The competing impact of network politics and hamula (clan) politics on perceived network performance: the case of minority networks

ABSTRACT

Studies of network governance have found that network management contributes to the effective performance of the network. Nevertheless, the influence of politics on the relationship between network management and network performance has not yet been examined. This is particularly true within traditional minority communities that have only partially adopted aspects of modernity and democratic governance. We conducted a survey among 160 Arab Israeli network members to examine the relationship between network management and perceived network performance by taking into account the effect of internal and external politics. Our findings show that network politics and hamula politics have a moderated mediation effect on this relationship. The study makes a theoretical contribution by emphasising the role of socio-cultural, managerial and political aspects of traditional communities when trying to develop local network governance. In these frameworks, the adoption of local network governance is challenging because democracy and network governance are still under-developed.

KEYWORDS:

• Local governance
• network management
• hamula politics
• clan politics
• network politics
• perceived network performance
• ethnic minority

Introduction

Government policymakers face a variety of what the literature has called ‘wicked problems’, defined as new, complex, persistent challenges. These daily problems involve moral, political and professional issues. Solving them is difficult because of the lack of technical knowledge about the nature and extent of the risks involved for individuals and society as a whole (Ritchey 2013; Van Bueren, Klijn, and Koppenjan 2003). According to the New Public Governance theory, networks might be an efficient and innovative way to find solutions to wicked problems. They allow actors from different sectors to work together to achieve their shared collective goals and outcomes that would not be achieved by one of the actors alone, by the marketplace or by the bureaucratic hierarchy (Ferlie et al. 2011; Rhodes 1997). The theory emphasises the concept of network governance, which implies governing through a self-regulating network consisting of interdependent and autonomous actors involved in inter-organisational and horizontal interactions. They direct the network with the goal of providing public services. They are not part of a bureaucratic hierarchy or dependent on the laws of the marketplace (Klijn and Koppenjan 2012; Osborne 2010; Stoker 1998; Torfing 2005; Uster, Beeri, and Vashdi 2018).

The second-generation scholars of the Network Governance Theory argued that network governance is prone to failure because it contains tensions that complicate the interaction process (Klijn and Edelenbos 2008; Sørensen and Torfing 2008). Thus, the complex interactions among the network members must be guided by various meta-governance mechanisms called network management strategies. Examples include developing methods that facilitate interactions between actors or creating rules and strategies to increase information exchanges (Klijn and Edelenbos 2008; Uster, Beeri, and Vashdi 2022).

Network governance theorists present empirical evidence that network management at the local level improves perceptions about network performance in societies with modern democratic and pluralistic values (Kickert, Klijn, and Koppenjan 1997; Klijn et al. 2015). However, much less is known about the effect of network politics on local network management in traditional communities that are struggling with the acceptance of these modern values. This struggle stems from the situation of ‘partial modernisation’ in which traditional values such as tribalism, which monopolises power and prevents collaborations, co-exist with the acceptance of modern trends (Abu Sharqie 2008; Rueschemeyer 1976). Hence, the theoretical assumptions of network governance may be valid only in certain contexts.

In Israel, as a result of the contact with the Jewish majority, Arab society has partially adopted modern democratic procedures. Nevertheless, the Arab minority is still a traditional society in which the hamula, meaning the clan or the extended family, and its politics play a major role in the local arena. As a traditional kinship social structure, hamula politics is a force outside the network. It affects local politics by obligating Arab local authorities to hire and interact with Arabs based on family relationships rather than professional skills and qualifications. In addition to the external challenge of hamula politics, Arab local authorities suffer from the internal challenges of a politicised, less democratic, less rational, less professional and less merit-oriented environment for local stakeholders and local networks (Ghanem and Mustafa 2009). Accordingly, we investigate how a situation characterised by hamula politics affects the relationship between network management and perceived network performance. In addition, we explore the role of network politics in the relationship between network management and perceived network performance in such traditional communities.

Our study makes a theoretical contribution to the network management literature by emphasising the need to consider the socio-cultural, political and managerial characteristics of local networks when examining the effectiveness of network governance in traditional ethnic groups. The success of network governance in these communities is particularly challenging because it is usually the central government and the majority culture that impose local governance on ethnic minorities rather than the latter adopting it gradually. In these communities, the local network members and leaders must deal with external challenges from the hamula, which strongly opposes attempts to wrest power from it and thus hinders the development of local governance.

We examined these dynamics using a program called Etgarim (Hebrew for Challenges) created by the Israeli central government to encourage Arab youngsters to participate in informal education programs. The goal of these programs is to improve the youngsters’ social and personal skills, enrich their formal education and provide them with skills that will help them integrate into the job market. Importantly, these programs were created through collaborations between various bodies and sectors, and managed through networks rather than being developed through the hamula and controlled by it. We surveyed 160 local network members of the Etgarim Plan in 30 Arab local networks.

The rest of the paper is structured as follows. The next section develops the theoretical bases of the study and the basic assumptions. Section three describes the methodology, the sample and data analysis. After presenting the results, we discuss the insights from the findings.

Theoretical framework and development of hypotheses

Network governance, network management and network performance

The theoretical roots of the New Public Management (NPM) are in studies that emphasise rational, intra-organisational management based on inputs and outputs. The governance mechanism of the NPM emphasises the competitive marketplace and contracts between independent contractors. The shortcomings of the NPM have prompted the development of the New Public Governance (NPG) (Osborne 2010). In contrast to the NPM, the NPG is rooted in notions from organisational sociology and network theories. Its focus is on the inter-organisational governance of autonomous and interdependent suppliers who have ongoing relationships based on trust and relational contracts (Osborne 2006; Stoker 1998).

Network governance has proven beneficial in dealing with local challenges. It can identify new problems through negotiations. It can also establish joint responsibility for policies, reducing resistance to their implementation. Finally, it helps decision-making by providing more needed information. Hence, network governance can potentially improve the performance collaborative networks (Kenis and Provan 2009; Provan and Kenis 2007; Sørensen and Torfing 2005).

Although network governance plays a crucial role in dealing with wicked problems, it has not escaped criticism. Sørensen and Torfing (2008) enumerated some persistent obstacles facing network governance. Examples include unresolved tensions and conflicts, weak and ineffective leadership, and external events and circumstances that can disrupt the policy process. Thus, the interaction between the network actors is complex. It can also produce conflicts mainly because there are different perspectives about the nature of the problems and the proper solutions. Hence, the cooperation between the actors requires a deliberate attempt to steer these interactions within the network.

Careful network management can help overcome these obstacles (Kickert, Klijn, and Koppenjan 1997; Rhodes 1997). Various networking management strategies must be used to develop and promote interactions between actors, creating better coordination and outcomes (Agranoff, McGuire, and Agranoff 2001; Klijn, Steijn, and Edelenbos 2010; Klijn et al. 2015; Provan, Fish, and Sydow 2007). Network management strategies are the tools managers use to guide the collaboration and help improve the network’s performance (Klijn, Steijn, and Edelenbos 2010; Klijn et al. 2015; Ysa, Sierra, and Esteve 2014). Kickert, Klijn, and Koppenjan (1997) argued that network management increases the governance potential of the network because the manager can improve the negotiation process within the network by mediating conflicts, planning processes, and engage in diplomacy. The need to understand and assess performance stems from the realisation that not all networks work (Torfing 2005).

Provan and Milward (2001) suggested that a network could be evaluated at three different levels. The first is the community level, which focuses on the contribution the network makes to the communities it serves. The second is the participants’ level, which considers the performance of the agents and clients participating in the network. The third is the quality of the network itself and its contribution to its constituent autonomous organisations. In this study we focused on the third level by considering the members of the network and their opinions and perceptions. Hence, we evaluated the performance of the network with regard to its positive results and achievements in meeting the strategic goals of the network’s collaborators, as reflected in the perceptions of the network members (McGuire and Agranoff 2011; Provan and Kenis 2007).

In addition to this goal, we also explored the relationship between network management and network performance. The literature on network performance provides empirical evidence about the positive strong relationship between network management and network performance, particularly in the eyes of its members (Klijn, Steijn, and Edelenbos 2010). Klijn, Steijn, and Edelenbos (2010) developed a PODSCORB model for effective managerial network strategies. The model consists of four categories:(1) creating temporary tools such as steering groups to facilitate interactions; (2) exploring where actors look for solutions and the information needed to develop new content such as new ideas; (3) connecting and enhancing interactions between actors in the network; and (4) processing rules such as increasing information exchanges designed to facilitate communications between actors in the network. These strategies are known to help overcome the disagreements between the network members about values and perceptions and improve the network’s stability, resulting in productive outcomes (Klijn et al. 2015). Taking into account the importance of network management for perceptions about network performance, regardless of the context, we hypothesise that:

H1:

The relationship between network management and network performance is positive.

However, one of the limitations of these studies is that they have generally been conducted in democratic societies in the Western world. There have been several attempts to conduct cross-national research to determine if the results hold. For example, O’Toole et al. (2007) examined whether managerial networking is a common pattern of governance among local government managers in the United States and United Kingdom. They noted that, while networking exists in both countries, it differs in various ways due to local political realities. Later on, Klijn et al. (2015) conducted a cross-national study of the relationship between network management and perceived network performance. They examined three countries – Taiwan, Spain, and the Netherlands – to determine whether findings from the US and the Netherlands would hold in Taiwan and Spain. They found that network management matters the most in all three countries, proving that regardless of the context, network management is crucial for achieving effective network performance. However, they also reported cross-national differences. For example, the effect of unexpected external events on perceived network performance was negative in Spain, positive in the Netherlands, and insignificant for Taiwan. They explained these variations based on differences in how long the countries had been using network governance, the extent to which it was used and the cultural setting. Klijn et al. (2015) noted that external changes and actions could also change the course of the interactions, increasing the required efforts and network management strategies. Therefore, they concluded that the influence of the external environment is vital because external events can make it hard to achieve effective results in the network. Based on this assumption, we examined how external politics operating outside the network, as well as internal politics within the network, affect the collaborators’ perceptions about the network’s performance.

As Torfing (2005) mentioned, the tension between the representative elected body and network governance is reflected in the issue of accountability. While in a representative democracy there is hierarchical accountability, network governance has horizontal accountability. Horizontal relationships in local governance require that local governments be plural and pluralistic, as Osborne (2010) suggested, in order to allow inter-organisational interactions. Nevertheless, once again, most of these studies have been conducted in Western democratic regimes where the values of pluralism and decentralisation are well-established principles. However, little scholarly attention has been paid to the emergence of local network governance in societies with partial modernisation and democratisation processes. For instance, Uster, Vashdi, and Beeri (2022) found that in Israeli local authorities, the degree to which the local authority is centralised explains the link between the network’s leadership behaviours and the network’s effectiveness. In Arab municipalities in particular such centralisation is counterproductive, impeding the network’s effectiveness.

Hamula politics and network politics

Based on these findings, we explored the role of external politics in determining the network’s performance in societies where such political forces play a major role. One example of such a society is Arab Israeli municipalities where hamula politics strongly affect the political and social dynamics (El-Taji 2008). Combining the two definitions of Lustick (1980) and Rosenfeld and Al-Haj (1990), a hamula is defined as a patrilineal descent group composed of all the members related biologically to the common greater father, or members who have related themselves socially to a certain hamula by fictive relatedness in order to obtain the advantages of the hamula’s protection and rights along with its responsibilities and commitments.

By the end of the 1948 War, the Palestinians moved from a majority to an ethnic minority due to the expulsion and flight of many of the middle and upper classes, and the urban landowning, professional and religious elite (Rosenfeld and Al-Haj 1990). This elite represented a small number of large and influential families located in the largest cities of Mandate Palestine (Lustick 1980). At the time, these families held superior political status (Rosenfeld and Al-Haj 1990; Lustick 1980).

The importance of kinship affiliation lies in the fact that it serves as the context for the allocation for goods, services, power, social interactions, and self-identification (Rosenfeld and Al-Haj 1990). This affiliation is particularly important at the local level because it grants legitimacy to the individual’s position in the local community. Thus, people’s place in the community depends on their family of descent. Khalaily and Navot (2020) claimed that the literature on tribalism, clans, and hamulas suffers from social and cultural biases. They suggested that few researchers consider the hamula as a political phenomenon. The authors maintained that the hamula as a social structure could affect the local political order, meaning it has the potential for massive corruption. Indeed, there are two proposed explanations for the continuation of hamula politics in Arab-Israeli local authorities: the state-society demarcation and the modernism-traditionalism dichotomy.

The state-society demarcation explanation emphasises how the fragmentation of the Israeli population along Jewish and Arab lines gained strength on the institutional level by mediating the hamulas’ competition for control, mainly in the local arena. The control of the Jewish state was also prominent in the education system. Al-Haj (1995) argued that the dominant group utilised the education system as a mechanism of social control. The ruling parties aimed at controlling Arab society and maintaining traditional structures. Instead of viewing education as a vehicle for the development and social change of the Arab minority, the dominant group denationalised and deprived the education system of any authentic content. In the same vein, informal education in Arab society is controlled and supervised by the Israeli state representing, in the end, the needs of the dominant group (Agbaria 2007). The rival relationship between Arab society, as an ethnic minority, and the Jewish state was reflected in the discrimination and limited resources and opportunities that were supplied to Arab local authorities. This lack of resources eventually transformed the Arab local authorities into an arena whereby hamulas compete for control of the limited resources, job opportunities, and political power.

The second explanation stresses that the mismanagement and inefficiencies in Arab local authorities are a by-product of the cultural dichotomy in Arab society between traditionalism and modernism. Culture-oriented approaches to organisations and management maintain that forces such as individualism versus collectivism influence the leadership styles and outcomes of decision makers (Hofstede 1984). As a cultural force committed to traditional culture, hamula politics weakens the performance of local governments because clans patronise their kin by conditioning jobs on clan ties rather than merit. They also grab state resources and pervade the modern state bureaucracy informally by creating informal, political, and economic rules that are not pluralist, equal, representative, transparent, or fully participatory (Collins 2004), leading to financial problems and mismanagement (El-Taji 2008).

In traditional-oriented local authorities that do not necessarily follow the path of modernisation and democratisation and that are heavily influenced by competing, external political powers, such as hamula politics, there will be more effort to use network management strategies. Increased network management strategies potentially improve the internal interaction and collaboration between the network members and reduce the disruptions caused by external politics. However, when interactions within the network improve as the result of network management strategies, members’ perceptions regarding the presence of hamula politics may actually increase because the clash between these two views becomes more evident and stark. Nonetheless, the less the network members feel the presence of hamula politics, the more effective they believe the network’s performance is. Therefore, the second hypothesis posits the negative mediating role of hamula politics in the relationship between network management and perceptions about the network’s performance (see Figure 1). Thus, we hypothesise:

Figure 1. Moderated mediation effect of perceived hamula politics and perceived network politics on the relationship between network management strategies and perceived network performance.

H2.

The more network management strategies are employed, the more the network members will report the presence of hamula politics. However, the less the network members feel the influence of hamula politics, the better they will rate the network’s performance

While Klijn et al. (2015) explored the effect of external events and changes on the relationship between network management strategies and network performance, Agranoff (2006) stressed that one of the lessons learned from the networks is that internal politics within and between the network members and collaborating organisations influences the success of the network. The research on perceived organisational politics (POP) emphasises the struggle internally for acquiring resources and personal benefits and achieving goals. Perceived organisational politics is defined as an individual’s sense that others are engaging in manipulative behaviours and coercive acts at the expense of others for short-term personal gains (Abbas and Raja 2014; Vigoda 2000). Therefore, organisational politics is a behaviour whose goal is to maximise one’s self-interests even if doing so contradicts collective organisational goals (Vigoda 2000).

Most studies regard organisational politics as a negative behaviour because it is associated with manipulation and illegitimate ways to fulfill personal objectives (Vigoda 2000). However, organisational leaders can mitigate this effect by educating subordinates, increasing their organisational understanding and making them feel that they are active, valued and have control (Ammeter et al. 2002). Whether organisational politics is regarded as positive or negative, its presence affects internal organisational beliefs, norms, perceptions and behaviours. Based on the definition of POP, we define the perception of network politics as the network members’ perceptions about the power behaviours within the network. These perceptions increase the impact of internal politics, regardless of whether the participants regard them as positive or negative.

Jolink and Dankbaar (2010) emphasised the importance of managerial practices promoting networking behaviours among actors. They claimed that the employees’ perceptions about their work environment depend on their managers’ encouragement of the value of networking. In other words, if the network’s managers support inter-organisational networking as a norm and often use network management strategies, employees may feel that internal network politics play a major role in their work. It should be remembered that network politics may have a positive effect. Thus, strong feelings about the presence of network politics may crowd out the impact of other competing, external forces such as hamula politics. In turn, hamula politics will have less of a destructive effect on perceived network performance. Thus, as Figure 1 indicates, we maintain that network politics moderates the relationship between network management and hamula politics. Within partially modernised democratic local networks (i) network politics weakens the impact of hamula politics and (ii) network politics and hamula politics both affect the relationship between network management and perceived network performance. Taken together, we posit that:

H3.

When network members sense that network politics plays a major role in the network, they are more likely to think that network management is being used and less likely to sense the presence of hamula politics.

Thus, our full model posits a moderated mediation relationship, i.e., network politics and hamula politics have a moderated mediation effect on the relationship between network management and network performance:

H4.

When network members sense that network politics plays a major role in the network, they are more likely to think that network management is being used and less likely to sense the presence of hamula politics. As a result, they believe that the network is performing well.

Methodology

The research context: network management in Arab-Israeli local authorities

Israel is considered a unitary state with two government levels: central and local. Israel traditionally takes a centralist and conservative view towards local authorities, particularly towards those located in the geographic and societal periphery of the country (Beeri 2021; Beeri and Razin 2015). Hence, in terms of institutional arrangements, the central government tends to intervene a great deal in the local arena (Beeri and Saad 2014). In contrast to the Jewish local governments, the shift from local government to local governance in Arab society is still unimplemented mainly due to the traditional political culture at the local level, resulting in the weakness of the division between politics and administration (Abu Sharqie 2008). Arab councils prioritise personal relations and interests over professionalism and good performance (Khamaisi 2007; Pollitt and Bouckaert 2004), significantly hindering collaborations in the provision of public services from different sectors. Ghanem and Mustafa (2009) claimed that the scarcity of voluntary organisations and civil society institutions in providing local public services leads to the poor management and quality of local services.

Recently, however, a shift to local governance has begun encouraging a culture of networking and collaborations (Beeri 2021). The Etgarim Plan, an informal education plan created pursuant to Government Resolution (GR) 922, seeks to promote the participation of Arab youth in informal education programs through network management and collaborations between various bodies from different sectors. The programs offer youngsters opportunities to improve their social and personal skills, supplement their formal education, and develop the leadership capabilities that will help them find jobs. However, Arab youths rarely participate in informal education. We can define this situation as a wicked problem because it is primarily socially based. In addition, it is particularly difficult to resolve because Arab society is a heterogeneous minority, making it hard to find uniform and all-encompassing solutions (Hadad Haj-Yahya and Rudnitzky 2018).

The five-year Etgarim Plan received 650 million NIS (new Israeli shekels, equal to about 25 American cents) of the 130 million NIS allocated to 67 local authorities on a per-capita basis, upon the submission and approval of an annual work plan submitted to the Youth and Society Administration of the Ministry of Education (YSA) (Schneider and Shoham 2017). In this program the YSA is considered the network organisation that administers the plan in local authorities. It consists of an Arab executive director, staff, and board that operates out of a physically distinct office in the Education Ministry. Although the YSA mandates the Etgarim Plan as part of the central government, it gives municipalities the autonomy to use this budget to provide regulated and approved informal education programs for their communities through collaborations to professionalise the informal education staff and increase the number of informal education providers. This decentralised policy will boost local governance in Arab society (Hadad Haj-Yahya and Rudnitzky 2018; Schneider and Shoham 2017). Hence, as a society dealing with politicisation on the local level while undergoing a process of adopting network governance, Arab local authorities are a good target for examining our hypotheses.

Sample and procedure

We surveyed 30 Arab local authorities out of 67 Arab local authorities implementing the Etgarim Plan managed as a network that contains members from the public, private and third sectors. The unit of the analysis was being a member of the network. In each local authority, there is a network of informal education suppliers and local school managers, collaborating with youth coordinators managed by the Youth Department head employed by the local authority and the YSA. The five-year program, which began in 2015, had 67 local authorities in its first year, excluding Druze and Circassian communities that represent 16 local authorities (Schneider and Shoham 2017). The sample did not contain the Jewish municipalities because the Etgarim Plan was approved only for Arab society. Due to time constraints and the sensitivity of the respondents when answering the survey, we were able to sample 30 local authorities from the Southern, Northern, Central, and Haifa districts. Each of these local networks contained 5–8 members. Given that we reached 30 local authorities out of 67, the response rate was 50%.

We asked the network members to evaluate the extent to which they felt that the Youth Department head was using network management strategies and their assessment of the performance of the Etgarim Plan in their local authority. They were also asked to evaluate their perception of the influence of hamula politics and the political environment within the network

In total, there were 160 members from different sectors. Their age ranged from 20 to 60 (Mean = 36.81, Std. = 8.72) and they were approximately evenly divided by gender (48% males, 51% females). With regard to education, 56.3% had a B.A. degree and 43% had an M.A. or higher. On average, they had held their position for five years (Mean = 5.84, Std. = 4.33). We obtained the data about the local authorities from Israel Bureau of Statistics (2015). The information included their socio-demographic index (Mean = 2.25, Std. = 1.00), population size (Mean = 22.99, Std.=.81) and their rating on the periphery index measuring their distance from the centre of the country (Mean = 4.46; Std. = 1.18). In terms of location, 51.9% of the local authorities were from the north, 25.6% from the centre or Haifa, and 22.5% were from the south.

Measures

In order to ensure the validity and reliability of our key concepts, we measured them using scales from previous studies. We then created an aggregated index of the scores (see Appendix A).

Dependent variable – perceived network performance

We measured perceptions about network performance based on the responses of the participants to items assessing the achievement of the plan’s goals. We asked the participants to indicate the extent to which each course or program was successful from 1 (failed) to 5 (very successful). McGuire and Silvia (2009) suggested that members of a network might vary in the degree to which they think the network has achieved their individual organisations’ needs and goals as well as the collective goals of the network. Hence, it is important to consider that assessments about the effectiveness of the network also depend on whether they feel their own needs have been fulfilled (Cronbach’s alpha=.86).

Independent variable – network management

We assessed network management using 11 items related to Klijn et al. (2015) four categories of arranging, connecting, exploring, and process rules. We asked the participants to indicate the degree to which they agreed with the statements on a Likert scale from 1 (strongly disagree) to 5 (strongly agree) (Cronbach’s alpha=.78).

The mediator – perceptions of hamula politics

We operationalised hamula politics using a scale developed by Hayek (2013) in Hebrew. The scale contains four items evaluated from 1 (strongly disagree) to 6 (strongly agree) that relate to the degree of interference of the hamulas, mainly in choosing the candidates in the primaries for local elections. The questionnaire was adjusted to the network environment. We also added two questions related to the influence of hamula affiliation on the general process of decision-making in local authorities (Cronbach’s alpha=.81).

The moderator – perceptions of network politics

We used the Perception of Organizational Politics (POP) Scale to investigate perceptions of network politics as a possible mediator. While networks are not considered organisations, they can be treated as a framework connecting different organisations aiming to maximise their interests and goals. The questionnaire focuses on the degree to which the participants view their network environment as political and, as a result, unjust and unfair (Kacmar and Carlson 1997; Vigoda 2000). The survey used 11 items of the POPs scale developed by Kacmar and Carlson (1997) and was adjusted to fit the network environment. The items are rated on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree) (Cronbach’s alpha=.75).

Control variables

We also included several control variables that might affect perceived network performance using data from the Israeli Central Bureau of Statistics. First, we measured the size of the local authority as the number of citizens (log). Second, we used the periphery index to determine whether the local authority was in the central or peripheral region of the country. Lastly, the local population’s socioeconomic demographics indicate the local authority’s socio-economic level in terms of its population, education, standard of living, and employment. The cluster is rated from 1 to 10, where 10 represents the highest socioeconomic status (Navot and Beeri 2018). Control variables also included the type of local authority and the district of the local authority. Finally, we asked the participants to indicate their gender, age tenure, education level, and job title within the network.

Robustness checks

Arguably, there are possible concerns regarding the reliability and validity of the data, measures and results, and about multicollinearity and common source bias, inter alia, due to our modest sample size and measurement instrument. To address these issues we used several procedures and conducted robustness tests. First, we collected information from at least five of the eight network members from local authorities that varied in terms of their geographical area, size of population and residents’ socio-economic status. Thus, the participants are representative of the population of the network management of Etgarim in terms of gender, age, position and affiliation. Second, we conducted several reliability tests, which revealed a high level of internal consistency of the measures (Cronbach’s alpha=.75 to .86). Third, we also used a non-response bias check for gender, age, education, position, and tenure. Most of the groups were homogeneous, meaning there were non-significant differences on Levene’s Homogeneity of Variance Test. The only exceptions were for women who felt that network management strategies were used a bit more often than men (M = 3.48 vs. M = 3.26) and more educated network members who regarded the role of hamula politics as stronger and network performance as poorer (M = 2.12, M = 3.22, respectively) than less educated network members (M = 2.11, M = 3.23, respectively). Fourth, we calculated variance inflation factors (VIF) to detect possible multicollinearity. The ranges of the VIF were very low (1.34 to 1.76 [of common thresholds of 3.0, 5.0 or 10.0]), indicating no concern about multicollinearity in our data. Finally, we checked for the presence of common source bias because we collected data for both the dependent and independent variables using the same measurement instrument. However, Harman’s single factor test for common source bias revealed a very low percentage of variance (% = 26.02 [much less than the recommended threshold of 50%]), indicating no problem with common method bias.

Data analysis

We used the Statistical Package for Social Sciences (SPSS) 23.0 software to analyse the data. We first conducted a descriptive analysis and determined the inter-correlations. Then, using multiple regression analysis, we examined the linear relationships between the independent and dependent variables. Finally, we tested the mediating, moderating and mediating-moderating hypotheses with the regression-based bootstrapping approach described in Hayes’ PROCESS macro (model #7).

Findings

Table 1 presents the means, standard deviations, bivariate correlations, and Cronbach’s α for all of the research variables. As the table indicates, the four main variables are correlated with each other. Notably, there was a significant positive association between the extent of the network management strategies used and perceived network performance (b = .558, p < .01). This finding supports H1 and indicates a zero-order level relationship. In other words, the more network management strategies the participants reported were used, the more likely they were to regard the performance of the network as successful.

Table 1. Multiple correlation matrix and descriptive statistics for the research variables (Cronbach’s α in parentheses).

	Mean (S.D.)	1	2	3	4	5	6	7	8	9	10	11	(α)
1. Network performance	3.23 (.70)												(.86)
2. Network management strategies	3.37 (.63)	.558**											(.78)
3. Network politics	2.62 (.68)	−.295**	−.481**										(.75)
4. Hamula politics	2.12 (.83)	−.388**	−.422**	.610**									(.81)
5. Age	36.81 (8.72)	−.085	−.159*	.084	.047
6. Gender	.51(.50)	.153	.180*	−.111	−.077	−.332**
7. Education level	2.43 (.49)	−.012	−.017	−.049	.003	.512**	−.375**
8. Tenure	5.84 (4.33)	−.073	−.145	.037	.564*	.762**	−.264**	−.064
9. Size of population (Log) level	4.06 (.81)	−.123	.048	−.001	.104	.125	−.029	−.144	.026
10. Local authority central-peripheral	4.46 (1.18)	.262**	.175*	−.108	−.285**	−.021	.176*		.107	−.172*
11. Local authority soc-eco status	2.25 (1.00)	.151	.076	−.150	−.306**	.016	.230**		.074	−.286**	.297**
12. Type of local authority	.36 (.48)	.159*	.063	−.054	−.319**	−.012	.137	−.176	−.0002	−.087	.362**	.213**

N = 160; *p < .05, **p < .01, ***p < .001.

Additionally, both perceived network politics and perceived hamula politics were significantly and negatively correlated with the employment of network management strategies. In other words, the more network management strategies the participants reported were used, the less likely they were to regard to the politics of the hamula as affecting the performance of the network (b=−.422, p < .01), and the less likely they were to regard the network environment as political (b=−.481, p < .01).

Another interesting finding is the significant positive association between perceived network politics and perceived hamula politics (b = .610, p < .01). However, perceived hamula politics seems to have a stronger relationship than perceived network politics. There were also negative correlations between perceived hamula politics and the local authority’s location in the centre or periphery of the country (b=−.285, p < .01) and its socio-economic status (b=−.306, p < .01). In other words, the more removed the local authority was from the national centre of the economy and politics, and the lower its socio-economic status the more its network members regarded the effect of hamula politics as extensive.

Next, we conducted a hierarchical multiple regression analysis to assess the relationship between network management strategies and perceived network performance. As Table 2 shows, the regression was conducted in three steps. In the first step, we included the micro-level factors to control for tenure, age, gender, and education. No significant effect appeared in the first step (F_(4,155) = 1.201; p=N.S.). In the second step, we controlled for the macro-level factors (socio-demographic cluster, size of population (Log), peripheral location and type of local authority). A peripheral location was a significant predictor (F_(4,151) = 3.090; b = .218; p < .01), indicating that the closer the network was located to the national centre, the better its perceived network performance. In the third step, we added the main variables and the model was significant (F_(3,148) = 8.05; p < .01). Only the effect of the extent of the network management strategies employed proved significant (ΔF = 21.372; b = .498; p < .01), above and beyond all other variables. These results support H1, which posited that the more network management strategies the participants reported were used, the more likely they were to regard the performance of the network as successful.

Table 2. Findings of a hierarchical multiple regression analysis (standardised coefficients) with perceived network performance as the dependent variable.

	Dependent Variable = Network Performance
Variable	Step 1 β	Step 2 β	Step3 β
Constant Micro-level factors	3.034	2.517	1.426
Tenure	−.006	−.014	−.034
Age	−.072	−.022	.045
Gender	.160	.101	.046
Education	.087	.099	.020
Macro-level factors
Socio-demographic cluster		004	−.005
Size of population (Log)		066	−.116
Periphery cluster		218*	.105
Local authority type		056	.027
Main variables
Network management strategies			.498**
Perceived network politics			.058
Perceived hamula politics			−.161
R^2	.030	.103	.374
Adjusted R^2	.005	.056	.328
F	1.201	2.178*	8.05***
ΔR^2		.73	.271
ΔF		3.090*	21.372**

N = 160; *p < 0.05; **p < 0.01; ***p < 0.001.

These interesting results led us to investigate the relationship between the main variables added in the third model. We used Hayes (2015) PROCESS (model #7) macro for SPSS to assess H2, H3 and H4. As the results in Table 3 indicate, network management was significantly and positively related to perceived network performance, reconfirming H1 (b = .520, P < .001). Next, we found a significant mediation effect. The use of network management was related to the perceived performance of the network through perceived hamula politics. Thus, as H2 posited, the more network management strategies the participants reported were used, the stronger their perceptions about the role of hamula politics (b = .653, P < .01). In turn, low levels of perceived hamula politics were significantly and negatively associated with perceived network performance. Hence, the weaker the perception of hamula politics, the stronger the perceptions about the performance of the network (b=−.157, P < .01).

Table 3. Conditional indirect effect (PROCESS model #7) of network management strategies on perceived network performance through perceived hamula politics contingent on perceived network politics.

	I: Dependent Variable= Perceived hamula politics R^2=.428***			II: Dependent Variable= Perceived network performance R^2=.343***
Predictor	B	SE	t	B	SE	T
Constant	−1.579	1.100	−1.435	1.677***	.456	3.678
Network management strategies	.653	.305	2.140**	.520	.082	6.378***
Perceived hamula politics				−.157	.061	−2.566**
Perceived network politics	1.700	.361	4.712***
Network management strategies X Perceived network politics	−.323	.107	−3.012**
	Conditional indirect effect of network management strategies on perceived hamula politics with simple slopes for perceived network politics
Perceived network politics	Effect		SE	LLCI		ULCI
−1SD	.008		.120	−.230		.245
Mean	−.172*		.094	−.358		.015
+1SD	−.423**		.114	−.648		−.197
	Indirect effect of highest-order interaction – Moderated Mediation
	B		SE	LLCI		ULCI
	.051***		.029	.008		.118

N = 160; p < .10*, p < .01**, p < .001***; All betas are standardised.

Table 3 also shows a moderating effect of perceived network politics. The relationship between network management strategies and hamula politics is contingent on perceived network politics, supporting H3. As Figure 2 illustrates, a simple slope analysis indicated that at low levels of perceived network politics, there was no significant association between the use of network management strategies and perceived hamula politics (b = .008, p=NS). However, there were significant and increased negative associations between network management strategies and perceived hamula politics at medium and high levels of perceived network politics (b=−.172, b=−.423, respectively, all p < .01). These findings suggest that when perceived network politics is medium to high, using more network management strategies helps reduce perceived hamula politics.

Figure 2. Conditional indirect effect of network management strategies on perceived hamula politics with simple slopes for perceived network politics.

Taken as a whole, network politics and hamula politics are competing elements that have a moderated-mediated effect on the relationship between network management and network performance (b = .051; [LLCI]=.008; [ULCI]=.118). When network politics is perceived as medium to high, network management strategies are effective in attenuating perceived hamula politics. In turn, the results indicate that in networks where the level of perceived hamula politics is low, network members are more likely to regard the performance of the network as effective. As expected, these results confirm H4.

Discussion and conclusion

Research has reported how network management contributes to perceptions about the performance of networks. However, most of the studies in this area have been conducted in modern societies where democracy and governance are well established and accepted. In contrast, we know little about whether this result holds in traditional societies that are still struggling to adopt modern, pluralistic values. To fill this theoretical and practical gap in the literature, we conducted a study in Arab Israeli local networks.

As the results show, in such traditional oriented contexts, where minorities struggle with democratisation and governance, using network management strategies does help improve perceptions about network performance. This evidence supports and accords with the major body of literature indicating that managerial strategies are crucial for improving perceptions about a network’s outcomes (Kickert, Klijn, and Koppenjan 1997; Klijn et al. 2015; Sørensen and Torfing 2008). Furthermore, internal and external politics have a role in shaping views about the network’s performance. Our findings revealed that network politics and hamula politics competed for influence in the relationship between the use of network management strategies and perceived network performance. Perceived hamula politics had a mediating effect on this relationship. While greater use of network management strategies was related to heightened perceptions about the impact of hamula politics, fewer perceptions about hamula politics were related to stronger perceptions about the network’s performance.

We also identified the conditions under which the impact of network management strategies on perceived hamula politics occurs and how this affected perceived network performance. Perceived network politics had a moderating effect. When network members reported high levels of network politics, meaning the presence of positive and negative interactions within the network increased, the use of more network management strategies was related to reduced perceptions of the influence of hamula politics. In turn, reduced perceptions of the influence of hamula politics improved the network members’ assessment about the performance of the network. Hence, when integrated as a moderator in the mediation model, perceived network politics had a positive association with network management strategies. More interestingly, it attenuated the impact of hamula politics and had a positive association with the perceived performance of the network.

The study showed that in partially modernised contexts, external politics are more harmful to perceived network performance than internal politics. Taken as a whole, when network management is properly implemented, network politics edge out hamula politics, resulting in a more professional climate. The Etgarim Plan could be managed properly because internal politics, which is usually regarded as a barrier (Vigoda 2000), reduced the negative effects of the external hamula politics on the members’ perceptions, eventually resulting in their assessment of the performance of the Etgarim Plan as effective.

However, the mechanism of the Etgarim Plan was incompatible with the managerial reality in the Arab local authorities. As a result, hamula politics affected the perceptions of its members as creating an unprofessional climate, leading them to assess the performance of the network as less effective. The moderate score on network management and perceived performance, especially in peripheral local authorities, mirrors the relationship between the central Jewish government and the local Arab minority. This relationship reflects the national struggle between the Israeli state and the Palestinian minority in Israel. While the state devised the Etgarim Plan based on the modern values needed for network governance, the Arab local authorities found it hard to adopt this type of management because the traditional social structure of the hamula impedes the move to local governance. Members of the Etgarim Plan, especially in the peripheral areas of Israel, found it difficult to change the internal political reality and encourage more professional management in the local arena. Their struggle exemplifies the broader struggle facing Arab society in moving from traditional values to modern values.

Limitations and recommendations for future research

This study has several limitations. Methodologically, the dependent variable was assessed based on the members’ perceptions, making it a subjective estimation. To minimise the impact of common source bias, we conducted nine interviews with regional supervisors who commented on the performance of the Etgarim Plan, which represents a more objective evaluation of the dependent variable. From a theoretical perspective, we assessed perceived network politics based on perceived organisational politics. However, networks are different from organisations. Thus, it is crucial to understand how the interaction between the members within networks operates to improve our understanding of the collaboration process. Furthermore, the study focused only on the Arab minority in Israel, which might limit our ability to generalise the results. Therefore, future studies should be conducted with other minority-populated local authorities looking to democratise and modernise. Finally, the study used a relatively small sample. An identical research design with a larger sample might yield different results.

Disclosure statement

No potential conflict of interest was reported by the authors.

Appendix Appendix A:

The survey

Please indicate the degree to which you agree with the following statements on a scale of 1 (strongly disagree) to 5 (strongly agree)

Network management strategies

In the Etgarim plan:

Groups of public stakeholders are involved through platforms for negotiation and debate.

Groups of private companies are involved through platforms for negotiation anddebate.

Civil-society groups are involved through platforms for negotiation and debate.

In this project, special attention has been paid to sharing diverse points of views.

During information collection, emphasis was placed on establishing starting points and common information.

The leaders of the project consulted the people who carried it out.

The leaders of the project took into account interpersonal relationships, their biases, and how they were generated and developed

When a deadlock was reached or problems arose, the management tried to find a common ground between the conflicting interests.

In this project, explicit agreements were reached about the organisation of thecooperation (managing groups).

The agreements for this project consciously envisaged the possibility of divertingfrom the plan, in the event that it proved advantageous to do so.

Perceived network politics

Please indicate the degree to which you agree with the following statements on a scale of 1 (strongly disagree) to 5 (strongly agree)

In the Etgarim Plan:

People attempt to build themselves up by tearing others down.

There has always been an influential group in this program that no one ever crosses.

Telling others what they want to hear is sometimes better than telling the truth.

Favouritism rather than merit determines who gets ahead around here.

Rewards come only to those who work hard in this organisation.

When one of the stakeholders offers help, it will be for their personal benefit, notreally with the aim of helping.

In the Etgarim Plan, I witnessed changes in the policy that benefit some of the stakeholders and not all of the stakeholders.

In general, promotion in the Etgarim Plan is the result of good performance.

Since I have been a member of the Etgarim Plan, I have not seen that salary policy or promotion are affected by politics.

Perceived hamula politics

Please indicate the degree to which you agree with the following statements on a scale of 1 (strongly disagree) to 6 (strongly agree)

In the Etgarim Plan:

In the project, most of the actors are hired based on their personal qualifications and merits and not on their hamula (familial) affiliation.

Decision making in Etgarim is affected by hamula affiliation.

The hamula affiliation in my city affects what happens in the Etgarim Plan.

Hamulas (extended families) in my city affect what happens in the Etgarim Plan.

Hamulas (extended families) in my city affect the manager of the the Etgarim Plan.

Hamulas (extended families) in my city affect the local coordinators of the the EtgarimPlan.

Perceived network performance

Please indicate the degree to which you agree with the following statements on ascale of 1 (strongly disagree) to 5 (strongly agree)

The stakeholders in the Etgarim Plan dealt with most of the problems of limitedparticipation in informal education in the city.

In general, the stakeholders in the Etgarim Plan contributed to achieving good results.

The stakeholders suggested solutions for the limited participation of Arab youth ininformal education programs.

The stakeholders in the Etgarim Plan succeeded in offering long-term solutions for theproblem of the limited participation of Arab youth in informal education.

The Etgarim plan succeeded in giving Arab youth equal opportunities for dreamingabout a better future.

The Etgarim Plan succeeded in developing the personal and social skills of the Arabyouth who took part in the plan.

The Etgarim Plan succeeded in developing the capability of Arab youngsters to plantheir future as they want.

The Etgarim Plan succeeded in giving Arab youngsters various skills to integrateoptimally into Israeli society.

The Etgarim Plan succeeded in giving Arab youngsters various skills to enter theworkforce.

The Etgarim Plan made the youths’ parents more aware of the benefits of effectiveleisure time and its contribution to the development of life skills and educationalachievements.

The Etgarim Plan succeeded in developing the personal and collective responsibility for the Arab youths’ environment and community.