is prepared to invest (more) knowledge, money and other resources. In this way steady relationships can arise. Trust also encourages the exchange of information. Nielsen (2007) adds that trust creates more flexibility. Zaheer, McEvily and Perrone (1998) make a distinction between interpersonal trust and inter-organizational trust. Inter-organizational trust is about trust between organizations, so that, for example, partners fulfill promises. Interpersonal trust is about the honesty, predictability and reliability of the contact person of an organization. There is a direct link between inter-organizational trust and performance of a network (Zaheer, McEvily and Perrone, 1998). The transaction costs can be lower, because the organizations offer each other more space when conflicts occur. In situations where there is little trust, mistrust can lead to conflicts and to ineffective

processes of information exchange. A disadvantage of only high interpersonal trust in network collaboration is that much depends on the personal contact between certain contact persons. When one of these people gets another job, this confidence disappears. This may be an indication of the ineffectiveness of network collaboration according to Terpstra (2004). When there is inter-organizational trust between the organizations, the trust is less personal and more guaranteed when a contact person falls away.

The exchange of confidential information is an important factor when it comes to trust (Noordegraaf, 2004; Aden 2018). There is often a reluctance to share intelligence between a secret service agency and organizations (Aden, 2018). The term secrecy can be defined as a “piece of information that is intentionally withheld by one or more social actor(s) from one or more other social actor(s)” (Scheppele 1988:12 in Aden 2018:982). The organizations that do share secret intelligence are connected by trust. However, the withholding of certain information generally has a negative influence on the perception of how honest and open the collaboration is and how much information is exchanged (Nielsen, 2007).

2.3.2c Goal consensus

Provan and Kenis (2007) indicate that objectives and goal consensus have been extensively debated in the organizational sciences in recent years. The outcome of this discussion is that a certain degree of goal consensus ensures the better functioning of organizations and inter-organizational links. Terpstra (2001) also refers in his article to the consensus on the goals as a criterion for effective network collaboration. In addition, Koppenjan and Klijn (2004) make the distinction between consensus about the problem and about the solution. One does not have to assume the other. A high consensus on the problem does not automatically mean that there is also a high consensus among the participants about the solution and vice versa.

The research by Provan and Kenis (2007) shows that the degree of consensus about the goals that are required depends on the form of control. Self-regulating networks are most effective when the target consensus is high. This high degree of goal consensus ensures few conflicts and everyone knows how to contribute to the goals. If

the target consensus is relatively low, Provan and Kenis consider the leader organization network and the NAO as more suitable steering forms.

2.3.2d Involvement

Provan and Milward (2001) mention the involvement of participants in network collaboration as a criterion for measuring the effectiveness of network collaboration. Provan and Kenis (2007) add that, in particular, self-managing networks require a high level of involvement from all or almost all organizations. The network members are responsible for managing internal relations and actions, but also for managing external relations.

If the participants are judged on their own effectiveness and if this also has a big influence, for example in the case of a subsidy, then the participants will be inclined to maximize their own effectiveness. This can be at the expense of the effectiveness of network collaboration in its totality. They therefore show strategic behavior. This may be a brake on the involvement of participants, while this involvement is necessary to tackle complex problems (Geuijen, 2011).

2.3.3 Output of network collaboration

The previous sections dealt with the structural and cultural characteristics in order to be able to say something about the process effectiveness. This section describes the ‘output’ of the network and mentions the fundamentals to say something about the effectiveness of the practical results that results from the network collaboration. This is referred to as result effectiveness. These criteria are used to give an answer to the third sub-question:

2.3.3a Practical products

One way in which effectiveness can be assessed at the network level is by looking at the scope of practical products provided by network collaboration. In this research, practical products are for example manuals, procedures, purchasing systems, etc. When only a small number of products are provided, clients are forced to go outside the formal network in order to fully fulfill their needs (Provan and Milward, 2001).

2.3.3b Absence of duplicate tasks

Provan and Milward (2001) mention another way to evaluate the effectiveness of network collaboration by measuring the absence of duplicate tasks. This happens when too many organizations and programs are involved in the provision of a particular service. This can lead to a confusing collection of services, with certain efforts often being delivered twice. Failure to perform duplicate tasks can therefore be regarded as a criterion for the effectiveness of network collaboration.

3. To what extent does this network meet the characteristics of Provan and Milward (2001) with respect to its result effectiveness?

2.4 Analytical framework

The criteria regarding the effectiveness of network collaboration as discussed above, are summarized in the tables below. These tables will be used as guidance to see how the C-UAS network has been emerged, how it operates and what the output is. With this knowledge it is possible to say something about the effectiveness of the chosen management model.

2.4.1 Sub-question 1 on type of network

The first sub-question will be answered by focusing on the different network models provided by Provan and Kenis (2007). This will be a more general view at the C-UAS network, to see what kind of network model is chosen. This knowledge will be used to form an answer to the first sub-question.

2.4.2 Sub-question 2 on structural and cultural characteristics

Since nothing can be said about the effectiveness of the network on the basis of the network model alone, other criteria are also important. These criteria are divided into process and result effectiveness. This section is about process effectiveness and is divided into structural and cultural characteristics of the network.

Structural Characteristics

Assessment criterion Effective when Ineffective when

The way the network is created

The network cooperation has emerged from bottom up, initiated by one of the participants.

The network cooperation is imposed from above, created by mandate

Existing network agreements

There are sufficient formal and informal agreements in network collaboration

There are not enough formal and informal agreements in network collaboration

Joining participants Sufficient new relevant partners can join, while stability is guaranteed by a core of organizations

Not enough new relevant partners can join and / or stability is not

guaranteed by a core organization

Number of participants The number of participants and the management form of the network

collaboration match

The number of participants and the management form of the network collaboration do not match

Cultural Characteristics

Assessment criterion Effective when Ineffective when

Frequency contact The contact between the

participants is frequent

The contact between the participants isn’t

frequent

Trust There is mutual trust

between participants

There is mistrust between participants

Goal consensus There is consensus among

the participants about the goals to be achieved

There is no consensus between participants about the goals to be achieved

Involvement The network participants

are involved in achieving the goals of network collaboration

There is a division in the extent to which network participants are involved in achieving the goals of network collaboration

2.4.2 Sub-question 3 on out characteristics

In addition to the criteria that make process effectiveness measurable, any potential results are also examined. On this basis the result-effectiveness can be determined.

Output Characteristics

Assessment criterion Effective when Ineffective when

Practical products

provided by the network

The network produces many products

The network produces few products

Absence of duplicate tasks

Tasks are not duplicated by the different network members

Tasks are duplicated by the different network members

2.4.3 Sub-question 4 on the managerial characteristics

After assessing the effectiveness on the basis of the different characteristics, the choice for a specific network model could also predict if the network is effective or not. By monitoring the predictors trust, number of participants, goal consensus and the need for network competencies used by Provan and Kenis (2007), a qualification may be given about the effectively of the chosen network model.

Managerial Characteristics

Assessment criterion Effective when Ineffective when

Effectiveness network-model

The self-managing

network is effective if trust is high, the number of participants low, the goal consensus high and the need for network competencies low

The self-managing network is not effective if trust is low, the

number of participants is high, the target

consensus is low and the need for network

3.

Research Design and Methodology

In the previous chapter the theoretical and analytical framework of this research were discussed. With this knowledge, this chapter will focus on the research design and the methods of data collection. The validity and reliability of this research are also discussed.

3.1 Research Design

The research design of this report is empirical and exploratory. Explorative research intends, as the name implies, to explore the research problem. In this research a theory will be used and applied to the Counter-UAS network in the Netherlands. This network is formed by multiple organizations in order to counter the threats of a UAS. This research does not intend to provide a conclusive solution to the problem, but tries to better understand how different organizations are working together to come up with these countermeasures.

This design is a useful approach in order to gather background information on a difficult topic and is also flexible because it can address different kinds of research questions (i.e. what, why and how). At last it provides an opportunity to explain and clarify existing concepts (Saunders, Lewis & Thornhill, 2012). The rationale behind this research is to create more knowledge in order to understand how the safety network operates. The intent is to interpret the way the network operates in light of the theories.

This research contains also an interpretive approach in order to understand the behaviour of the C-UAS network in providing countermeasures. The interpretive understanding, also referred to as ‘verstehen’ by Max Weber, is a German term that looks for meanings and motives behind people’s actions (Whitley, 1984). In public administration and other social science research there are roughly two ways of scientific knowledge gathering. These are referred to as positivistic versus interpretative, or as quantitative versus qualitative (Haverland & Yanow, 2010 in Korsten, 2011:8). The positivist approach assumes that reality can be observed independently of the observer and that it is objectively known. As a result, a

researcher can convincingly demonstrate that one factor (independent variable) causes another (dependent variable). There must be a causal and effect relation. In the case of the assessment of the effectiveness of the C-UAS network, it is hard to come up with independent variables that directly cause other variables. Therefore, this research will have an interpretative approach. In interpretative research, reasons for a phenomenon are presented, and not the causes (Haverland & Yanow, 2010). Interpretive research does not begin with hypotheses or with variables but prefers to work with 'expectations' or 'hunches' (Haverland & Yanow, 2010:93 in Korsten, 2011:9). In this research the expectations are that certain indicators influence the effectiveness of the network. The goal of this research is therefore to understand an empirical case by applying theory.

The choice for using qualitative research is due to the fact that qualitative research is a good way to perform detailed research. With this form of research, a better understanding can be given of complex situations and can contribute by understanding decisions or how organizations interpret these decisions. As last, it could also reveal possible tensions between organizations. This kind of information is difficult to gather from a quantitative research (Bryman, 2012).

3.2 Operationalization

The ‘translation of abstract concepts (…) into less abstract concepts that can be

detected, classified, and measured in the empirical world” (Toshkov, 2016: 100) is

the definition that is used in this research for the analytical framework in chapter two. The sub-questions as described in the theoretical framework are used to formulate an answer to the main research question. In order to examine to what extent the network theory can help to explain how the network operates and if this can be seen as effective, it is important to first have a clear picture of how the network operates in general. Which organizations are part of this network and what are their talking points? This overview is gathered after an exploratory conversation/interview with one of the members of the network. Afterwards, a (semi) structured interview has been drawn up on the basis of the information provided.

3.3 Methods of data collection

The data provided in this research is collected through interviews and document analysis, of which the interviews form the main data source. The interviews are ‘in-depth’ and semi-structured. This type of interview is much less rigid than structured. The goal of a semi-structured interview is to explore a topic more openly and to allow interviewees to express their opinions and ideas in their own words. Depending on the answers given by the interviewee, there is room for follow-up questions, questions on another topic or to come back at a certain subject (Bryman, 2012). The big advantage of this freedom during the interview is that the interview can sometimes take surprising turns. Because the interviews are partly structured, a list with talking points was made in advance, but these questions were not leading the conversation. First of all, the interviewees were given the opportunity to introduce themselves and explain what role their organization has within the network. Based on the criteria used in this research to determine effectiveness, questions were asked about - for example - trust in each other, or about the presence of formal agreements. By asking the same questions to each respondent, something can be said about the organization and effectiveness of the network. The interviewees did always have room to expatiate themselves to certain subjects. This led to interesting insights that were unknown in advance about for example, the way in which the network deals with issues regarding the responsibility for airspace. The complete topic list can be found in Appendix A.

Before the start of each interview, the interviewer will introduce himself, explain the purpose and length of the interview and explain how the data will be used.

Besides the semi-structured interviews, a literature study and document studies are conducted. The literature study will be conducted on the basis of an analysis of scientific literature, professional literature, earlier research, existing regulation and media messages with regard to the use of UASs. Due to the fact that UASs have received much more attention in recent years, many reports have been made. However, these are mostly aimed at regulations, at UAS in general or on privacy concerns. Not much has been written about countermeasures yet, perhaps due to the fact that a lot of information is seen as confidential. One of the documents on UASs is an exploratory research into unmanned aircraft. On behalf of the government, the Research and Documentation Center (WODC) conducted an exploratory study into unmanned aerial vehicles (UAVs) in 2015. General questions were asked about the

anticipated possibilities and threats of ‘drones’. This research formed a good basis to understand the status quo of UASs in the period until 2015 (WODC, 2015). Other documents are legislation and regulations to gain insight into the complexity surrounding the use of UAS. In addition, many (international) news pages have been consulted to see how other countries are dealing with the challenge surrounding the emergence of countermeasures.

3.4 Participants

Because the C-UAS is an internal network within the NCTV, it is difficult in advance to get in touch with the various organizations. Hence the snowball sampling as a non-probability sampling method is used, whereby after each interview with an expert the interviewee was asked if he or she knew another expert who is willing to do an interview (Bryman, 2012). In this research a variation of an elite interview will be used to gain the information needed. In an elite interview, a special type of respondents is involved in the research. These persons are influential, prominent and knowledgeable in an organization. Zuckerman (1972) uses the term ‘ultra elites’ to describe individuals are already considered elite because of their title (i.e. senator or president). McDowell (1998) takes it broader and describes individuals working at different levels and specify them as elite because they are “highly skilled, professionally competent, and class-specific” (McDowell, 1998: 2135). When referring to this research, elites are not the appropriate term as to those who were interviewed. Therefore, the interviewees are seen as ‘experts’. This is a variation on the elite interview and involves interviewing people who are well informed about certain issues and / or are well socialized in certain locations or social situations. Ultimately, ten interviews were conducted with experts from the following organizations1:

− Ministry of Infrastructure and Water Management (hereinafter; I&W)

In the first instance, the focus of this study was to investigate the regulation of UASs. In total there have been three interviews with this Ministry. Although the subject was viewed from a different angle at a certain point in time, these conversations gave a clear picture of the complexity of the various interests.

− Dutch Air Traffic Control (LVNL)

In order to enter into dialogue with other stakeholders in addition to the safety network, the LVNL has been interviewed.

− NCTV

Since the working group - and therefore the network - was set up from an initiative of the NCTV, this organization was also interviewed.

− The National Police

The interview with the National Police at the Department of Technological Development and Expertise was very useful in view of all the examples discussed and to hear the vision of 'the Police' within the network.

− DSI (Special Interventions Unit)

As member of the network, the Special Interventions Unit was also asked to talk about the network. It was interesting to see how an organization with both Defense and Police personnel deals with this subject.

− Ministry of Defense

The conversation with the responsible department of the Ministry of Defense was very interesting because of the details of the conversation. It was not just about the criteria that were present or not, but a lot of additional information was given. This was very welcome in order to get a complete picture about UAS.

− Ground based Air Defense Command (hereinafter DGLC)

On the DGLC I was invited to talk with the 'coreteam' that Defense has developed. During the interview, a lot of details were given about how they work with UASs and what kind of counter systems there are on the market.

− Royal Netherlands Marechaussee (hereinafter KMar)

conversation was very open and honest. As a result, almost everything could be asked, which therefore benefited the research.

3.5 Data analysis

At the end of each interview, the conversations were processed as soon as possible. When the interviewee agreed with a recording of the interview, the conversation was transcribed afterwards. Only in one case did the interviewee object, so a report was made on the basis of notes. In two other cases was decided in advance not to use recordings and to make notes instead. Both the transcript and the reports were sent back to the interviewees for approval. In this way it is prevented that certain answers have been misinterpreted or need to be supplemented with more information (Davidson, 2009). After all respondents had agreed, added or modified the transcripts or reports, the results were coded. This coding process started with marking certain text fragments in each printed interview, according to a coding scheme. This coding-scheme corresponded with the different criteria used in the theoretical framework. In this way, the answers with the same codes can be merged into one category. This gives a clear overview of the respondents who gave the same answer, so that the results can be compared more easily with each other and with the theory to draw a convincing conclusion. This coding process was done by hand.

3.6 Confidentiality

In this study respondents were asked for opinions about other organizations, trust, feeling etc. These factors are important and vital to be able to say something later about the effectiveness of the entire network. However, these are also sensitive statements and therefore, efforts will be made to ensure the confidentiality of each respondent. But assuring complete confidentiality means that everything the respondents said cannot be repeated or used in this research. That makes it difficult for the researcher to use that information. A solution for this problem is to ensure the respondents no identifiable information will be disclosed and try to protect the identity of the respondents (Wiles, Crow, Heath & Charles, 2008). In the case of the C-UAS network is it sometimes inevitable that some information is reducible to the organization in dispute. Therefore, in the chapter with the analyses, more general statements are made in order to ensure - to the best of my ability - that no respondents

are identified at about statements they have made. When there are doubts about certain findings, the transcripts can be requested.

3.7 Validity and reliability

In order to get reliable results, both internal and external validity must be guaranteed (Bryman, 2012). External validity focuses on the generalizability of the research. Explorative studies generate often information that stands on interpretation. Therefore, one could say that this type of information is subject to bias (Bryman, 2012). These kinds of studies also make use of a number of interviewees or data that may not represent the population. So therefore this kind of research cannot be generalized to other problems (Saunders, Lewis & Thornhill, 2012).

Internal validity means the extent in which the observations suit the theoretical ideas that are researched. In other words; to which a research actually measures what it wants to measure (Bryman, 2012) During this research, data will be collected through (semi) structured interviews to find out the experiences of those involved in counter UAS measures. To ensure the internal validity in this research, the practice should match the theoretical conclusions. Unfortunately, it is not entirely possible to test all of the conclusions in practice, since the working group is private. With access to the network and observations from close by, the research will score higher on internal validity. Besides, it has been agreed with the working group to present the results to them, so the network can do something with the findings. To ensure the reliability of this research, all gathered data and interviews would be made accessible in order to see how this research occurred and to perform the same research again.

4.1

Findings: setting the scene

The findings are presented in two parts. The first part of the findings provides some background information on the use of UASs and the need for countermeasures. This information gives a better understanding of the complexity of the problems the network has to deal with and what the network entails. The second part of this chapter will examine the effectiveness of the network on the basis of various criteria mentioned in chapter two. It is worth emphasizing again at this point, that, in the interest of guaranteeing confidentiality, it was consciously chosen to show the respondents' quotes in italics, without mentioning the respondent in question. In other cases, when it concerns quotes from studies, reports or media etc., this will be provided with “parentheses”.

The first part begins with the different types of UASs and an overview of what the opportunities of UASs contain. The use of UASs is increasing, therefore it is useful to mention for wat purposes UASs can be used. This knowledge is necessary to see how big the market for UAS is and therefore the size of the potential dangers. These dangers are elaborated in the next section to see how UASs could be used in the wrong way. By identifying threats, the necessity of countermeasures becomes clear. The recent incidents at London Gatwick Airport are used to illustrate the importance of these measures and explain why it is complicated to have a good system that keeps UASs off airports. The last section examines the need for a combined approach of multiple organizations; the network approach and gives an answer to the kind of network model is chosen.

4.1.1 Types of UASs

There are many types of UASs, but in general you can make a distinction between fourdifferent UASs:

• COTS (Commercial of the Shelve)

This type of UAS comes directly from the shop and is unmodified.

This type of UAS is specifically developed for a specific purpose. Most of the time this is a military purpose.

• MOD (Modified)

This is a type of UAS that is modified after being purchased. Someone purchases a UAS in the shop and then modifies it to something he or she wants.

• DIY (Do It Yourself)

The last type is the DIY UAS. This type is fully equipped with different parts of other UAS or devices.

The tricky thing about the last two types is that it is unknown what you are dealing with when you see it flying and this makes it difficult for setting up countermeasures. 4.1.2 Opportunities

Opportunities can be found in many sectors. In the public sector, for instance, many ministries are using UASs. Some tasks that are performed with the help of UASs are crime prevention, reconstructions of crime scenes, emergency aid, dyke inspections, anti-fraud, border control and environmental, agricultural controls etc. In the private sector there are mainly opportunities for citizens, businesses and science. At the moment, the most important opportunities lie mainly in the use of UASs in combination with cameras. There are also sufficient opportunities for UASs that are equipped with other sensors, such as heat sensors for the detection of cannabis plantations, drones with medicines, drones for rescue operations in the sea and drones with spray liquid for agricultural projects (WODC, 2015).

4.1.3 Challenges (and threats)

As is often the case with inventions and technologies, besides the advantages of UASs, there are also disadvantages and threats when using them. This is the case for example in aviation. Where pilots always feared a collision with a bird, a danger has been added since several years; the UAS. "A pilot encountered a UAS at a height of 5 kilometers," says Joost van Doesburg of the Dutch Road Flyers Association. "If you hit such a hard thing at 600 kilometers per hour, for example at your cockpit window, you really have a big problem” (Hoogerwaard, 2018). In order to create a clear picture

of the different threats, here are some general categories of threats mentioned in the interview with DGLC:

• Reckless behavior

From the overviews that Human Environment and Transport Inspectorate (ILT) / ABL periodically provides of UASs related to ILT or police reported incidents, it appears among other things that reported incidents with third parties mainly have a relation with recreational used UASs. A logical explanation may be that this group is insufficiently aware of the risks. For example, one does not realize that “just looking at how high a UAS can”, could lead to a collision with a commercial aircraft (ILT, 2018:2).

• Intentional nuisance

With intentional nuisance is meant that the UAS-operator is flying somewhere where he or she deliberately causes inconvenience. This is the case when someone is flying over an accident side, where for instance a trauma helicopter tries to land. When the helicopter pilot sees a UAS, he needs to abort the landing (RTL Nieuws, 2018).

• Propaganda

Propaganda is one of the strong forces of the Islamic State (IS) (Wever, 2016). In that light, you might think that IS could make aerial videos of a certain target. Just showing a video can create a certain threat. In addition, they can also show what they are capable of with a UAS, for instance flying near on of the residents of the Royal family.

• Harassment

For example, this is the case when someone deliberately spies on the neighbor who is sunbathing in the garden. The privacy aspect will not receive further attention in this study, but is an important point of attention when using a UAS.

• Smuggling

Primarily, prisons and borders suffer from this threat. A UAS can be used to easily fly over fences / walls to drop a package, without being seen (BBC, 2018b).

• ISR (intelligence / surveillance / reconnaissance)

Civil UASs can also be used to make recordings to use these to gather information about a specific area. This can be seen as a threat, because one does not know the intent of the use of this knowledge.

• (Direct) Attack – with or without payload

When UASs are used to carry out an attack they can be provided with, for instance, an explosive or a substance. When the UAS flies close to the target, the operator detonates the explosive or distributes the substance. This is called a UAS with payload. When someone uses a UAS to carry out an attack by only using the UAS, then it is a UAS without payload. The operator then flies the UAS directly to a person or object.

4.1.4 Counter UAS Measures

To give a clear picture of what UAS countermeasures entail, the recent problems around London Gatwick will be used to illustrate these. Shortly before Christmas, the airport was closed for more than 36 hours because there were more than 50 drones seen. Many people were asking why it took so long to have proper counter-UAS technology deployed. The reason is somewhat more complicated and requires a certain explanation about the

working of UASs. Fig. 4. Illustrative numbers of the UAS (drone) disruption (BBC, 2018a).

Most drones are controlled by signals transmitted by the pilot, via radio signals, to the UAS in the air. In the event of the loss of these signals - for example because the UAS is flying too far away from the pilot - the UAS will basically become unmanageable. The UAS can respond in different ways: the aircraft returns to the place of takeoff,

lingers in place until the radio connection is restored, or automatically lands on the ground. In fact, this would mean that when these signals can be (intentionally) disturbed, a UAS could be stopped if it flies in a place where it is not allowed. Manufacturers of so-called UAS jammers make grateful use of this knowledge They developed equipment that deliberately disrupts the pilot's radio signal, for example by sending radio waves of a large intensity to the device. There is one problem: the law in the Netherlands and in Europe prohibits the use of jammers:

“In accordance with the Dutch Telecommunications Act article 15.1 it is not permitted to disrupt radio signals in the Netherlands. The disruption of GPS signals is also prohibited given the fact that many (vital) systems depend on these signal", according to the Agency Telecommunications (Jager, W. 2018). It is not surprising that jammers are not (yet) allowed. If you jam a GPS-signal, that could affect many devices. If you see a normal type of UAS (DJI-Phantom) fly and want to jam its

signal, it has to jam at 2.4 GHz. If this signal falls out, then the Wi-Fi can also fail for a while. But also certain systems of the Police operate at 2.4 GHz. Thus, the impact

of the jammer can be massive. Even for the military it is difficult to have jammers available, let alone to use it. The reason for this is that jammers are not part of the

‘means of force’ that military personnel are allowed to use (DGLC, 2018).

For the same reason, the police forces in London couldn’t use a counter UAS device like jammers (Farmbrough, 2018). They were considering shooting the UASs out of the sky, but decided not to do this because of concerns about stray bullets (BBC, 2018a). That is why the military was asked to help with this problem. They have more knowledge and experience, and moreover material, to be able to get the UASs out of the air. How exactly this has been done has not yet been confirmed.

As a result of these events, they are currently working in the UK to see whether they can change the legislation on the use of countermeasures by the authorities in order to defend airports for instance and to equip them with signal-blockers or hunter drones (Murphy, 2019).