The Link of Functional Differentiation and Innovation

(1)

The Link of Functional Differentiation and Innovation

An analysis of the relationship between organisational complexity and the occurrence of participatory innovations in municipalities

Master Thesis, 10 January 2018

Leiden University, the Netherlands Faculty of Governance and Global Affairs

Author: Hendrik Ewens (s1804103) Supervisor: Dr. Joris van der Voet Second reader: Dr. Petra van den Bekerom

(2)

1 Abstract

In the 21st_{century, the public sector is faced with several challenges. Mega trends like globalisation}

and digitalisation are affecting the work of public organisations, but also do individual claims of citizens for increasing participation. The implementation of innovation is one way for public organisations to adapt to the changing environment and to meet new demands from citizens. However, it has been debated whether the organisational structure of public organisations is even supportive for innovation. Public organisations are recognized as a complex framework, which might harm the adoption of innovation. This is contrasted by findings of scholars, who state that complex organisational structures can facilitate innovation. Proceeding from this stand, this study investigates if the organisational complexity of a public organisation is positively related to the occurrence of innovation. The introduction of participatory budgeting will be used as an example of public innovation, as especially in the last decade participatory budgeting initiatives were implemented in municipalities in Europe. However, still a minority of municipalities adopted this innovation. One attempt of explanation for this variation are differences in the organisational complexity. Organisational complexity consists of two variables in this research. Functional differentiation is the first and expresses how many functional departments an organisation has, while organisational size is the second and consist of the number of employees. This study focuses on municipalities (N=394) in North Rhine-Westphalia (Germany) where participatory innovations in the budgeting process occurred in the last years. By means of binary logistic regression analyses, this study examines the relationship of organisational complexity and participatory innovations. The study has two main empirical outcomes. Firstly, the results highlight the importance of functional differentiation for a municipality, as it is positively and statistically significant related to participatory budgeting. Every additional functional department increases the likelihood of the occurrence of this innovation by the factor 1.098. Secondly, organisational size has no statistically significant relationship with participatory budgeting. One interpretation of the empirical findings is that it is not the number of people of an organisation that matters (i.e. quantity), but the functional expert knowledge they can bring in the organisation (i.e. quality). In addition, several indicators point in the direction that the size-innovation relationship is not statistically significant, as it is mediated by functional differentiation. Moreover, the focus from researchers on environmental variables to explain participatory innovation cannot be supported from this study, as organisational structure variables seem to be more relevant for participatory innovation than environmental factors. While it can be concluded that a higher functional differentiation of a public organisation might be beneficial for innovation, it competes with other values like efficiency and parsimony. Practitioners have to deal with this value trade-off, which perhaps cannot be solved.

(3)

2

1. Introduction

As the public sector has a central role in society, it must deal with a changing environment and changing requirements. Through current trends like digitalisation and globalisation, this change might be even accelerated. One solution to close the gap between the own organisation and new tasks is to create and adopt innovation (Bekkers, Edelenbos, & Steijn, 2011). If new problems cannot be solved with old solutions, innovation can be helpful. This might be problematic, as there is the opinion that the public sector is in general slow, backward and less innovative compared to the private sector (Raadschelders & Vigoda-Gadot, 2015). In recent years, public administration scholars tried to examine this field in order to find evidence for innovation in the public sector and came to the conclusion that the public sector is an environment, where innovation can emerge (Bekkers et al., 2011).

While there is a consensus among scholars that the public sector is also innovative, there is still a lack of evidence, which circumstances influence the occurrence of new ideas. One focus of studies in this field is to research the influence of the organisational structure on the occurrence of innovations in the public sector (Jakobsen & Thrane, 2016). Originating from normative standpoints and findings from the private sector, mechanistic organisations might be less innovative than organic ones (Burns & Stalker, 1961). Public organisations share characteristics with mechanistic organisations (Rainey, 2014) and one could assume that their rigid framework might be a reason for less innovation. However, organisations can also be examined by analysing specific parameters of organisational structures rather than on dichotomous characterisations. Specific parameters are for instance centralisation, formalisation and specialisation. While specialisation is assumed to be positively related to innovation, centralisation and formalisation might harm the innovativeness of an organisation (Damanpour, 1991; Zornoza, Boronat, & Cirpes, 2007). Other scholars focus on organisational complexity in their studies and they concluded that organisational complexity has a positive relationship with innovation (Aiken & Hage, 1971; Damanpour, 1991, 1996).

Not only a closer look at organisational structures is needed to find answers in the relationship between organisational context and innovation, but also a more detailed view on innovation itself. Even though more research has been done in this field, the definition of innovation varies or is often omitted; whereby the existing definitions have in common that they describe innovation as a novelty for the organisation (De Vries, Bekkers, & Tummers, 2016). Due to this broad description and the variation of findings, existing empirical research often examines specific types of innovation; such as the creation of a new product or service (product innovation), the changing of an existing process (process innovation) or an exchange between the organisation and the environment (ancillary innovation) (De Vries et al., 2016). Which type of innovation is pursued, depends on the problem the public organisation wants to solve.

(5)

4 For the sake of resolving problems of democratic deficits and to increase legitimacy, especially municipalities create new ways to deal with citizens. One new way consists of participatory budgeting: where citizens are involved in the process of budgeting and determine on what the public budget is spent on (Geissel, 2009). In general, a participatory and inclusive approach is used to increase the quality of democracy and, hence, gives citizens the opportunity to be part of political decisions (Cabannes, 2004). However, there are differences in the implementation of participatory budgeting among countries. For instance, in Germany, the goal was to reduce public spending and to encourage citizens to come up with ideas to cut expenses in a consultative phase without decision making rights (Geissel, 2009). This consultation is restricted to the part of the creation of a budget draft, which is the responsibility of the municipality.

Participatory budgeting is an innovation for municipalities. In a regular budgeting process, citizens are not actively involved (Lee, Johnson, & Joyce, 2013). The decisions, how to spend the budget, are mainly taken by civil servants and politicians. While civil servants make proposals for budget spending and work on the draft of the budget plan, a legislation body consisting of politicians decides finally on the budget (Lee et al., 2013). In a budgeting process with participatory elements, citizens are involved, as they can bring in their own proposals for the budget plan (Ebdon & Franklin, 2006; Geissel, 2009). In addition, from a time perspective, it is a relatively new concept for European states. For instance, in Germany, participatory budgeting started in the 1990s with isolated examples of only a few municipalities and remained a small percentage after the millennium (Geissel, 2009). However, during the last years the number increased and it can be expected that more than 300 municipalities in Germany are dealing with the introduction or have already introduced this innovation (bpb & SKEW, 2015). Similar developments could be detected in other European countries (Röcke, 2014). Hence, this form of budgeting spread especially during the last 10 years among municipalities in Europe. Even though the number of municipalities with participatory budgeting increased, the majority has still not implemented this type of innovation yet.

One reason for the different distribution of participatory methods in the budgeting process of municipalities could be related to organisational differences. As stated above, scholars observed that more complex structures in public organisations are positively related to the occurrence of innovation (Damanpour, 1996). This study assumes that a public organisation, which is more complex, will more likely adopt an innovative budget process. As a result, this research will look specifically at the relationship between organisational complexity and the occurrence of innovation in the drafting of a municipal budget.

(6)

5

1.1 Problem definition and research question

It is easy to claim that the public sector must become more innovative. Since innovation is the outcome, scholars and practitioners are trying to find means to create innovation. In the field of the innovation research, it is recognized that there are circumstances, which promote the occurrence of innovation and others, which constrain innovation (Damanpour, 1991). Public administration scholars try to test these findings in a public context (De Vries et al., 2016).

There are several organisational variables that might be related to the occurrence of innovation. One way to examine organisational structure is to focus on centralisation, formalisation and specialisation (Jakobsen & Thrane, 2016). Out of the aforementioned variables, Jakobsen and Thrane (2016) state that from the less existing empirical evidence, it can be concluded that only centralisation and specialisation are related to the occurrence of innovation. In addition, it could also be argued that there is a relationship between the organisational complexity and the occurrence of an innovation (Damanpour, 1996). The measurement of organisational complexity is broader, as it does not exclusively focus at one specific parameter. Due to empirical findings, it can be expected that the complexity of an organisation has a positive relationship with innovation (Damanpour, 1996).

In recent years, participatory innovation, where both citizens and civil servants contribute to the fulfilment of public tasks, became more popular and spread in European municipalities (Geissel & Newton, 2012) One example of this innovation is participatory budgeting and gives citizens the possibility to have a voice in the creation of a municipal budget (Ebdon & Franklin, 2006; Hong, 2015; Rossmann & Shanahan, 2011). However, some municipalities use this innovation, others do not. At the present time, it is not fully discovered why there are differences in the usage of participatory models (Geissel & Newton, 2012). Applying the existing research on innovation on this specific type, the organisational structure might influence the usage of participatory elements in the budget process. The focus on budgeting is especially interesting, as the distribution of resources is one of the most powerful tools of the public sector (Wildavsky, 1961) and the drafting of the budget is the core responsibility of the municipality. This research aims to discover the relationship of organisational structure and the occurrence of participatory innovation in the budgeting process among municipalities.

From the existing research, it could be expected that a more complex organisational structure in a municipality leads to a higher likelihood of the adaption of participatory budgeting. Complex organisations are often larger and have more resources that can be used for the creation and adoption of innovation (Damanpour, 1992, 1996). Moreover, they tend to be more differentiated, which promotes expert knowledge and cross-fertilisation among experts (Aiken, Bacharach, & French, 1980; Damanpour, 1996). Empirical findings from Damanpour (1996) indicate a positive relationship of

(7)

6 organisational complexity and innovation. However, there is only few empirical research about this relationship in a public context and further investigations are needed in order to support this argument. Hence, this study will focus on organisational complexity as an explanatory variable for the divergent usage of participatory elements in the budgeting process by examining the following key research question:

“To what extent does organisational complexity affect the use of participatory innovation in the drafting of a municipal budget”?

1.2 Academic and social relevance

The content of this thesis can enlarge the existing research in the field of public administration. In general, the innovation research in a public sector context is relatively new and not fully developed yet (Torfing & Triantafillou, 2016). This is particularly true for participatory innovations, which became more popular in the last years and are not comprehensively examined (Pateman, 2012).

In addition, this thesis contributes to the existing research, as it will be designed differently compared to the majority of other studies. Public administration scholars focus on normative or conceptual arguments to explain the occurrence of innovation in the public sector (De Vries et al., 2016). More empirical research is needed to test these assumptions. The little existing empirical research is mostly qualitative and, therefore, De Vries et al. (2016) emphasized in a recent meta-analysis the need for quantitative studies in the field of innovation in the public sector. As a result, this research will follow an empirical large-N design to collect and analyse quantitative data.

For the society and practitioners in municipalities innovativeness is also relevant. Being innovative is not an end in itself, but can help to achieve specific organisational objectives or promote societal goals (Bekkers et al., 2011; Torfing & Triantafillou, 2016). With a view to the participatory innovation, these goals could be to create legitimacy, more democracy or even to save financial resources (Geissel, 2009). A better understanding of the interaction of organisational structures and the occurrence of participatory innovation is beneficial for practitioners to make a decision about their organisational design. While oversight agencies claim for less complex organisational structures to achieve goals of parsimony (GPANRW, 2014a), academic findings point in the direction that complex structure have beneficial outcomes such as innovativeness (Damanpour, 1996). Outcomes of this study could advise municipalities to structure their organisation carefully in line with their context and their goals. If this study finds evidence for a positive relationship of organisational complexity and innovation, managers in municipalities would have arguments for keeping complex structures rather then follow blindly the claim of oversight agencies for implementing simple structures.

(8)

7

1.3 Structure of the thesis

After the introduction follows a theory part in Chapter 2. The goal is to provide an overview of the existing literature regarding the research question, to set a conceptual framework with definitions and to state hypotheses. Chapter 3 explains the used research design and the data collection strategy. The dependent variable and independent variables will be operationalized. In the results part, Chapter 4, the collected data will be described and analysed. In the final Chapter 5, results of this study will be discussed and interpreted. The thesis ends with concluding remarks on its academic and practical implications and offers a possible avenue for future research.

(9)

8

2. Theory

This chapter will set a theoretical foundation. It starts with research findings on innovation in the public sector and a distinction of innovation in private and public organisations. In addition, a definition of innovation and the description of different types of innovation form part of this theory chapter. Participatory innovation is the dependent variable and will be discussed as a specific form of public sector innovation. Afterwards, organisational structure will be described in general and referring to the occurrence of innovation. To narrow the focus, organisational complexity with its variables functional differentiation and organisational size will be defined and explained as specific parameters of organisational structure.

2.1 Innovation in the public sector

Both public and private sector are dealing with innovations, but are driven by different motives. Companies aim to be innovative to increase their competitivity and their chances of higher earnings due to new products or services. Even though the public sector is not facing the same competitive context, there are rising expectations from the outside to have an innovative public sector (Lewis et al., 2014). Just like in private organisations, innovations in the public sector are seen as a way to deliver better services (Walker, 2006). In contrast to the private sector, the public sector needs innovation to solve major problems that are relevant for the future of the society (Bekkers et al., 2011; Torfing & Triantafillou, 2016). Both perspectives, either private or public, may lead to the same outcome: a new process or a new service. However, they differ in their underlying motivation. While the private sector is driven to make benefits, the public sector uses innovation to improve their services to be prepared for future challenges and to have a connection to its environment (Bekkers et al., 2011).

It is often claimed that the public sector is not innovative (Torfing & Triantafillou, 2016). However, this statement cannot be confirmed. The statement has its roots in perceptions that people have about public organisations (Raadschelders & Vigoda-Gadot, 2015). Public organisations are often recognized as rigid organisations, which might harm innovativeness (Rainey, 2014). Furthermore, the democratic process with the involvement of many people and the goal of reaching consensus could also hamper the upcoming of innovations in the public context (Bekkers et al., 2011). Despite this, there are also reasons why the public sector could be innovative. One main argument for this statement is that the public sector is actually innovative in practice (De Vries et al., 2016; Mazzucato, 2013; Walker, 2014). The public sector is capable to be innovative due to its internal diversity and its complex environment with divergent goals (Bekkers et al., 2011). Research has shown that complexity in an organisation can actually foster innovation, as it creates space for expert knowledge and collaboration (Hage, 1999). In recent years, there is an increasing academic interest in studying public sector innovation and from

(10)

9 the existing research, it can be concluded that the public sector is a place where innovation can occur (Bekkers et al., 2011; De Vries et al., 2016).

At this stage, we have to clarify what innovation actually is. Whereas the term “innovation” is quite often used in our daily lives, most of the scientific papers are not defining innovation and if they do, the definition is mostly general (De Vries et al., 2016). In their meta-analysis, De Vries et al. (2016) also conclude that the definitions of innovation mostly consists, first, of a novelty and, second, of the adaption of this novelty. This can be exemplified by Walker (2006, p. 592), who defines innovation as “a process through which new ideas, objects and practices are created, developed or reinvented and which are new and novel to the unit of adoption.” This definition will be used for this thesis, as it is in line with the contemporary research by using the terms novel and adaption.

While the definition of innovation might be more clarified, the concept of innovation still remains abstract and needs to be narrowed down in categories. The definitions of innovation are quite broad due to the fact that innovation can have many forms. There are three main types of innovation: process innovation, product or service innovation and ancillary innovation. Process innovations aim to change the management or organisation of processes (Bekkers et al., 2011). Often they are subdivided in administrative process innovations or technological process innovations (Walker, 2006). The second type of innovations are product or service innovations. This type aims to create new products or services (Bekkers et al., 2011). A third type are ancillary innovations. According to Walker (2006), these are forms of collaboration, where the success depends on individuals or organisations from outside the organisation.

Participatory innovation in the budgeting process, the focus of this research, can be classified as an ancillary innovation. For the success of a participatory budgeting, it is a requirement that people from the outside participate (Röcke, 2014). It has to be mentioned that ancillary innovation can also have other descriptions. For instance, this type can also be linked to open innovation, which implies collaboration with actors from the outside of an organisation and has its roots in the private sector (Chesbrough, 2006; Mergel & Desouza, 2013). Scholars, who focus on participatory budgeting, label ancillary innovation as participatory innovation, which has a similar meaning, but the aim is to increase legitimacy by involving citizens (Geissel, 2009; Röcke, 2014). In particular, participatory innovation can be seen as a subtype of ancillary innovation, as it fits under the broader term. In the following chapters, this research will link participatory budgeting to participatory innovation, as it describes more precise that not only people or organisations from the outside, but citizens are needed for this type of innovation.

(11)

10

2.2 Participative elements in the budgeting process

Budgeting is a core task of public policy and determines the distribution of resources, priorities and power. For decades, one main part of public policy is the distribution of resources (Rubin, 2017). Incoming taxes and revenues are redistributed by politicians and civil servants to specific tasks. Moreover, budgeting indicates which tasks will be done by the government and also which priority these tasks have (Rubin, 2017). A budget is needed to fulfil a specific public task. Problematic is that the resources of the public sector are on the one hand limited, but on the other hand the public sector faces increasing tasks and problems (Lewis et al., 2014). Having influence on the public budget decision-making process, demonstrate some sort of power. Hence, it is not a surprise that the interplay between politicians and civil servants in budgeting stayed mainly untouched in the majority of public organisations (Sintomer, Herzberg, & Röcke, 2008).

Both traditional budgeting and participatory budgeting follow a budget cycle. This cycle consists basically of four parts: drafting, legislation, execution and auditing (Lee et al., 2013). In traditional systems, the draft is done by the municipality, the legislation by the elected legislative body, the execution through the municipality and in the auditing stage, the budget will be checked by regulatory authorities and afterwards published (Lee et al., 2013). In traditional budget systems, the only part where citizens have to be involved, is the publication of the budget decision (Lee et al., 2013). However, this involvement is more a providing of final results than actual citizens’ participation. Table 1 illustrates the budget cycle in the traditional and the participative system.

Table 1: Comparison of traditional and participatory budgeting

Budget cycle stages Traditional budgeting Participatory budgeting Drafting Municipality creates the draft. Municipality creates the draft, but

citizens have to be consulted. Legislation Legislative body decides on the

budget draft.

Legislative body decides on the budget draft.

Execution Municipality executes the draft. Municipality executes the draft.

Auditing

Publication of the budget plan. Regulatory authorities check the budget plan.

Publication of the budget plan. Regulatory authorities check the budget plan. Citizens receive

feedback what happened with their proposals.

In comparison to the traditional budgeting, the additional involvement of citizens makes participatory budgeting an innovation. In recent years, the traditional budget process became more diverse due to the adoption of participatory elements (Ebdon & Franklin, 2006). Even though the phases of the budget cycle remain the same, the introduction of participatory budgeting increases the number of

(12)

11 participants in this process and gives them more influence. Thus, participatory budgeting can be defined as the involvement of citizens in the decision-making process of the public budget (Geissel, 2009; Hong, 2015; Rossmann & Shanahan, 2011). In German municipalities, this involvement takes place mainly in the drafting phase, whereby the citizens are able to make proposals for additional spending or saving potentials (Geissel, 2009). It can be concluded that participatory budgeting is a participatory innovation, as it is crucial that citizens are involved in this process.

Although participatory budgeting is not new, it is still innovative for Western public organisations and academics (Michels, 2011). The current research focuses more on the outcomes of participatory budgeting for the early adopters and examines it from normative viewpoints (e.g. Baiocchi & Ganuza, 2014; Michels, 2011; Rossmann & Shanahan, 2011). Hence, slowly adopting countries like Germany are missing. Their participatory budgeting development is still in the adoption stage and their usage of participatory approaches might be different compared to the early examples (Röcke, 2014) Hence, it is unclear what drives participatory budgeting in countries that adopted participatory budgeting recently. This research aims to discover this gap.

2.3 Organisational structure and its relationship with innovation

Organisational structure is one of several conditions, which might influence organisational performance and in particular innovations. Another attempt of explanation for the occurrence of innovation is related to characteristics of the environment. The environment consist of different forces (e.g., political, economic, technological), which affect the way the organisation works and regulate the access to additional resources (Ebdon & Franklin, 2006; Jones, 2007; O’Toole & Meier, 2015). Other explanations focus on internal factors, like leadership or goal-setting (Ebdon & Franklin, 2006; Jaskyte, 2011; O’Toole & Meier, 2015). However, there is also a group of scholars, who consider organisational structure as an explanation for the occurrence of innovation (Burns & Stalker, 1961; Damanpour, 1991; Walker, 2008; Zornoza et al., 2007). This research will have a closer look at organisational structure and its influence on innovation.

The structure of an organisation is a way to manage an organisation and it has an influence on performance. Mintzberg (1979) defines organisational structure as “the sum total of ways in which it [i.e. the organisation] divides its labor into distinct tasks and then achieves coordination among them” (p. 2). The traditional public administration was structured based on Weber’s bureaucracy: a stable organisation with clear rules to achieve efficiency (Pollitt & Bouckaert, 2011). Burns and Stalker (1961) did not follow this ideal type approach, but argued that the organisational structure is dependent on contingency factors. They classify organisations as mechanistic or organic. Mechanistic organisations consist of a more stable structure, while organic organisations are built of a looser structure (Burns & Stalker, 1961). In a mechanistic organisation, there is a clear chain of command, specific job

(13)

12 descriptions and communication among equals (Burns & Stalker, 1961). Hence, there is less collaboration and decisions are mostly taken on the top levels of the hierarchy. In comparison, organic organisations resemble a network, containing less hierarchy and more collaboration (Burns & Stalker, 1961). Employees are able to participate in decisions and can also exchange knowledge. As a result, organic organisations tend to be more innovative and capable for change than mechanistic organisations (Burns & Stalker, 1961).

In more recent research, scholars focused on specific parameters of organisational structure. This research follows the approach of current public administration meta-analyses (e.g. De Vries et al., 2016; Jakobsen & Thrane, 2016) that build their arguments on organisational structure mainly on management research conducted in the private sector. In the literature, the specific variables of organisational structure are predominantly the following: specialisation, formalisation and centralisation (Damanpour, 1991; Walker, 2008; Zornoza et al., 2007). Specialisation reflects on the amount of specific knowledge and skills, which are needed in the organisation (Damanpour, 1991). Formalisation states the degree of regulation by written rules that activities in the organisation undergo (Zornoza et al., 2007). Centralisation expresses to which extent decisions are made on the top levels in the hierarchy (Jakobsen & Thrane, 2016). It can be expected that specialisation, formalisation and centralisation are related to the occurrence of innovation.

Specialisation is expected to be positively related to innovation. Specialisation allows employees to focus on a specific field or task and enables them to get expert knowledge (Jones, 2007). It can be assumed that employees with a more narrowed focus, can also be more innovative, as they gain specific knowledge about their task area (Aiken & Hage, 1971). Moreover, if these experts work together with other experts, they obtain different perspectives and are able to come up with innovative ideas (Hage, 1999). Formalisation might have a negative relationship with innovation. If the job role is clearly defined, it will provide employees with less leeway for the execution of the task and leads to increased routine rather than innovation (Burns & Stalker, 1961). Another argument for the negative influence of formalisation is that employees are more focused on following all the given rules than on being creative if the degree of formalisation is high (Hage & Aiken, 1970). Centralisation is expected to have a negative influence on innovation. In a centralized organisation, members of the organisation are mainly not part in the decision making, as it is a task of the top management. However, a more decentralized and inclusive structure allows employees to be informed and inspires them to come up with new ideas during the decision making process (Damanpour, 1991).

There are also other approaches to examine the organisational structure. Several studies investigate the relationship between organisational complexity and innovation, as they argue that these variables are positively related (Aiken & Hage, 1971; Damanpour, 1991, 1996). Complex organisations consist of

(14)

13 two parameters: a complex structure and a large size (Damanpour, 1996). Complex structures include many and different departments, where expert knowledge can emerge (Damanpour, 1991). In the same way as specialisation, these experts are more likely to find solutions for problems in their narrowed field (Aiken & Hage, 1971). Moreover, if the different departments work together, these experts enlarge their own knowledge with different perspectives from other functional departments, which leads to a cross-fertilisation (Aiken & Hage, 1971). Organisational size is the other parameter of organisational complexity and can consist of the budget or the number of personnel (Aiken & Hage, 1971). Organisational size as a specific parameter can have a positive or negative relationship with innovation (Camisón-Zornoza, Lapiedra-Alcamí, Segarra-Ciprés, & Boronat-Navarro, 2004; Damanpour, 1992). It can be argued that size is positively related with innovation, since large organisations have more resources and through a larger number of employees they also have more expertise (Damanpour, 1996). On the other hand, large organisations tend to be more formalised and less flexible for adopting new ideas (Damanpour, 1996).

We have expressed several approaches to conduct organisational structure and in the next sub-chapter, we will describe why we choose to examine organisational complexity and explain its concept in more detail.

2.4 Organisational complexity and its influence on innovation

In this research, we will focus on organisational complexity. There are three reasons for this decision. Firstly, it is expected that municipalities vary in their complexity. They might have a different way how they differentiate their units and they also vary in their size. As a result, there might be enough variation among these indicators to explain differences in the adoption of participatory innovation. Secondly, the relationship between innovation and organisational parameters like formalisation, centralisation and specialisation has already been examined in the private and public sector. The meta-analyses from Damanpour (1991) and Jakobsen and Thrane (2016) express that this research field has already been covered. Through the focus on organisational complexity, this study could give a different perspective on organisational structures in the field of public administration research that has not been extensively examined. Thirdly, information about organisational complexity can be gained through the examination of more objective information. While it is common to measure formalisation, centralisation and specialisation through perceptions, organisational complexity can be analysed by investigating objective data like organisational charts or statistics (Pennings, 1973). Thus, it is assumed this objective measurement of organisational structure can contribute to the existing research body. Organisational complexity is a concept which is more abstract. It has been already briefly described that organisational complexity consists of mainly two sub-variables: structural complexity and size (Damanpour, 1996). Structural complexity can have different meanings, but it is mostly related to the

(15)

14 degree of differentiation on several levels (Damanpour, 1996). Organisations can be differentiated in a spatial, occupational, hierarchical and functional way (M. P. Blau, 1970). Spatial differentiation is the physical separation of units. Occupational differentiation is closely linked to division of labour and counts the amount of specialist in an organisation (Damanpour, 1996). Hierarchical differentiation measures the vertical levels of an organisation (Damanpour, 1991), while functional differentiation expresses the number of departments on a horizontal level (Aiken et al., 1980). This research will focus on functional differentiation. Previous studies have already been argued that functional differentiation is appropriate to represent complexity and to predict innovation, whereas spatial and hierarchical differentiation were less suitable (Damanpour, 1996). Occupational differentiation will be excluded, as we assume that municipalities will not vary enough on occupational job descriptions. Municipalities have similar tasks and civil servants in municipalities follow usually a comparable formal education. As a result, there would be less variation among this variable, which makes it an inappropriate indicator for innovation. However, municipalities have discretion on how to structure the organisation in a horizontal way and, hence, in this research, structural complexity will be recognized as functional differentiation.

It can be expected that functional differentiation is positively related to the occurrence of innovation. The more an organisation is functionally differentiated, the more departments with special tasks it has. As a results, in these departments, employees with specific knowledge emerge (Aiken & Hage, 1971; Baldridge & Burnham, 1975). This step alone leads already to more innovation. Due to their narrowed focus, these experts can come up with new ideas in their specific field (Aiken & Hage, 1971). Plus, in these departments, collations of experts are created and they can learn from their colleagues and cross-fertilize each other (Aiken & Hage, 1971). Nevertheless, an organisation does not only consist of separated units of experts. The diverse departments have to work together to achieve the goals of the organisation. Problematic is that due to the differentiation and its diversity, departments might have conflicts about goals or resources (Baldridge & Burnham, 1975). Hence, the departments have to work together to overcome this complexity (Hage, 1999). Due to this collaboration, different departments gain additional viewpoints of the other experts, which can lead to a cross-fertilisation between the departments (Baldridge & Burnham, 1975). In their meta-analyses, Damanpour (1991) and Zornoza (2007) found empirical evidence for a positive relationship between functional differentiation and the occurrence of innovation. In addition, Damanpour (1987) tested how functional differentiation is related to several types of innovations and found the strongest positive relationship between functional differentiation and ancillary innovation, which contains participatory innovation. From the theoretical reasoning and empirical findings, the following hypothesis can be formulated:

(16)

15 The second variable of organisational complexity is organisational size, which is also assumed to be related to innovation. As stated earlier, the size of an organisation consists mostly of the number of employees or the size of the budget of an organisation (Aiken & Hage, 1971). In the literature, there are mixed results for the direction of the size-innovation relationship (Camisón-Zornoza et al., 2004; Damanpour, 1992).

There is a group of scholars, who state that organisational size has a positive relationship with innovation. Larger organisations tend to have complex and various resources, especially with a view on its personnel (Nord & Tucker, 1987). Thus, they have a larger pool of potential expertise and different represented viewpoints. In particular, these additional personnel resources consist of staff with technical know-how and it is more likely that an organisation adopts technical or technically-supported innovation (Damanpour, 1992). In addition, due to its size and the additional resources, large organisations can take more risks (Damanpour, 1992). The failed adoption of an innovation hurts a large organisation with many resources less than a small organisation with a tight budget.

However, other arguments point in the direction that organisational size has a negative influence on the occurrence and adaption of innovations. One main argument is that smaller organisations are more flexible than large organisations (Damanpour, 1996). Through this flexibility, it is easier for them to manage change and adopt innovation. In contrast, large organisations are often described as bureaucratic and rigid, which might even harm the innovation. The second main argument is that more formalisation is needed to control a large organisation (Hitt, Hoskisson, & Ireland, 1990). We have already explained previously that formalisation is expected to have a negative effect on innovation (Burns & Stalker, 1961; Damanpour, 1991). Less flexibility and a higher formalisation of large organisations leads to less leeway of employees to bring in ideas and it might be more complicated to adopt innovation compared to smaller organisations.

In this research, it is expected that organisational size is positively related to innovation. The main reason for this is that empirical evidence from two meta-analyses support a positive relationship between organisational size and innovation (Camisón-Zornoza et al., 2004; Damanpour, 1992). In addition, Damanpour (1987) found that organisational size is especially positively related to ancillary innovation. It can be concluded that empirical evidence points in the direction that organisational size might have a positive relationship with the occurrence of participatory budgeting. Given these argumentations, the following hypothesis will be tested:

H2: Organisational size has a positive influence on the occurrence of participatory innovation. In the next chapter, the methodology of this study will be illustrated.

(17)

16

3. Research Design and Data Collection

In this chapter, the methodology of this thesis will be explained. Firstly, the research design and the case selection will be described. Afterwards, the methods for data collection of this research will be discussed. This is followed by the operationalisation of the used variables. Thereafter, this research will be evaluated in terms of reliability and validity. Lastly, the analysis strategy will be explained.

3.1 Research design and case selection

This research consists of a large-N design. It is appropriate to detect general relationships of variables by using large numbers and statistical methods (Toshkov, 2016). The goal of this research is to investigate if there is a relationship between organisational variables and the occurrence of innovation. Moreover, this research will test assumptions from organisational theory and apply them to a specific type of innovation. Large-N studies are suitable for testing and applying theory, rather than developing new ones (Toshkov, 2016). This research design does not only fit due to the general framework of this research, but also aligns with the research gap. As research about innovations in the public sector is dominated by qualitative studies (De Vries et al., 2016), this quantitative approach is able to enlarge the existing research.

The cases for this study are municipalities. By exclusively looking at municipalities, we ensure that the investigated organisations have an equal mission and execute similar tasks. Furthermore, participatory budgeting is not common on a federal or state level, which leaves only municipalities as a potential case. In this study, municipalities in the German state North Rhine-Westphalia will be investigated. This selection has three reasons. Firstly, we focus on Germany, as participatory budgeting is not common there yet and can still be seen as an innovation. Secondly, we narrow the focus down on one state, as not all of the states have significant numbers to test our assumptions. However, in North Rhine-Westphalia, 27 % of the municipalities have some experience with participatory budgeting. From this follows that there are enough municipalities with participatory budgeting to analyse in a large-N design. Thirdly, it is also expected that municipalities in North Rhine-Westphalia differ among the explanatory variables, which allows to test our assumptions in this population. North Rhine-Westphalia consists of 396 municipalities and all of them will be tested.

3.2 Methods for data collection

The data will be collected by using existing statistics and by the analysis of documents. Data regarding the dependent variable, the occurrence of participatory budgeting, will be exclusively gathered by using an existing statistic. In a joint project, Bundeszentrale für politische Bildung (bpb) and Servicestelle Kommunen in der Einen Welt (SKEW) offer online on buergerhaushalt.org statistics about the usage of participatory budgeting in German municipalities. Bundeszentrale für politische Bildung is

(18)

17 an organisation of the Federal Ministry of the Interior (Bundesministerium des Innern, n.d.) and Servicestelle Kommunen in der Einen Welt was founded by the Federal Ministry of Economic Cooperation and Development (SKEW, n.d.). Hence, both are relatively independent public organisations and we can assume that they offer reliable and transparent data. The explanatory variables will be collected by analysing documents and by using data from statistical agencies. For gathering the organisational variables, the organisational charts of municipalities will be analysed. In many cases they might be available online. However, if the organisational chart is not accessible on the homepage, the human resource or organisational department will be contacted in order to receive the information. Assumed controlling variables, such as slack resources, number of citizens, political support and citizen characteristics will be built on data of the statistical office of the state North Rhine-Westphalia and the Federal Statistical Office.

3.3 Operationalisation of variables

In this section, the variables will be operationalised. The outcome variable is participatory budgeting. In this research, we make use of the study from buergerhaushalt.org and follow their classification. They checked the webpages and asked municipalities if they had experiences with participatory budgeting and put them in specific classifications that reflect on the stage of participatory budgeting (bpb & SKEW, 2015). We will code the outcome variable binary. 1 means that there exists participatory budgeting and 0 illustrates that there is not yet participatory budgeting present. The existence of participatory budgeting consists of the classifications: decision, early form, introduction, continuation and Abstellgleis from the categorization of buergerhaushalt.org.

Firstly, participatory budgeting is present if the municipality decided to implement participatory budgeting (i.e., decision phase). The stage decision is given if participatory budgeting is decided by the municipal council (bpb & SKEW, 2015). The decision is not older than two years, but participatory budgeting has not been introduced yet (bpb & SKEW, 2015). Secondly, if an early form of participatory budgeting can be detected, it will be counted as existing participatory budgeting (i.e., early form phase). The municipality claims that it is using participatory budgeting, but there are no formal interactive participatory processes (bpb & SKEW, 2015). An example is that citizens can send their ideas to the municipality via email or through an online form, but there is no public discussion about these ideas (bpb & SKEW, 2015). Thirdly, the introduction of participatory budgeting counts as the present of this innovation (i.e., introduction phase). For the introduction stage it is necessary that citizens are informed about the budget and that there is a public consultation (bpb & SKEW, 2015). Citizens can evaluate proposals from the municipality, but they also can bring in own proposals and afterwards the municipality has to inform the public what they have done with the citizens’ proposals (bpb & SKEW, 2015). Fourthly, if the municipality uses participatory budgeting continuously, participatory budgeting

(19)

18 is present (i.e., continuation phase). A continuation stage is present if participatory budgeting was used for the third time or more often (bpb & SKEW, 2015). Lastly, if participatory budgeting has been used in the past, but is no longer applied, it is also counted as the presence of this innovation and is labelled as Abstellgleis (bpb & SKEW, 2015). Abstellgleis is a German phrase and means to put something by the side, you already used or had contact with. It might be discussable to code Abstellgleis as the existence of the innovation. However, in this research it will be coded as the presence of participatory budgeting, as those municipalities have been innovative in the past. The actors in the municipality discussed this topic, decided on it and implemented it. In addition, following the definition of innovation, it is about the novelty and adoption of ideas. Even in the case where participatory budgeting is no longer used, they still had adopted this type of innovation. Moreover, this study is about the occurrence of innovation and not about the evaluation of its effectivity or sustainability. Hence, Abstellgleis will be classified as the occurrence of participatory budgeting.

Two classifications of buergerhaushalt.org will be summarized as the absence of participatory budgeting: discussion and no status. Discussion means that the introduction of participatory budgeting is discussed within the municipality by a political party or non-governmental organisation (bpb & SKEW, 2015). The claim is not older than two years and there was no decision by the municipal council so far (bpb & SKEW, 2015). As it is only discussed, the innovation is not yet adapted and cannot be counted as present. The other absence indicator is no status. No status means that the introduction of participatory budgeting is not an issue in this municipality and there is no participatory budgeting observable (bpb & SKEW, 2015). The complete coding is illustrated in Table 2.

Table 2: Coding of the dependent variable

Participatory budgeting Status Explanation

Yes (1) Decision Council decided to start participatory

budgeting (PB), but it is not introduced yet. Early form Not a fully interactive form of PB, as it lacks

formal consultation. Introduction PB has been introduced.

Continuation PB has been introduced and pursued. Abstellgleis PB has been used, but is no longer applied No (0) Discussion A political party or NGO brought in an

initiative about PB in the municipal council. No status No PB is applied.

(20)

19 usually measured by the total number of departments under the top management (Aiken et al., 1980; J. R. Blau & McKinley, 1979; Damanpour, 1991). This study will follow the operationalisation of Aiken et al. (1980) where all departments that report directly to the major or an alderman (i.e. political level) and consist of more than two persons were measured. To make this operationalisation more precise, we add that subunits will not be counted, as they are not independent and belong to another unit. It might be problematic to distinguish between units and subunits. Hence, we will focus only on the hierarchical level under the mayor to avoid including subunits. In addition, departments with no function will not be counted. For instance, some municipalities show in their hierarchical chart names of persons as a unit without a functional description. Often these persons are aldermen, who have a leading position in a municipality and are the head of political portfolios. As we are interested in public administration, political portfolios will not be counted. To sum up, the independent units in the hierarchical level under the mayor with a function and more than two employees will be counted. In the appendix, some examples of organisational charts illustrate the operationalisation and measurement of functional differentiation.

The second explanatory variable is organisational size. Size can be measured in different ways, which can lead to different results (Kimberly, 1976). It can be operationalised by the number of employees, the budget, input and output volume or physical capacity (Damanpour, 1996). In this research, organisational size will be measured by the number of employees. Physical capacity and the input and output volume (i.e., number of clients) might not be appropriate measurements for municipalities, as they are more related to private companies. The number of employees seems to be a reliable and valid measurement, as municipalities are obligated to send their personnel statistics to a statistical state office. This process ensures some quality and comparability of the dataset. From these statistics, the full-time equivalent employment will be used to illustrate the number of employees.

It is assumed that slack resources, political support, size of the population and characteristics of the population (i.e., income and political activity) also have an influence on the occurrence of innovation. As a result, we have to control for these variables to obtain valid findings.

The slack resources might be a variable, which has an influence on the occurrence of innovation. It has been argued that organisations that have additional resources available are more likely to be innovative (Aiken & Hage, 1971; Damanpour, 1991). They have more financial resources to invest in ideas and can take the risk of failure, as they have a greater financial leeway compared to less solvent organisations (Damanpour, 1991). In addition, due to its financial position, organisations with higher slack resources can also adopt ideas in advance (Damanpour, 1991). As a result, it seems appropriate to control for slack resources. Aiken and Hage (1971) measured slack resources with a focus on the budget and earnings of an organisation. In this research, we will measure slack resources negatively.

(21)

20 We assume that municipalities with a high debt, have less slack resources. Information und Technik Nordrhein-Westfalen (IT.NRW) offers data about the debts of municipalities. The debt per citizens ratio of the municipality will be used to measure negatively slack resources.

Previous studies have also illustrated the importance of political support for participatory innovations. Röcke (2014) argues that the political support for this innovation influences the adaption and success of participatory budgeting due to the dominant position of politicians in the policy system. Through the separation of power, the main decision-making responsibility lies in legislative bodies. Ebdon and Franklin (2006) state that participatory budgeting is about the collaboration of politicians and citizens and both actors need some kind of a supportive attitude. These theoretical argumentations were strengthened by empirical findings that support the relationship of political support on the occurrence of participatory budgeting (Brun-Martos & Lapsley, 2017; Röcke, 2014). It might not be feasible to measure individual political support. Hence, this research will control for the involved political parties in the municipality. In general, the Christlich Demokratische Union (CDU) is the only major German party, which is not in favour of direct democracy (Mehr Demokratie e.V., 2017). The reason could be that this party is classified as conservative. It can be assumed that the more votes the CDU gets for the municipal legislation body, the stronger their legislative influence is and as a result the less likely is participatory budgeting. Political support will be measured negatively by the percentage of votes for the CDU in a municipal legislation body. The data is provided by the office of the Federal Election Commissioner.

Another control variable is the size of the population. The size of the population could have an effect on the organisational structure of the municipality and on the adaption of participatory budgeting (Andrews, Boyne, Meier, O’Toole, & Walker, 2011; Ebdon & Franklin, 2006). The tasks for a municipality might be more complex if the number of clients is high. For instance, it makes a difference to organise the waste disposal for 20.000 or 1.000.000 citizen. This means that municipalities might design their organisational structure in line with the number of clients to fulfil their tasks. Plus, it can also be expected that a large number of citizens also leads to a larger diversity of demands. Contingency models have already been discussed in this study and as a result it can be assumed that the size of the environment is related to organisational variables. Furthermore, the size of the population might have a direct relationship with participatory innovation. Ebdon and Franklin (2006) argue that participation is more common in larger cities. The reason for this is that larger cities have also a more diverse population and the municipality wants to mediate conflicting claims by allowing participation (Protasel, 1988). Moreover, citizens of larger cities have been found to steer for more political participation (Nalbandian, 1991). To overcome different demands and to satisfy the additional need for participation, larger municipalities are more likely to introduce participatory budgeting. This

(22)

21 variable will be measured by the number of citizens that have a primary or secondary residence in the municipality. The data is provided from the Federal Statistical Office.

The socio-economic environment can also influence the occurrence of innovation. The income of citizens might have an effect whether people want to engage in participatory models (Irvin & Stansbury, 2004). For people with higher incomes it might be easier to attend to meetings for the budgeting, while low income families focus more on survival and care of the family (Irvin & Stansbury, 2004). In an environment with high income, people might be more willing to engage in participatory budgeting and also claim for more participation. Income of citizens will be measured by the average available income that citizens can use for saving or consumption, which is collected by the statistical office of North Rhine-Westphalia. Another socio-economic variable is political activity. Citizens, who are more active in politics, might be steering for more participatory models (Geissel, 2009). The involvement of citizens in politics can be operationalised by the voter turnout. It is assumed that municipalities with a higher voter turnout, also have a higher politically active citizenry. In this research, the political activity will be measured by the voter turnout for municipal elections. We use statistics from the Federal Election Commissioner for this variable.

Other typical variables that might influence the adaption of innovations are the type of the organisation or the legal environment (Damanpour, 1991; Ebdon & Franklin, 2006). Due to the focus on municipalities in one state of one country, the type of the organisation and the legal environment remain stable and must not be included. Table 3 illustrates all used variables.

Table 3: Summary of the variables

Variable Source Description of the variable

Participatory budgeting Buergerhaushalt.org Participatory budgeting map 2017

Municipalities that are going to introduce or have introduced participatory budgeting Functional differentiation Organisational charts of

municipalities in North Rhine-Westphalia (from their homepages; collected in November 2017)

The sum of administrative units that consist of more than two people and are in the hierarchy level under the mayor

Organisational size IT.NRW

Personnel statistics 2016

Full-time equivalent of employees

(23)

22

Slack resources IT.NRW

Debt statistics municipalities 2016

Debt ratio in Euro (public debt/population)

Political support Federal Election Commissioner Official election results

2014

Share of votes for the CDU in percentage for the municipal elections

Size of the population Federal Statistical Office Administration structure 2015

Number of citizens with primary or secondary residence

Wealth of the citizenry IT.NRW

Income statistics 2015

Average available income in Euro of a citizen in a

municipality Political activity of citizenry Federal Election Commissioner

Official election results 2014

Voter turnout in percent from municipal elections

3.4 Reliability and validity

This research design and the operationalisation of the variables has strengths and weaknesses regarding reliability and validity. This research is expected to be reliable, as it follows clear concepts and definitions from previous studies (Neuman, 2014). All variables, except of functional differentiation, are measured by using existing datasets from independent public organisations. These datasets are transparent and accessible for everyone, which simplifies the replicability of this study. Only the values for functional differentiation are not available and must be collected. However, for this data only official and accessible documents will be analysed. The organisational charts can be found on the homepages of the municipalities, which makes this measurement more reliable. A weakness regarding the reliability is related to the measurement of the organisational charts, as the versions we studied, may not be archived on the homepage. This might complicate the repetition of this research. Internal and external validity is also given in this research. It can be assumed that one strength of this research is the internal validity for three main reasons. Firstly, this research is expected to have face validity. Face validity is present if researchers accept this measurement as appropriate to measure the construct (Neuman, 2014). Due to the application of existing concepts and operationalisations from recognized scholars, it can be assumed that this research has face validity. Secondly, content validity is another advantage of this research. Content validity expresses to which extent the measurement covers all aspects of the definition (Neuman, 2014). Especially the core variables have this content

(24)

23 validity, as their measurements contain all parts of the definitions. Thirdly, the usage of objective measurements increases the internal validity. Some studies of organisational structure exclusively rely on a questionnaire approach, which is based on perceptions of individuals (Pennings, 1973). In this study, all the variables are objectively measured, which avoids potential biases that results from the usage of subjective measurements. One limitation regarding the internal validity is that not all indicators are congruent from the year of observation. However, most of the available data is recent. In addition, we assume that the variables are relatively stable over time. Another weakness is that some organisational charts were published without a creation date and it is not clear if this is the actual structure, which is used in reality. It could be that organisations do not keep their homepage up to date and the gathered data would lack internal validity. External validity is another part of validity, which expresses if the results can be generalized (Neuman, 2014). It might be appropriate to generalize from this research for all states in Germany, as all states are relatively similar due to the same legal environment and a mutual culture. However, it could be that there are unknown parameters, which influence especially one state, but not another. This would lead to wrong conclusions. Furthermore, to apply outcomes of this study to other countries might be problematic, as they may face a different environment and have distinct political and administrative systems (Pollitt & Bouckaert, 2011). In summary, this research design might be an opportunity to investigate the research question and test the hypotheses. Especially the strengths regarding the reliability and internal validity outweigh potential limitations and as a result support the execution of this study.

3.5 Analysis strategy

To examine the relationship between the explanatory variables and participatory budgeting, a regression analysis will be used. A regression analysis is a statistical technique to investigate the effects of the explanatory variables on the outcome variable (Neuman, 2014). A regression analysis is advantageous, as it can analyse different correlations among different variables in one model, and it is possible to control for other explanatory variables (Neuman, 2014). Due to the fact that the dependent variable is binary, this research will analyse the data with a binary logistic regression analysis. The binary logistic regression analysis measures if the explanatory variables are related to the likelihood that participatory budgeting will be present with the value 1 (Studenmund, 2014). This regression will be done by using the software SPSS 24. The analyses of the data follow in the next chapter.

(25)

24

4. Data Analyses and Results

In this chapter, the results will be illustrated and analysed. As a first step, the variables will be separately examined by analysing their descriptive statistics. Afterwards, the correlation of the different variables among each other will be presented in a correlation matrix and explained. Lastly, binary logistic regression analyses will be conducted to test the hypotheses.

4.1 Descriptive statistics

Table 4 illustrates the descriptive statistics of the used variables. The first row describes participatory budgeting. It is coded as dummy variable, whereas 1 reflects on the presence of participatory budgeting and 0 on the absence. The mean with 0,16 is rather low, which implies that the majority of the municipalities in North Rhine-Westphalia have not applied participatory budgeting. Yet, 65 out of 396 municipalities have some form of participatory budgeting introduced.

Table 4: Descriptive statistics

N Minimum Maximum Mean Std. Deviation

1 Participatory budgeting (1 = yes) 396 0 1 ,16 ,37

2 Functional differentiation 394 2 49 8,98 7,95

3 Organisational size 396 25 16.300 521,88 1381,71

4 Slack resources 396 ,00 10.266,02 2.184,65 1.654,62

5 Political support 396 ,18 ,76 ,42 ,09

6 Size of the population 396 4.236 1.060.582 45.114,94 89.395,33

7 Wealth of the citizenry 396 15.313,26 36.381,83 22.008,02 2.747,93

8 Political activity of citizenry 396 ,39 ,72 ,54 ,06

The second row represents functional differentiation. For 394 municipalities values for functional differentiation could be detected by analysing organisational charts on their homepages. Two municipalities had no organisational chart on their homepage and were contacted via e-mail to get this information. However, they did not respond. The number of administrative units under the hierarchy level of the mayor ranges from 2 to 49. The average municipality in this sample has around nine functional departments. Given the wide range of values and the relatively high standard deviation, the assumption that there is variance among the functional differentiation of municipalities in this sample is confirmed.

Organisational size is expressed by the fulltime equivalent employment and ranges from 25 to 16.300. Thus, there is a relatively large span of municipalities with regard to the organisational size. The average municipality has around 522 full time equivalent employees. In addition, the standard

(26)

25 deviation is relatively high. It can be concluded that this population represents a large span of municipalities in terms of organisational size and has variation among the observations. This was also assumed in previous chapters and has been proven by the descriptive statistics in this sample.

The fourth row represents the slack resources. It is measured negatively by the debt per citizens ratio of a municipality in Euro. The debt ratio ranges from €0 to over €10.266,02. The mean is €2.184,65. Twelve municipalities had no debt, whilst the majority of municipalities are in debt. However, due to the wide range of debts and the broad distribution, municipalities have variance among their slack resources.

Political support is expressed negatively by the share of votes for the conservative party CDU. This share has a minimum of 18 % and a maximum of 76 % among municipalities in North Rhine-Westphalia. On average the CDU has a 42 % of the share of the votes in a municipality. It can be concluded that the CDU is relevant party, which can influence politics in many municipalities. This argument is strengthened by the relatively low standard deviation. Hence, in the majority of municipalities, the share of votes for the CDU is close to the mean, which is already quite high with 42 % share of votes. The sixth row represents the size of the population, which is measured by the number of citizens. This number ranges from 4.236 to 1.060.582. Thus, this span expresses that in this state there are relatively small municipalities, but also relatively large ones. The average municipality has 45.115 citizens. Due to values for the mean and the standard deviation, it can be concluded that there are relatively large municipalities, but their number is quite low. However, there is enough variance among this variable to test our assumptions.

The average wealth of citizens is illustrated in the seventh row. It is measured by the average income a citizen has for saving or consuming purposes in the municipality. The average income ranges from €15.313 to €36.382, while the mean is €22.008. Whereas the standard deviation is not relatively large, there is still some variation in this population.

The last row illustrates the political activity of the citizenry, which was measured by percentage of voter turnout. The voter turnout in the municipality ranges from 39 % to 72 %, while the average voter turnout is 54 %. The standard deviation is relatively low, which means that the majority of values are distributed closely around the mean. However, there is variance and outliers, which makes it appropriate to test this variable.

4.2 Correlations among the variables

Table 5 illustrates pairwise correlations among the used variables. A correlation describes if the relationship between two variables is positive or negative and also how strong this relationship is

The Link of Functional Differentiation and Innovation