To Learn or Not to Learn: How External Factors Influence Organizational Learning Processes of Dutch Safety Regions

(1)

1

To Learn or Not to Learn:

How External Factors Influence

Organizational Learning

Processes of Dutch Safety

Regions

Justin Selier, BSc. – s2038242

10-JUN-2018

Thesis CSM (Final Version)

Supervisor Thesis: drs. W. Broekema

Second Reader: dr. S. Kuipers

(2)

2

1. Introduction

Even though globalization has changed the value of local and regional governance (Flint & Taylor 2011: 253; Murray & Overton 2015: 151), crises prevention and reduction in the Netherlands is still mainly being managed on a regional scale (Ministerie van Veiligheid en Justitie 2013). In the Netherlands, the twenty-five ‘Safety Regions’ (Dutch:

Veiligheidsregio’s) have the legal responsibility to maintain an adequate professional

organization that can deal with crises and disasters (Ministerie van Veiligheid en Justitie 2013: 13). The Dutch National Government, however, tests these ‘Safety Regions’ on their performance, as the ‘Inspectorate of Justice and Security’ publishes the ‘Status of Emergency Response’ (Dutch: Staat van de Rampenbestrijding1+2_{). This report tests on what factors of} emergency response individual safety regions can and must improve. In the latest version of these reports (Inspectorate of Security and Justice 2016a), the researchers show that multiple safety regions still, six years after they were introduced, do not meet national standards, while other safety regions do meet these standards. These differences have formed the main reason academic research into Dutch safety regions’ performances is necessary.

As the ‘Status of Emergency Response’ identifies different issues for different safety regions, it can be concluded that the individual safety regions have their own individual strong and weak points. The report does not address why the different safety regions perform differently, nor for what reasons they improve or not. This paper proposes research to find out whether and what underlying external factors lead to improvements in performance by addressing the research question: ‘What external factors can explain whether Dutch Safety Regions learn or not?’. To do so, a particular focus will be put on learning from earlier Status of Emergency Response reports, and addresses the sub-question: ‘To what extent does crisis management performance of Dutch Safety Regions improve’?

Through finding these external factors, we gain a better academic understanding of how local and regional governments deal with the responsibility for emergency response. It can provide insights in organizational learning from a perspective that has not yet been used in the

academic debate. Also, through an analysis of factors that have been discovered by other authors, the current academic debate on factors contributing to organizational learning will be

1_{The 2016 Version can be found here:}

https://www.inspectie-jenv.nl/Publicaties/rapporten/2016/12/07/staat-van-de-rampenbestrijding---rapporten

2_{The 2013 Version can be found here:}

https://www.rijksoverheid.nl/documenten/rapporten/2013/05/24/rapport-staat-van-de-rampenbestrijding-de-regioprofielen

(5)

5

tested. Societally, addressing influencing factors that lead to organizational learning for safety regions could help these safety regions with improving on their performance, through which society should gain more protection from crises and disasters. This study could therefore teach the Safety Regions what external factors should be considered when they are aiming for improvements. These external factors derive from earlier academic studies and will now be tested in the field of the Safety Regions. Also, besides recommendations to the safety regions, external actors, who are expected to have an impact on external factors, gain knowledge on how they are able to improve a safety region’s capabilities.

1.1. Safety Regions as Crisis Management Organizations

In 2010, the ‘Law Safety Regions’ (Dutch: Wet Veiligheidsregio’s) was implemented3_{. This}

law created twenty-five regions that became responsible for internal crisis management. The safety region was created to form an organization in which those in charge of crisis

management, namely the mayors of individual municipalities, were able to cooperate. The main responsibilities for this new organization are: (i) making inventory of fires, disasters and crises, (ii) advise those in charge during crises, (iii) prepare for crises and crisis prevention, (iv) maintaining the fire service and regional health service, (v) create an alarm room, (vi) buy and share vehicles and other necessary equipment and (vii) create a platform to share

information between all relevant crisis management and prevention actors (Ministerie Veiligheid en Justitie 2013: Artikel 10).

In practice, this means that the Dutch safety regions do not consist of a fixed number of employees or other crisis managers. The safety region is namely a platform for crisis and disaster management and prevention and can therefore be related to many different actors, including those from private entities. It is important to know that the organizations must set their own policy goals and that they must meet nationally set standards (Ministerie Veiligheid en Justitie 2013: Artikel 14+15). It is therefore possible that these crisis management

organizations perform differently, as they hold different points of view on what policy plans work best.

(6)

6

2. Theory

2.1. Crisis Management and Crisis Prevention

Rosenthal et al. (1989: 9) define a crisis on the basis of three characteristics. Firstly, it is a situation in which the perceptions of a severe threat to society rise. Secondly, the situation contains a high degree of uncertainty. Thirdly, a response to the situation is quickly needed, as such there is a high degree of time pressure. A crisis therefore is almost always a large-scale incident that requires multiple actors to cooperate in order to manage it. Crisis management is often seen as “a holistic process involving prevention, planning, acute response, recovery and learning” (Boin & McConnell 2007: 52). This process plays a major role in every

organization, including governments, that has to manage crises.

To prevent large-scale incidents from happening, several measures can be taken. Academic theory on crisis prevention is, however, widespread, as many authors believe in different approaches. The main consensus lies in the fact that many authors, including Turner (1976) and Boin & McConnell (2007), see a failure of foresight as the main causes of crises. Organizations, including governments, should, per Turner (1976), invest in intelligence gathering. When an organization foresees a crisis, it is namely able to implement changes to avoid the crisis. This implies that organizations that are responsible for crisis prevention should set up policies that provide a clear outline on how specific intelligence is gathered. If this policy is namely effective, the organization should be able to prevent crisis accordingly. Besides foreseeing crises, organizations can also learn from previous incidents, disasters and crises in their quest to prevent a new one. Smith & Elliott (2007) use Turner’s (1976) notion of ‘cultural readjustments’ to argue that organizations should invest heavily in learning from crises. The difficulty, Smith & Elliott (2007: 521) argue, is, however, that learning from a crisis is supposed to take place after a crisis, which, in itself, is a time in which the legitimacy of the organization is being questioned. Learning often takes place in the form of identifying malfunctioning policies, systems and protocols. These mistakes are a defect of knowledge, which can be restored by learning from a performance or inquiry report. Smith and Elliott (2007: 533) namely indicate that “learning occurs at an individual and group level and that it requires the recognition that all is not well with our current state of knowledge.” One can see the performance and inquiry reports as sources of intelligence, of which Turner (1977) and Boin & McConnell (2007) argue are important in the realm of crisis prevention.

(7)

7

2.2. Organizational Learning

Within crisis management, organizational learning thus plays an important role. Per Stern (1997) however, the concept of learning is much debated in the academic debate. The question that many scholars try to answer is “who or what learns” (Stern 1997: 70), questioning the scale level of learning. The essence of learning is that it “is a continuous process which is inherent in the very being of organizations” (Nicolini & Meznar 1995: 738). Learning involves acquiring knowledge and skills (Stern 1997: 69) and organizations are continuously doing so (Nicolini & Meznar 1995). Organizational learning is, per Nicolini & Meznar (1995: 738), a process that often happens unnoticedly, as organizations are

transformed through “actions and enactions”. Organizational learning is learning that occurs on a very specific scale, namely that of a meso-level, in which indidivuals in a small social-structure learn and share new knowledge through communication (Stern 1997). Through combining the insights of Stern (1997) and Nicolini & Meznar (1995), in this study

organizational learning is defined as a continuous process of small social-structure knowledge

sharing that improves an organization’s performance.

This does not mean, however, that the continuity of organizational learning per definition leads to an ever-increasing performance of organizations. Internal factors can cause what is described by Nicolini & Meznar (1995: 732) as “organizational unlearning”. In essence, Hedberg (1981) argues that malfunctioning internal systems, such as bottle-necks in

communication, self-interest by managers and the complexity of the organization can cause an environment in which new knowledge is either not acquired or not shared with relevant

others.

What thus remains is what factors contribute to organizational learning. Whereas Hedberg (1981) identifies factors that hinder organizational learning, multiple studies have succeeded to distinguish a variety of factors that ought to help an organization in increasing its

performance. Besides numerous internal factors, scholars have tried to figure out what external factors contribute to organizational learning. In the remainder of this theoretical framework, the most common external factors that are expected to positively influence a process of organizational learning will be identified and explained.

2.3. Political Control

In the ideal world, crises do not occur and when they do, they would never reoccur due to organizations learning from previous and others’ mistakes. In the real world, however, “many

(8)

8

scholars are markedly sceptical about the learning capacity of policy-makers and

governmental organizations and argue that governments learn poorly or slowly at best” (Stern 1997: 69). The reasons behind the learning processes of governmental organizations after crisis situations are researched by Stern (1997) through questioning (i) what factors contribute to learning and (ii) what factors hinder learning.

Stern (1997) firstly argues that crises provide a great opportunity for organizational learning. Crises namely provide a new way of thinking for relevant policy makers and other actors; they create a shift in the general discourse. Stern (1997) provides numerous examples, such as the Chernobyl power plant disaster and the Cuban Missile Crisis, through which policy

makers changed their ideas on security. The political nature of crisis is, per Stern (1997), vital in the question as to whether organizational learning takes place or not. Issues being highly politicized can be an important incentive for organizations to start a process of learning. From this point of view, organizational learning would be facilitated by a disastrous situation and/or its great political attention.

However, Stern (1997) believes that there are multiple factors that hinder organizational learning, even after or during crisis situations. Stern (1997) firstly explains how the so-called ‘threat-rigidity hypothesis’ can lead to stiffness in the way decision makers think. This is due to the process of people’s field of view being narrowed and power being centralized during crises: when this happens, “dominant modes of thought” (Stern 1997: 77) take the upper hand, meaning that organizational learning is very unlikely to occur. Secondly, defensiveness and opportunism of decision makers leads to an absence of organizational learning, as these decision makers defend their position during crises by arguing that it was never their

responsibility and during success arguing it was fully their responsibility. To start a process of organizational learning, those in charge must admit something was wrong in the first place, which seems to be a tough acknowledgement.

One way to overcome these internal struggles towards organizational learning is by

organizations being subject to external or political control. Stern (1997) already demonstrated how greater political attention can lead to a higher chance of organizational learning taking place. Ebrahim (2005: 59) argues that accountability, that includes organizational learning, is a “process of holding actors responsible for actions”. Because of this, external actors, such as political organizations on different scale-levels, can hold an organization accountable for learning. Political control can thus, with the help of Ebrahim (2005: 59) and Stern (1997) be defined for this research as being the process of holding organizations responsible for their

(9)

9

performance. It is expected that a higher amount of political control leads to a higher chance

of organizational learning taking place. Organizational learning can namely be induced by political actors holding organizations accountable for their performance, and if the

performance can be increased, the political realm will most likely demand actions to do so.

Hypothesis 1: The higher the amount of political control, the more organizational learning occurs.

2.4. Media Attention

Stern’s (1997) theory on the barriers to learning from crisis situation provide multiple internal or organizational factors that explain why learning from crises does not occur. External factors, however, play a less important role in Stern’s (1997) thesis, apart from the political environment being one important factor. To find external factors that possibly influence organizational learning, a look at Broekema et al. (2017) will help. They provide an overview of factors that have contributed to organizational learning, both internally and externally to organizations. One of the factors that plays an important role in hindering organizational learning is, per Broekema et al. (2017: 329), sense-making, and especially media attention to great issues. They (ibid.: 329) argue that “the many interpretations that circulate in the media together with large streams of subjective and ambivalent information make it difficult to formulate concrete crisis lessons.” Organizations would thus benefit from few media attention, as this would help formulate more concrete lessons to be learned, making the process of organizational learning easier.

On the other hand, if no media attention is given to certain organizational issues, what are the societal incentives for organizations to improve on such issues? Dekker & Hansén (2004) argue that media attention can lead to organizational learning, given that the information spread by the media is concise and clear. “Not all information [presented by the media]

appeared suitable for translation into organizational lessons. Information was often inaccurate, ambiguous, or contradictory. At times, therefore, organizations did not see the wood for the trees. Access to information thus benefits organizational learning, provided that the message is clear and unambiguous. Otherwise, it can easily contribute to information overload.

(Dekker & Hansén 2004: 221).” The latter part of Dekker & Hansén’s (2004: 221) quote fits in seamlessly with Broekema et al.’s (2017) theory. Dekker & Hansén (2004) contrary to Broekema et al. (2017) see, however, a possible situation in which media attention can lead to organizational learning.

(10)

10

Using both Broekema et al. (2017) and Dekker & Hansén (2004), media attention can be defined as information presented by the media that provide insights in the performance of an

organization. Information produced by media that do not cover any issues or positive aspects

of an organization are to be neglected when researching organizational learning as this information is highly unlikely to provide the organization with an incentive to improve on their performance, or, in other words, invest in organizational learning.

Hypothesis 2: The more clear and unambiguous media information is available, the more organizational learning takes place.

2.5. Financial Situation

A third factor that could influence organizational learning is also identified by Broekema et

al. (2017: 333). The budget at hand for the organization is likely to impact the extent to which

organizations invest in organizational learning. Many organizations, including non-governmental organizations (NGO’s) and non-governmental agencies, receive funding from external organizations (Ebrahim 2005). The budget of these non-profit actors is therefore heavily reliant on external actors. Broekema et al. (2017) argue that cuts in the budget have led to quicker and more adequate organizational learning processes by governmental institutions. These budget cuts were namely the result of inadequate policy, making

organizational learning an incentive to get this money back in the future. Budget cuts due to inadequate management can thus lead to organizational learning, given that there is still enough budget to implement change. (ibid.).

The financial situation of organizations, however, entails much more than just budget cuts and increases. The way in which the budget at hand is invested is also likely to impact

organizational learning. One can hypothesise that investments in a learning department are much more likely to positively impact organizational learning than investments in new office buildings. Defining financial situation with an eye on organizational learning must thus take into account the way in which money is spend and the amount of money that is actually made available by external actors. The definition of the financial situation that will be used in this thesis is therefore the amount of money that is made available to a non-profit organization by

external actors and the way in which only this money is spent.

Using the theory of Broekema et al. (2017) and some common sense, a decrease in the amount of money that is available should lead to organizational learning, as this provides the

(11)

11

organization with an incentive to perform better in order to regain their old budget. The money that is given to an organization should be invested in factors that can contribute to organizational learning, such as policy evaluations, more man-power or exercises in order for organizational learning to actually take place. For such investments to take place, the external actors still need to provide the organizations with an adequate budget.

Hypothesis 3: Budget cuts lead to organizational learning, but only if some of the remaining budget is invested in factors that can contribute to learning.

2.6. Focusing Events

The factors mentioned above, namely political control, media attention and the financial situation of an organization, are all subject to one factor that can influence them. Birkland (1997) describes how so-called ‘focusing events’, crises and disasters that are directly linked to an organization, can change the political agenda in such a way that specific policies in the same realm as the ‘focusing event’ can change due to new political priorities. In other words, after a focusing event, the political control is likely to be increased. This seems to line-up with Stern’s (1997) thought on crises as organizational learning facilitators. Such ‘focusing events’ are often situations that can be classified under Rosenthal et al.’s (1989) description of a crisis.

Besides impacting political control, focusing events also lead to heightened media attention. The heightened media attention can go two ways, either the information provided is clear and unambiguous, as such it can assist the impacted organization with organizations learning (Dekker & Hansén 2004), or the information is unclear and contradictory, which would be a barrier to organizational learning (Broekema et al. 2017). Focusing events can also lead to changes in the financial situation of organizations as priorities have shifted. The organization might invest more heavily in man-power and policy change to prevent another (similar) focusing event from occurring. Such investments could lead to more organizational learning. The definition of a focusing event is given by Birkland (1998: 54) and is “an event that is

sudden; relatively uncommon; can be reasonably defined as harmful or revealing the possibility of potentially greater future harms; has harms that are concentrated in a particular geographical area or community of interest; and that is known to policy makers and the public simultaneously”. For organizations whose sole responsibility it is to prevent

(12)

12

management took place. For such organizations, namely, the mere occurrence of a crisis is not a focusing event as it is the organizations responsibility to manage it. When mismanagement took place, then it can be defined as a focusing event for crisis management organizations.

Hypothesis 4: The occurrence of a ‘focusing event’ (Birkland 1998) will impact political control, media attention and financial situation. Subsequently, it impacts organizational learning.

Summarizing the findings of the academic debate, whether organizational learning takes place is subject to multiple external factors. One reoccurring theme is that heightened political control to issues with the organization can be a major motivation for change, whether this control stems from a crisis situation (Birkland 1997; Stern 1997) or not. Other factors that might explain why organizational learning takes place are cuts in budget due to inadequate management (Broekema et al. 2017) and the subsequent patterns of investment and few but clear and unambiguous media attention to issues (ibid.; Dekker & Hansén 2004). Table 1 shows the four hypotheses that derived from the analysis of the academic debate and by which authors they were identified.

H# Hypothesis Authors

1 The higher the political control, the more organizational learning occurs.

Stern (1997); Ebrahim (2005)

2 The more clear and unambiguous media information is available, the more organizational learning takes place.

Dekker & Hansén (2004); Broekema et al. (2017)

3 Budget cuts lead to organizational learning, but only if some of the remaining budget is invested in factors that can contribute to learning.

Broekema et al. (2017)

4 The occurrence of a ‘focusing event’ (Birkland 1998)

will impact political control, media attention and financial situation. Subsequently, it impacts organizational learning.

Birkland (1997 & 1998)

Table 1: The Hypotheses derived from an analysis of the academic debate on organizational learning.

(13)

13

3. Methodology

To research the question ‘What external factors can explain whether Dutch Safety Regions learn or not?’, a comparative case study design will be used. In section 3.1., the reasons for this choice will be put forward. Firstly, the two cases will be selected on the basis of the two ‘Status of Emergency Response’ reports (2013; 2016a), as these reports shows how every single one of the twenty-five Dutch Safety regions has performed in both the period of 2010-2013 and 2014-2016. By doing this, the performance of both Safety Regions during these two periods can be identified, which allows to find one Safety Region that has experienced

organizational learning and one that has not. By comparing the influence of external factors on both a Safety Region that has and a Safety Region that has not experienced organizational learning, its differences in external influencing factors could explain how this came about. Secondly, to understand why these safety regions learn, a process-tracing method will be used. By using process-tracing, an essentially better understanding of how these external factors influence organizational learning can be achieved.

Vennesson (2008: 232) explains how the use of process-tracing techniques contribute to the understanding of causality and causal mechanisms. “Process tracing also provides an

opportunity to combine positivist and interpretivist approaches in the making of a case study, allowing the researcher to explore both the causal ‘what’ and the causal ‘how’” (ibid). The aim of this thesis is to find out what external factors contribute to organizational learning by Dutch safety regions, but also how this causality plays out. By using process-tracing as a method, that is explaining why one variable has had a certain outcome (Collier 2011), these question can be answered.

The internal validity of this research should generally not become an issue, as a case study is a design in which the conclusions match reality as closely as possible when methodologically done right. By choosing two different cases, it will be possible to identify differences and examine similarities, which in turn can teach the extent to which external factors influence organizational learning. The external validity is, however, less guaranteed. As the thesis seeks for differences and commonalities between just two of twenty-five safety regions, it will be impossible to generalize those findings for all safety regions in the Netherlands. The focus of the generalizability should thus be on the organizational similarities between all safety regions, as presumably all safety regions are subject to the same external factors. A further study could prove these findings. Also, the identification of external factors should be more easily generalizable.

(14)

14

3.1 Comparative Case Study Design

A case study is defined by Yin (2014: 16) as “an empirical enquiry that investigates a

contemporary phenomenon (the “case”) in depth and within its real-world context, especially when the boundaries between phenomenon and context may not be clearly evident.” The word context plays an important role in the decision whether to choose for a case study design or not. As the aim of this research is to find out why performances differ in similar regions, the context of these regions plays an important role. This research thus fits Yin’s (2014: 16) definition of a case study.

The comparative case study design is most fitting when researching the contextual and policy reasons behind the performance of Dutch Safety Regions. First of all, the number of safety regions is too small to even consider a quantitative approach. For a quantitative analysis, a larger number of cases (N) is necessary to gain significant results. Besides a lack of N’s, as there are just twenty-five safety regions, research into external factors and their subsequent impact on organizational learning asks for a thorough research beyond the ‘to what extent’ question. In this paper, the aim is to find out ‘why’ safety regions experience organizational learning, a question that is not suited to be answered through quantitative analysis, as such an analysis can only indicate ‘that’ they do. Especially the process-tracing method is most fitting in this case study design.

Second, a comparative case study design is, for this research, more preferred over a singular case study design. The findings from a comparative case study design are “considered more compelling, and the overall study is therefore regarded as being more robust” (Yin 2014: 57). Yin (2014) argues, however, that the major downside of a comparative case study design over a singular case study design is the amount of work the researcher has to commit to. This downside has no further implications on the quality of research into the performance of safety regions. Moreover, a comparative case study allows the findings from one safety region to be valued through the findings of another one: especially when considering a dichotomous variable, whether organizational learning took place, and when two cases are being researched, counterfactuals increase the internal validity of the conclusions (Sprinz & Wolisnky-Nahmias 2004: 25).

The critiques on case study design are often ontological in nature. Appleton (2002) reviews literature on case study designs and argues that there is no clear definition of a “case”. Yin (2014: 31) does not give a clear definition of a case indeed, yet he describes how the selection of the case, or the unit of analysis, which in a case study design are the same entity, is

(15)

15

dependent upon the research question the researcher tries to answer. This means that Appleton (2002) is right in the sense that there is no clear understanding of what a ‘case’ is. In this research however, the cases and units of analysis are clearly defined as a Dutch Safety Regions, which are politically bounded entities with defined responsibilities. As such, the cases in this research are not subject to vagueness.

3.1.1. Within-Case Level of Analysis

Rohlfing (2012) describes two different sorts of comparative case-study analyses, namely the within-case level and the cross-case level. The cross-case level is most fitting when the

research question focusses on discovering causal effects, when the main idea of the research is to show whether variable X influences variable Y in any way, and will therefore be used in this research. For the research of this paper, the cross-case level would be fitting if the main question that sought to be answered is only what external factors lead to a change in the performance of Safety Regions. The within-case level of Rohlfing (2012) is best applicable to this study, as it seeks to find what chain of events, or what process, cause the effect that has been determined by the cross-case level of analysis. In other words, the ‘why-question’ can be answered through within-case research. This study tries to find out why Safety Regions’ performance changes negatively or positively due to external factors. Therefore, the within-case level of analysis is being used.

3.2. Case Selection

To find out what causes the relationship between crisis management policies and the Safety Regions’ performance, it is important that as few other variables as possible are at play. The case selection is therefore of utmost important; the cases that will be selected should differ in policy, differ in performance, but differ in as few other aspects as possible. To find two of such cases, a list of so-called ‘scope conditions’ has been put up. Identifying a number of scope conditions should generally lead to a lesser need of control of causes (Yin 2014: 34). The time-frame of the scope conditions is set at 2016: the selected Safety Regions must have met the scope conditions in 2016 in order to be applicable to be researched.

The first scope condition is that the Safety Region should not be completely urbanized. It should therefore not be responsible for one of the four largest Dutch cities, namely

Amsterdam, Rotterdam, The Hague or Utrecht. It is expected that urbanized Safety Regions have to deal with very specific urban challenges, that cannot be generalized and would therefore be inapplicable to compare, especially to more rural Safety Regions.

(16)

16

The second scope condition that has been used to select cases is that the Safety Regions should not be too small geographically. For this research, this means that the safety regions that will be selected should cover at least half of a Dutch province’s geographical size in terms of both the area (in square kilometres) and population. The smaller the Safety Region, the easier it will be for the Safety Region to comply with national standards on response time and cooperation with the municipalities the Safety Region is responsible for, which are both indicators for their performance.

Thirdly, in order for the safety regions to be useful to answer to research question, one of them must have experienced organizational learning and one of them must not. Table 2 provides and overview of the scores of all of the twenty-five Dutch safety regions. It shows that the score of thirteen safety regions decreased, while the score of twelve safety regions increased. Of course, the lower the starting point, the easier it presumably is to increase on the score. In order to make sure that the performance change is noticeable, the selected cases should differ at least one point in 2016 compared to 2013; one of the cases must have scored one point higher and the other one point lower. Also, all the cases that scored below 10 in 2013 will be disregarded, as it was presumably too easy to improve on such a score as indicated by table 2, in which all such safety regions show an increase in score.

(17)

17

The Performance Scores of the 25 Dutch Safety Regions as Measured in 2013 and 2016

Table 2: The Performance Scores of the 25 Dutch Safety Regions as Measured in 2013 and 2016 (source: Inspectorate Security & Justice 2013, -2016a) (edited by Author)

Safety Region Score 2010-2012 Score 2013-2015 Change Brabant-Noord 5.188 11.667 6.479

Zuid-Limburg 5.750 10.333 4.583

Noord- en Oost- Gelderland 5.150 9.500 4.350 Gooi en Vechtstreek 6.450 10.667 4.217 Utrecht 8.308 12.000 3.693 Midden- en West-Brabant 10.850 13.500 2.650 Noord-Holland Noord 9.933 12.333 2.401 Flevoland 9.470 11.500 2.030 Zaanstreek-Waterland 9.725 11.333 1.608 Fryslân 11.088 12.500 1.413 Limburg-Noord 10.363 11.000 0.637 Haaglanden 11.063 11.667 0.604 IJsselland 11.725 11.500 -0.225 Zeeland 10.988 10.667 -0.321 Hollands Midden 12.913 11.833 -1.079 Gelderland-Midden 13.558 12.333 -1.224 Amsterdam-Amstelland 11.688 10.333 -1.354 Kennemerland 12.725 11.167 -1.558 Gelderland-Zuid 11.875 10.167 -1.708 Groningen 14.525 12.500 -2.025 Drenthe 11.383 8.667 -2.716 Zuid-Holland Zuid 15.038 12.167 -2.871 Twente 14.188 10.833 -3.354 Brabant-Zuidoost 12.225 8.833 -3.392 Rotterdam-Rijnmond 15.715 11.667 -4.048 Average 10.875 11.227 0.352

(18)

18

The two Safety Regions that have been selected, and thus meet the two scope conditions explained above, are Safety Region Fryslân (VRF) and Safety Region Drenthe (VRD). Table

3 shows how these two safety regions compare on geographical variables. Not only do these

two Safety Regions meet the scope conditions, they also share a border and both are congruent with a province. This will make the cooperation variable more interesting and useful to research, as the cooperation policy of the one Safety Region will influence the cooperation performance of the other. This will reduce interference of the data coming from other Safety Regions. The geographical composition of those two Safety Regions will be described in the Analysis chapter of this study. Secondly, these safety regions score different on their performance, which is ultimately necessary to find causal mechanisms.

Variable Drenthe Fryslân

Size of the Population (2016)* 488.269 646.040 Size of the Region (km2)* 2.683 3.250

Capital City Assen Leeuwarden

Relative Location in the Netherlands

North East North

*Source: CBS (2018)

Table 3: Geographical Comparison of the Selected Cases

3.3 Operationalization

The main research question identifies multiple variables: ‘Financial Situation’ (X1), the ‘Amount and Clarity of Media Attention’ (X2) and the ‘Amount of Political Control’ (X3) are expected to explain ‘whether organizational learning took place’ (Y). Secondly, the

occurrence of a focusing event (X4) might have impacted the variables X1-3. The ‘Status of Emergency Response’ report has already identified the data regarding variable Y: it explains how the safety regions have performed on numerous indicators during two periods. The report considers policy, as it researches the policy goals and whether these goals match national legislation. The report does not compare external factors between safety regions and whether change or the absence of change in the performance of the safety regions can be explained through these external factors. In short, the report can be used to identify variable Y, namely

(19)

19

the performance of the safety regions and its changes, but lacks information on the independent variables.

As the Status of the Emergency Response report does not cover the independent variables, another way to gather and analyse them must be found. From the theory, it has become clear that multiple independent variables play an important role in whether organizational learning takes place or not. In this sub-section, these variables will be operationalized. As such, they can be used to research their impact on changes in Safety Regions’ performances. Also, it must become clear how and when organizational learning has taken place. For all of these variables, the data that will be gathered will have a specific time frame. All relevant data must come from either between March 1st 2010 and October 31st 2012 (period one), or March 1st 2013 and October 31st 2015 (period two). As such, they meet the same time frame as the Status of the Emergency Response report. This is necessary because this report is the main source on whether organizational learning took place or not.

3.3.1. Operationalization of Organizational Learning

Most of the relevant terminology has been operationalized in the Status of Emergency Response report already. Learning has, however, not yet been operationalized. When an organization’s performance has increased (through small-structure knowledge sharing), learning has taken place. This will be measured by comparing the Status of Emergency Response 2013 with the 2016 version: if the safety region in total scores better on the indicators, an improvement in their performance has occurred. To do so, the indicators that are used in both reports must match. Unfortunately, the Inspectorate of Justice and Security (2013; 2016a) has used a different method to measure scores in the 2016 version. Appendix A shows how these different methods have been integrated into one score. Deducting the 2016 score from the 2013 should indicate if and to what extent organizational learning has taken place: if the score is positive, then it has increased and learning has thus occurred. If the score is negative or equal, their performance has decreased or stayed the same and organizational learning is absent.

A second way to analyse the data from the Status of Emergency Response reports is to identify whether the policy recommendations that were published in the 2013 version were followed in the years after. The Status of Emergency Response (Inspectorate of Security and Justice 2016a) 2016 version shows whether, in the eyes of the Inspectorate, the safety regions have followed the policy recommendations that were made three years earlier. The more policy recommendations were followed, the more organizational learning took place. It is

(20)

20

important, however, to note that, as some safety regions might have received less policy recommendations than others due to their policy already being deemed effective, quantifying the number of policy recommendations that were followed does not indicate how ‘good’ a certain safety region is. It merely indicates how much and in what aspects they have improved.

3.3.2. Operationalization of Financial Situation

Firstly, to gather data on the financial situation of the Safety Region, a close look at the budget that it was provided with, equally to what Broekema et al. (2017) conduct in their study, will be taken. The data on the budget of the safety regions will be taken either from the website of the safety region itself, or anywhere else where the google search term ‘budget safety region [Drenthe/Fryslân] [2010-2016]’ (Dutch: Begroting Veiligheidsregio

[Drenthe/Fryslân] [2010-2016]) yields the most relevant result. The budget is allocated and

explained on a yearly basis.

In the analysis, the year to year budget changes will be analysed. In the process-tracing section of the analysis, a close look at where the allocated budget was invested in will be taken. By doing so, it becomes clear how the budget of the two safety regions has changed during the relevant time frame, and possibly how this change might have had an impact on the safety regions’ ability to experience organizational learning.

3.3.3. Operationalization of Political Control

Political control will be measured by researching political debates in the most relevant city council of the Safety Region. As the Safety Regions that have been selected, Fryslân and Drenthe, both share the same borders as the similarly named provinces, the most important municipality is operationalized as the municipality of the capital of the province: Leeuwarden in the case of Fryslân and Assen in the case of Drenthe. Debates of these councils are, by law, recorded. The data gathering will thus be done through a thorough analysis of past political debates. To make this more doable, for both city councils, a search term of veiligheidsregio will be conducted on a webpage that contains all relevant political data.

During the data gathering, an issue occurred when researching the safety region Drenthe. The city council of Assen, namely, only stores the current year’s data on an open access website. This means that for relevant data of 2010 to 2016, a physical visit to a document storage must be paid. To avoid this issue, it has been decided that Drenthe’s largest city, Emmen, will be researched on their political control over the safety region, rather than Assen.

(21)

21

The analysis of these data will mainly be done by looking at how often a debate was about the Safety Region on a yearly basis. Secondly, the nature of the debate will be coded. In the process-tracing section of the analysis, the political control will be further analysed by researching which political actor gave attention, who the attention was directed to and whether the attention was positive or negative.

3.3.4. Operationalization of Media Attention

To gather data on the scope and direction of media attention, news items from regional newspapers safety regions will be counted (amount of attention) and coded (amount of different interpretations of the event), as such, it matches closely to Broekema et al.’s (2017) and Dekker & Hansén’s (2004) hypothesis (number two) that more media attention and different interpretations of the events can lead to less learning taking place. To find data, an online database of historic newspapers will be consulted. This online database, ‘De Krant van Toen’4_{, provide access to the regional newspapers of both Drenthe and Fryslân. In the search} bar of the database, for each time-period, the term veiligheidsregio [Drenthe/Fryslân] will be used. Then, the researcher will filter out all articles that do not concern performance by reading them all.

The analysis of the data will mostly be done by providing a summary of one sentence of these news items. These sentences will then be coded into either positive media attention or

negative media attention and will be, if applicable, matched to another quote from another news article. Matching similar summaries should help in finding out how unambiguous the attention was. Then, again, a year to year analysis can be conducted showing in which years media attention to safety regions was more present and how unambiguous it was over all those years.

3.3.5. Operationalization of Focusing Events

Crisis situations can be reasonably thought of as focusing events. What crises have in common is that they are almost always a large-scale incident that requires multiple actors to cooperate in order to manage it. It is here that the safety regions can play an important role. For a crisis to have an impact on organizational learning by safety regions, its occurrence must be mismanaged. It is the safety region’s responsibility to manage crises, and, if they do so decently, the crisis situation is unlikely to impact any of the three variables researched above. Only when crises are mismanaged by a safety region, they are likely to impact the

(22)

22

political control, financial situation and media attention of/to the safety region. To find out if and how such focusing events have had an impact on organizational learning in the safety regions Drenthe and Fryslân, a two-step process will be used. Firstly, the amount of ‘focusing events’ will be counted by using the number of Incidents in the safety regions. GRIP-incidents, an abbreviation of Coordinated Regional Crisis Management Procedure (Dutch:

Gecoördineerd Regionaal Incidentbestrijdings Procedure), are categorized on six different

levels: GRIP-1 being a small-scale incident and GRIP-5 and GRIP-Rijk being large scale and national incidents. In order for a GRIP-incident to fit Rosenthal et al.’s (1989: 9) definition of a crisis, it must be at least a GRIP-3 or higher. Otherwise, the incident does not pose a “severe threat” (Rosenthal et al. 1989: 9).

When the number and types of GRIP-3 or higher incidents is known, it is possible to check whether the safety region has managed the incident properly. To do this, several official reports, whether inquiry reports or others, will be used to analyse the safety region’s

performances. If the performance is deemed inadequate, it can be concluded that a focusing event took place.

Table 4 provides an overview of this subsection. It serves as a systematic summary of how the

factors that are expected to contribute to organizational learning have been made measurable and through which sources this data will be gathered.

(23)

23

Overview of the Operationalization of Factors

Factor Indicators Source

Organizational Learning -Scores provided by the Inspectorate

-Amount of recommendations followed

Status of the Emergency Response (2013; 2016)

Financial Situation Amount of money received from external actors on year to year basis

Internal Documents on budget

Political Control Amount of times performance of safety region is discussed in municipal council

Municipal Council Website

Media Attention How often and how positive or negative the safety region has received mentions in regional newspapers

‘De Krant van Toen’ website

Focusing Events The amount of 3 or GRIP-4 incidents that the safety region has mismanaged

Status of the Emergency Response (2013; 2016) plus other (inquiry) reports

Table 4: Overview of the Operationalization of Factors

3.4. Validity and Biases

Rowley (2002) discusses how four tests are used to determine the quality of empirical social research. These tests concern the construct validity, the internal validity, the external validity and the reliability of the research design and analysis. In this sub-section, each of these four tests will be applied to the research design described in this chapter. As such, it becomes clear that this research design meets the quality standard of empirical social research.

Firstly, the construct validity test is about “establishing correct operational measures for the concepts being studied” (Rowley 2002: 20). The main way to get to such ‘correct operational measures’ is to link the concepts to and data collection questions to propositions (ibid.). In this research, the concepts have clearly been linked to propositions or hypotheses that derived from an analysis of the academic debate. Also, reducing subjectivity, which is another

(24)

24

important issue that the construct validity test concerns, is being reached by using external sources for all variables regarding organizational learning. It is, in this research, not up to the author to decide which safety region experienced organizational learning, but by using different reports, this variable is being measured as objectively as possible.

Secondly, the internal validity is being reached when “a causal relationship whereby certain conditions are shown to lead to other conditions” (ibid: 20) is researched. The causal

relationship that this study researches is between multiple external variables that derived from previous academic work and organizational learning in Dutch safety regions. As the variables have already been identified as having an impact on organizational learning in other studies, it can be concluded that this relationship is non-spurious, making this study internally valid. Thirdly, the external validity of this case study design, which concerns generalization (ibid.) can be reached, per Rowley (2002: 20) through “analytical generalisation in which a

previously developed theory is used as a template with which to compare the empirical results of the case study. If two or more cases are shown to support the same theory, replication can be claimed.” This study uses previously developed theories on organizational learning and external variables that possibly impact it and applies it to organizations that have not yet been studied in such a manner, namely Dutch safety regions. This narrative fits Rowley’s (2002) idea on the external validity of a case-study design.

Fourthly, the reliability is about “demonstrating that the operations of a study, such as the data collection produced can be repeated with the same results” (ibid: 20). To do so, it is important that every step of the research is well-documented and well explained. To reach the desired level of reliability, the collection of the data as well as its analysis will be thoroughly explained in this study.

(25)

25

4. Results

The results of this study will be presented in this chapter throughout multiple sub-sections. Firstly, the outcome of the dependent variable, namely organizational learning, will be

analysed for both cases. Thereafter, the cases will be analysed on each of the four independent variables consecutively. Finally, the results of those sub-sections will be compared and the differences and similarities between the two cases highlighted. This structure fits a systematic approach that should enhance the replicability of the study and, in the process, meet the academically desired level of reliability of empirical social research (Rowley 2002).

4.1 Organizational Learning

Whether organizational learning has taken place or not has been measured by the Inspectorate of Security and Justice (2013; 2016a). The results of comparing the two ‘Status of the

Emergency Response’ reports have already been used in the case selection (sub-section 3.2.) of this paper. In this paragraph, the results of the comparison between the two reports will be looked at for solely the two selected safety regions, Drenthe and Fryslân, and how they compare to the average and to each other.

Table 2 shows the recalculated (see section 3.3.1. and Appendix A) scores for all safety

regions in the Netherlands over both the periods 2010-2012 and 2013-2015. The table shows that, understandably so, the safety regions that scored below average in 2013 in general are the safety regions that showed the most improvements in 2016. This is possibly due to the fact that the lower the score in the first place, the more room for improvements there is. On the other hand, what is more unexpected and striking, is that the safety regions that scored (well) above average in 2013 show a decrease in score in 2016. This could be an indication that organizational unlearning (Nicolini & Meznar 1995: 732) is very much present in Dutch safety regions. A more logical explanation, however, is that the Inspectorate of Security and Justice has used different indicators to measure the same performance indicators. A thorough look at the results shows that the Status of Emergency response in 2013 consisted of much more 100 percent scores on particular indicators than the 2016 version, in which a 100 percent score was seldom given.

(26)

26

4.1.1 Organizational Learning of the Safety Region Drenthe

In order to avoid making wrongful conclusions on the basis of this renewed methodology of the Inspectorate, identifying whether organizational learning has taken place has to be done relative to the other safety regions. The safety region Drenthe in 2013 scored half a point higher than the average of all safety regions. In 2016, however, they score two-and-a-half points lower than the average of all safety regions. It can therefore be safely concluded that there is no indication that the safety region Drenthe has experienced organizational learning based on the scores given by the Status of the Emergency Response reports.

Another indicator of whether organizational learning has taken place in Dutch safety regions is whether they followed the policy recommendations made by the Inspectorate of Security and Justice in 2013. For the safety region Drenthe, the Inspectorate made three policy

recommendations in 2013 (Inspectorate Justice & Security 2013: 55). The first one is that the role and functionality of the so-called ‘Regional Operational Team’ (ROT) should get a clear definition. Article 2.1.4 of the ‘Law Safety Regions’5_{provides a clear outline of who the ROT} should consist of and what their specific tasks should be. The safety region Drenthe has, in the eyes of the Inspectorate, not provided a clear enough regional application to this section. The second recommendation is that the safety region Drenthe should “come up with a solution for the response time so activities can start earlier” (Inspectorate of Security and Justice 2013: 55). Finally, the safety region Drenthe should improve on its information management by deploying a so-called ‘netcentric working model’. This model is “a way of working in which all relevant teams and organizations share information as quickly as possible with each other. This information is continuously being summarized to keep it clear for everyone involved” (IFV 2015: 1).

In 2016, the Inspectorate of Justice and Security concludes that just one of the three recommendations had been followed. The sole recommendation that was followed was the implementation of the netcentric working model. The Inspectorate (2016b: 10) argues that the information management system of the safety region Drenthe has improved, due to their information managers following a national course on netcentric working. As for the other two policy recommendations, the safety region Drenthe has decided to continue on the path they were working on before the Status of the Emergency Response report was published in 2013, namely to continue to adapt their ‘Drenths Model’. As such, they did not specify the roles of

(27)

27

the Regional Operational Team, nor did their response time issues get the attention the Inspectorate desired it would get.

Taking into account both the scores given on the multiple performance indicators in 2013 and 2016 and the fact that just one recommendation was followed by the safety region Drenthe, it can be concluded that no process of organizational learning took place between 2013 and 2015 in the safety region Drenthe. In the remainder of this chapter, it will be determined whether this was due to any of the four variables that were identified by author studies.

4.1.2. Organizational Learning of the Safety Region Fryslân

Returning to table 2 (section 3.2.), it becomes clear that the safety region Fryslân has performed quite uniquely when comparing the Status of the Emergency Response reports from 2013 and 2016. Fryslân, namely, is the only out of twenty-five safety regions that managed to improve its score with at least one point in 2016 after having scored at least eleven points in 2013. The organization also, together with the safety region Haaglanden, is only one of two safety regions that scored above the national average in both reports. Having explained in section 4.1.1. that getting an improvement on an already decent score was very tough to accomplish due to the Inspectorate of Security and Justice using a renewed

methodology to measure safety regions’ performances, it becomes clear that the safety region Fryslân has done extremely well in terms of their performance. Due to this difficult to gain increase in performance, it can be concluded that there are indications that organizational learning has taken place in the safety region Fryslân.

Secondly, looking at the recommendations made by the Inspectorate of Security and Justice in 2013, the Inspectorate concludes (2016c: 11) that the safety region Fryslân has followed all three recommendations. The first recommendation that the Inspectorate of Security and Justice (2013: 43) made was that the safety region Fryslân should organize a practice session or simulation concerning all main actors of crisis management in the region. Since 2013, however, most probably due to this recommendation, the safety region Fryslân organizes a so-called ‘system-test’ on a yearly basis. Secondly, the safety region Fryslân was recommended to include the municipalities in their training and practicing policies. Since this

recommendation, the safety region Fryslân has included teams of all municipalities that are responsible for safety and security of the population in its multi-disciplinary training and practicing. Thirdly, similar to the safety region Drenthe, Fryslân was recommended to

(28)

28

implement the ‘netcentric working model’ to improve on information management practices. In 2016, the Inspectorate (2016c: 43) concludes that information management has become a central issue in the safety region’s policy and that regular meetings are being held between people responsible for this issue. They conclude that the recommendation was therefore followed.

Given that the safety region Fryslân has managed to improve on an already above average performance score in 2016, and that they have followed all three recommendations made by the Inspectorate of Justice and Security (2013; 2016c), it can safely be concluded that the safety region Fryslân has experienced organizational learning. In the remainder of this sub-section, the influence of political control, media attention and the financial situation of/to the safety region Fryslân will be analysed to find out how this organizational learning process came about.

4.2. Political Control

4.2.1. Political Control of the Safety Region Drenthe

It is expected (hypothesis 1) that more organizational learning takes place when there is more political control. As the previous sub-section has indicated that the safety region Drenthe has not experienced organizational learning, for the hypothesis to be confirmed, little or no political control must have taken place. On the other hand, the safety region Drenthe scored relatively high on their performance in 2013, so the political control at that time might have been high. To research how the political control of the safety region Drenthe has evolved, data from the municipality of Emmen has been coded and researched.

In period one (March 2010 to October 2012), the municipality of Emmen has ‘controlled’ the performance of the safety region Drenthe six times. Table 5 shows the search results of the search term ‘performance safety region’. It indicates at what date and what specific element the municipality controlled of the safety region Drenthe. It shows that in three of the six instances political control was present, it was about finances and the justification of money spent.

Out of the other three, the document of May 12th 2011 is the most notable document that shows political interest in the performance of the safety region Drenthe. The second title of the document is “research of possibilities to improve the quality of crisis management” (VRD 2011: 1). The document is an internal analysis conducted by the safety region itself on weak

(29)

29

and strong points of the performance of the safety region Drenthe. The report was sent to the council of Emmen and was discussed in the meeting of May 12th 2011. According the Gemeente Emmen (2011: 9), one party, namely the ChristenUnie (Engish: United Christian

Party), asked for a discussion on this document. However, as stated in the records of this

meeting, due to technical difficulties, the document could not be opened at the time and the discussion on it was therefore cancelled. This instance does, however, indicate that political control over the safety region Drenthe was present at the time. Including the other documents and discussions, it can be concluded that the municipality of Emmen ‘controlled’ the safety region Drenthe six times in period one.

Political Control of the Safety Region Drenthe by the Municipality of Emmen in Period one and two

Table 5: Political Control of the Safety Region Drenthe by the Municipality of Emmen in Period one and two (sources can be found in Appendix B)

In period two (March 2013 to October 2015), there is one fewer instance in which the

municipal council of Emmen discussed the performance of the safety region Drenthe. The one

Date Source Control VRD Over About

15/09/2011 Gemeente Emmen YES Budget 12/05/2011 Gemeente Emmen YES Performance

31/03/2011 Gemeente Emmen NO Municipal Safety 23/02/2012 Gemeente Emmen NO Environment 12/06/2012 Gemeente Emmen YES GGD - Budget

12/04/2012 Gemeente Emmen YES Regionalisation Fire Brigade 12/06/2012 Gemeente Emmen YES GGD - Performance

04/11/2010 Gemeente Emmen YES Justification Money Spent

12/09/2011 Gemeente Emmen NO Municipal Safety 15/09/2011 Gemeente Emmen NO Municipal Organization 12/05/2010 Gemeente Emmen NO National Developments 06/12/2011 Gemeente Emmen NO Status National Health Care

---18/12/2014 Gemeente Emmen NO Municipal Rules & National Law 06/02/2014 Gemeente Emmen YES Budget

12/03/2015 Gemeente Emmen YES Performance of the Board

29/10/2015 Gemeente Emmen NO Change in the Law 25/09/2015 Gemeente Emmen YES Budget

13/06/2013 Gemeente Emmen YES Summary of the Year 2012 16/04/2015 Gemeente Emmen YES New Policy VRD

13/05/2014 Gemeente Emmen NO Finances GGD

04/12/2013 Gemeente Emmen NO Reaction to National Policy 13/05/2015 Gemeente Emmen NO Municipal Finances

(30)

30

instance that is missing is about a budgetary question, as, in period two, the municipality twice controls the budget and thrice another phenomenon. The most striking instance of political control in period two occurred on April the 16th 2015, when the municipal council discussed at length the new policy that the safety region Drenthe had published. On April 16th, a representative from the safety region Drenthe presented this new set of policies in front of a commission that included members from all of Emmen’s municipal political parties. The records of this meeting show that the most political parties are satisfied with the new policies regarding the new risk analyses, yet some, like the ‘Seniorenpartij’ (Party for the Elderly) and ‘LEF!’ ask the representative why certain risks are not included (Gemeente Emmen 2015: 1-3). This can be classified as a prime example of political control: the safety region presents new plans and the municipal council questions its potential effectiveness.

Taking into account this instance of clear political control and the fact that there is merely one less issue discussed by Emmen’s municipal council regarding the performance of the safety region Drenthe, it can be concluded that the amount of political control stayed the same when comparing period one and period two. This analysis shows that the municipality has the tools to control the performance safety region Drenthe. It is, considering how relatively new the organization of the safety region is, therefore interesting and perhaps worrisome that the council only used this tool eleven times in a five-year period. It can be concluded though that the decrease in performance of the safety region Drenthe is not due to less political control as hypothesised in hypothesis 1.

4.2.2. Political Control of the Safety Region Fryslân

As the previous subsection has made clear that the organization of the safety region Fryslân has experienced organizational learning, it is possible that, per hypothesis one, this is due to more political control taking place in period two. To find out how much political control the safety region was subject to, official documents like transcripts from meetings and debates from the municipal council of Leeuwarden have been coded and analysed. This subsection will show whether there was a difference in political control of the safety region Fryslân between period one and period two.

(31)

31

Political Control of the Safety Region Fryslân in Period One

Table 6: Political Control of the Safety Region Fryslân in Period One (source: Gemeente Leeuwarden6₎

Table 6 shows that, in period one, the number of times the municipal council of Leeuwarden

discussed or was informed of the safety region Fryslân is eleven. Only five of those times, however, the discussion can be categorized as political control. Similarly to the safety region Drenthe, about half of those times the political discussion involved finances and budgetary decisions. Out of the other three forms of control, the municipal council discussed a risk assessment conducted by the safety region Fryslân twice. A debate on this risk assessment, which involved potential risks that would harm not only Leeuwarden, but the entire safety region and province of Fryslân, led to no further changes. In fact, the chair of the council meeting, the mayor of Leeuwarden, concluded in a letter to the safety region (Gemeente Leeuwarden 2010: 1) that “the council of Leeuwarden had no further comments, nor regarding the risk assessment nor regarding the policy plans.” This shows that political control took place: the council discussed a product produced by the safety region and agreed that it was sufficient. It is, however, striking that the municipal council used their power to control (products of) the safety region responsible for the crisis prevention and management of their municipality only once in a three-year period.

6_{All documents can be accessed through:}

https://leeuwarden.notubiz.nl/zoeken?keywords=%22Veiligheidsregio+Frysl%C3%A2n%22&search=send&limit =10&gremia_html=&documenttype_html=&from=01-03-2010&until=31-10-2012&form_mode=simple (Retrieved May 11th_2018).

Date Source Control VRF Over About

19/05/2010 Gemeente Leeuwarden No Shared Services

14/06/2010 Gemeente Leeuwarden Yes How to Structure the Safety Region

13/09/2010 Gemeente Leeuwarden Yes Risk Assessments Conducted by Safety Region (debate) 27/09/2010 Gemeente Leeuwarden Yes Risk Assessments Conducted by Safety Region (vote)

07/12/2010 Gemeente Leeuwarden No Fire Services will not be Responsibility of VRF

13/04/2011 Gemeente Leeuwarden No Financial Expenses of the Municipal Council

24/10/2011 Gemeente Leeuwarden No Safety Region will cost the Municipal Council more

20/02/2012 Gemeente Leeuwarden No A Desired Change in how Certain Services are Named

26/03/2012 Gemeente Leeuwarden No ibid , but now the vote

13/06/2012 Gemeente Leeuwarden Yes Financial Situation and Municipal Expenses to VRF 25/06/2012 Gemeente Leeuwarden Yes ibid .

To Learn or Not to Learn: How External Factors Influence Organizational Learning Processes of Dutch Safety Regions