Bureaucratic structures and organizational commitment: findings from a comparative study of 20 European countries

ABSTRACT

Do civil servants in some countries have higher organizational commitment? Is there any substantial cross-national variation in the form and degree of commitment? Good governance studies show a positive link between Weberian bureaucracy and favourable macro-level outcomes. However, previous comparative research is silent regarding cross-national differences of individual bureaucrats’ attitudes and their relationship with national bureaucratic structures. Employing social exchange theory, we argue that closed civil service systems produce higher commitment in senior public officials than open systems do. Using two large data sets in 20 European countries, we find closed systems are associated with continuance and normative commitment.

KEYWORDS:

• Bureaucratic structure
• civil service systems
• Weberian bureaucracy
• organizational commitment
• social exchange theory
• comparative public administration and management

Introduction

The last two decades have seen a reappraisal of Weberian bureaucratic structures and of the significant role bureaucracy plays in shaping public policies, their implementation, and the related socioeconomic outcomes (Dahlström, Lapuente, and Teorell 2012a; Dahlström and Lapuente 2017; Evans and Rauch 1999; Fukuyama 2013; Miller 2000; Miller and Whitford 2016; Olsen 2006, 2008; Painter and Peters 2010; Rauch and Evans 2000; Rothstein and Teorell 2008). In particular, a large body of cross-national and sub-national studies show that politically autonomous and impartial bureaucratic structures (i.e. Weberian bureaucracy), where civil servants are recruited through merit examination and have tenure protection, have an empirical link with positive macro-level outcomes including socioeconomic development (Evans and Rauch 1999; Nistotskaya, Charron, and Lapuente 2015; Rauch and Evans 2000), corruption prevention (Charron et al. 2017; Dahlström, Lapuente, and Teorell 2012a), regulatory quality and entrepreneurship (Nistotskaya and Cingolani 2016), scientific productivity (Fernández-Carro and Lapuente-Giné 2016), innovation outputs (Suzuki and Demircioglu 2018), civic actions (Cornell and Grimes 2015), political legitimacy, satisfaction with government, and support for democracy (Boräng, Nistotskaya, and Xezonakis 2017; Dahlberg and Holmberg 2014; Rothstein 2009), and administrative effectiveness (Dahlström and Lapuente 2017).

While this line of good governance literature, mostly grounded in political science, has enhanced the understanding of the role of Weberian bureaucracy, we still know little about individual civil servants and how their attitudes are influenced by the administrative structure from a comparative perspective. On the other hand, a substantial body of the public management literature has examined the determinants of individual attitudes mainly within an organizational setting, not from a cross-national perspective. Despite the recent increase in attention to comparative and contextual factors in public management (Meier, Rutherford, and Avellaneda 2017; O’Toole and Meier 2015), the field is said to neglect the national characteristics of bureaucracies and a broad view of governance, assuming that ‘all states are alike’ (Milward et al. 2016, 312; Roberts 2018). In order to fill this gap, our paper aims to connect existing good governance literature with the public management literature, taking the individual-level outcomes into account.

This study examines how characteristics of national bureaucracy are associated with the degree and form of organizational commitment in senior public officials. The Weberian bureaucratic model is multidimensional, consisting of 1) a formal structure (e.g. hierarchy and specialization), 2) administrative procedures and processes (e.g. compliance to formal rules, strong emphasis on law), and 3) a personnel system (e.g. meritocratic recruitment, seniority, and tenure protection) (Dahlström and Lapuente 2017; Gualmini 2008). We focus on the personnel system, following previous studies (Cornell and Grimes 2015; Dahlström and Lapuente 2017; Evans and Rauch 1999; Nistotskaya and Cingolani 2016; Oliveros and Schuster 2018; Rauch and Evans 2000). In particular, we look at the degree to which recruitment and promotion systems of public officials are open or closed (Lægreid and Wise 2015; Dahlström and Lapuente 2017). Closed systems are characterized by the career distinctiveness of public service (Christensen 2012; Lœgreid and Wise 2007; Peters 2010; Rauch and Evans 2000). Public service careers are restricted through formalized exams, public employees enjoy lifetime tenure protection, and special labour regulations are applied to public sector employees. On the other hand, open systems feature career mobility of officials who switch between public and private sectors, more diverse and flexible access to the public sector, and less distinction between the public and the private.

To study employee commitment, we rely on the concept of organizational commitment (Meyer and Allen 1991; Meyer et al. 2012). Organizational commitment has been widely used as a significant indicator for employee attitude. According to Meyer and Allen (1991) and Wiener and Vardi (1980, 86), commitment can take three forms: (a) affective commitment (i.e. emotional attachment to organizational goals and values); (b) continuance commitment (i.e. commitment based on the side benefits and costs of leaving); and (c) normative commitment (i.e. feelings of obligation to remain with the organization). Specifically, we employ social exchange theory as a theoretical framework (Gould-Williams and Davies 2005; Gould-Williams 2007; McClean and Collins 2011; Nishii and Mayer 2009; Wayne, Shore, and Liden 1997; Van De Voorde, Paauwe, and Marc 2012).

Social exchange theory focuses on the exchange of obligations or the reciprocal relationship between employees and their organizations. Social exchange is defined as ‘voluntary actions of individuals that are motivated by the returns they are expected to bring and typically do in fact bring from others’ (Blau 1964, 93). When employees feel that their organizations are committed to them such as providing job security, training opportunities, fostering collegiality, they ‘seek a balance in their exchange relationships with organizations by having attitudes and behaviors commensurate with the degree of employer commitment to them as individuals’ (Wayne, Shore, and Liden 1997, 83). We argue that the characteristics of closed systems are more likely to produce such an exchange of obligations between organization and public officials than open systems, which leads to higher organizational commitment. Using two unique large comparative data sets in 20 European countries – the COCOPS Top Executive Survey (Hammerschmid 2015; Van de Walle et al. 2016) and the QoG (Quality of Government) Expert Survey (Dahlström et al. 2015a) – we find that closed systems are associated with continuance and normative commitment, but not with affective commitment. Thus, closed systems enhance commitment, but not all forms of commitment equally.

This paper first explains the theoretical framework for this study. The second section offers the hypotheses tested in this study while highlighting how variation in commitment is associated with characteristics of various civil service systems. The third section explains the data and methods of this study, followed by a fourth section containing results and analysis. Finally, this paper ends with discussion, conclusions, and limitations.

Weberian bureaucracy and open/closed civil service system

We focus on human resource management aspects as they are one of the core components of the Weberian model. Previous studies suggest that one significant way to characterize the personnel system of bureaucracy is to look at the extent to which systems are ‘open’ or ‘closed’ with respect to recruitment and promotion (Dahlström, Lapuente, and Teorell 2012b; Lægreid and Wise 2015). Characteristics of the open system include greater flexibility in recruitment and promotion of public officials, a high degree of job mobility between the public and private sector, a focus on selecting the best candidate for each position (i.e. position-based systems), and regulation of public organizations by general labour laws. On the other hand, a closed system entails regulated entry and mobility patterns, a low degree of public and private sector mobility, a focus on formalized entry to the public service, seniority systems, lifetime employment (i.e. career-based systems), and special labour laws that regulate public sector employees (Dahlström, Lapuente, and Teorell 2012b; Lægreid and Wise 2015; Selden 2012). By open/closed civil service systems, we mainly refer to the de facto status of HR practices rather than de jure (Dahlström and Lapuente 2017).¹ This is because previous studies report a discrepancy between law and practice in human resource management. For instance, meritocratic laws are often bypassed. Schuster (2017) points out the discrepancy between merit-based civil service management in law and practice, using a comparative data set of 117 countries.

One of the core elements of the classic Weberian bureaucracy entails meritocratic recruitment, tenure protection, the separation of public service careers from private sector careers, and a low degree of mobility between public and private organizations (Byrkjeflot, Du Gay, and Greve 2018; Evans and Rauch 1999; Gualmini 2008). Therefore, the Weberian model is more similar to the closed civil service system than open systems (Dahlström and Lapuente 2017). However, in the real world, human resource management practices in many countries have departed from the classic model, adopting more private-sector HR practices and open civil service systems (Dahlström and Lapuente 2017; Lægreid and Wise 2015). One reason for such variation is that the Weberian model has been challenged by reform efforts which reduce the distinctiveness of public service careers and make public organizations more like private organization, most clearly exemplified by New Public Management (NPM) approaches (Lægreid and Wise 2015; Gualmini 2008). Several studies have referred to the degree to which bureaucracies are open or closed (Auer, Demmke, and Polet 1996; Bekke and Meer 2000; Peters 2010). However, the data collection effort of a group of researchers at the Quality of Government Institute, Sweden, has enabled researchers to see quantification of characteristics of national bureaucracy covering over 150 countries (Dahlström et al. 2015a, 2015b), including the degrees of closedness/openness of civil service systems across countries (Please see Dahlström, Lapuente, and Teorell [2012b] for details).

Organizational commitment as work morale and attitudes

This study focuses on organizational commitment as a measurement for the relationship between employees and their organizations. Research on organizational commitment has developed significantly over the past four decades (Becker 1960; Meyer and Allen 1991; Mowday, Porter, and Steers 1982; Porter et al. 1974; Salancik 1977). Managing organizational commitment by fostering employee morale is a crucial concern for public managers, since higher levels of involvement in that organization’s activities (i.e. high level of organizational commitment) are expected to lead to positive workplace outcomes such as work effort, productivity, and performance (Moldogaziev and Silvia 2015). In fact, a substantial body of research has shown that organizational commitment is positively related to work motivation, work effort, productivity, and performance (Anderfuhren-Biget et al. 2010; Boardman and Sundquist 2009; Caillier 2013; Locke 1997; Moynihan and Pandey 2007; Wright 2004).

Results of management studies suggest that organizational commitment is multi-dimensional. According to Allen and Meyer (1990) and Meyer and Allen (1991), organizational commitment can take three forms: (a) affective commitment refers to employees’ emotional attachment to, identification with, and involvement in the organization; (b) continuance commitment refers to commitment based on the costs that employees associate with leaving the organization; (c) normative commitment refers to employees’ feelings of obligation to remain with the organization. Specifically, the nature of each form of commitment shows a significantly different effect on employee outcomes.²

A variety of theoretical explanations have been offered for why organizational commitment occurs, and numerous constructs have been examined as its antecedents. Previous research has shown that commitment can be explained by a broad range of antecedents such as motivational factors, individual factors, organizational culture, managerial level, sector, institutional context, politics and power, political environment and administrative reform, public service motivation, leadership, goal clarity and empowerment, and performance appraisal systems (Dick 2011; Moldogaziev and Silvia 2015; Moon 2000; Park and Rainey 2007; Stazyk, Pandey, and Wright 2011; Taylor 2007; Wilson 1999; Yang and Pandey 2008). However, individual and organizational variables are the main targets of scholarly interest, resulting in the failure to relate these variables to macro factors such as bureaucratic structure (Egeberg 1999). Country-level institutional factors have been relatively overlooked, although some literature has considered cultural factors in the variation of commitment (Fischer and Mansell 2009; Meyer et al. 2012; Randall 1993). Thus, we know little about these differences in the context of public organizations. In particular, the major constraints limiting cross-national studies are lack of data and resources for an extensive data collection process, and more importantly the barriers in getting access to public managers across countries (Chordiya, Sabharwal, and Goodman 2017). In particular, while several studies examine cross-national variation in public employees’ attitudes and values (Esteve et al. 2017; Fernández-Gutiérrez and Van de Walle 2019; Lapuente and Suzuki 2017; Jeannot, Van de Walle, and Hammerschmid 2018), none of the studies assess the effect of bureaucratic structure on commitment.³ Thus, despite the large volume of research on organizational commitment and Weberian bureaucracy, the two concepts have seldom been examined together.

Social exchange theory, closed civil service system, and organizational commitment

How can we explain a link between closed/open civil service systems (i.e. their HRM practices and activities) and organizational commitment? We employ social exchange theory as a theoretical framework. Organizational researchers have often used the social exchange concept to explain how an organization’s investments in human resource (HR) activities and the organizational environment will elicit positive work attitudes and behaviour (Gould-Williams 2007; Gould-Williams and Davies 2005; McClean and Collins 2011; Van De Voorde, Paauwe, and Marc 2012). According to Blau (1964), social exchange can be defined as ‘voluntary actions of individuals that are motivated by the returns they are expected to bring and typically do in fact bring from others’ (Blau 1964, 93). On this basis, employees who positively value HR activities will reciprocate through showing attitudes and behaviours that are valued by the organization (Eisenberger, Fasolo, and Valerie 1990; Gould-Williams 2007; Van De Voorde, Paauwe, and Marc 2012). These employee reactions to HR activities depend on employees’ perceptions of how committed the employing organization is to them (Eisenberger, Fasolo, and Valerie 1990; Gould-Williams and Davies 2005; Romzek 1990; Wayne, Shore, and Liden 1997). For instance, Eisenberger, Fasolo, and Valerie (1990) describe that ‘positive discretionary actions by the organization that benefited the employee would be taken as evidence that the organization cared about one’s well-being’ (1990, 51). Therefore, when employees perceive that organizations value and deal equitably with them, they will reciprocate these ‘good deeds with positive work attitudes and behaviors’ (Aryee, Budhwar, and Chen 2002, 268; Gould-Williams and Davies 2005, 4; Haas and Deseran 1981). We argue that, in closed civil service systems, public officials are likely to have such reciprocal relationships or exchanges of obligations between public officials and their organization, which leads to greater employee commitment.

HR activities in a closed system and continuance commitment

First, we expect a positive relationship between closed systems and continuance commitment. Based on social exchange theory, we argue that HR practices and activities in closed systems produce more reciprocal relationships between public organizations and officials than open systems do, which lead to greater continuance commitment. In a closed system, public officials typically spend their entire career in the public sector and invest their time and resources acquiring public-sector specific skills and knowledge. Outside employment opportunities are typically limited, thus public officials may sense fewer viable alternatives or side benefits in a closed system. On the other hand, open systems feature career mobility of public officials who switch between public and private sectors, more diverse and flexible access to the public sector, and less distinction between the public and the private.

According to Becker (1960), Allen and Meyer (1990), and Meyer et al. (2002), the magnitude and number of side benefits that employees recognize, including retirement funds and medical benefits, positively affect continuance commitment. Furthermore, the lack of employment alternatives also increases the perceived costs associated with leaving the organization. The fewer viable alternatives that employees believe are available, the stronger will be their continuance commitment to their current employer. We argue that a number of HR practices grounded in closed systems such as tenure protection, seniority rule, employee benefits such as retirement fund and superior health-care plans, and public-sector specific knowledge and skills produce a greater sense of obligation in civil servants to remain in their organization. Moreover, we also hypothesize that if employees perceive a lack of career alternatives and incurred sunk costs, they would want to give in return to their organization, leading to increased continuance commitment. Thus, we argue that HR activities and customs in closed systems are likely to enhance public officials’ perceptions of the benefits they receive from their organizations and associated costs of leaving their organization, which are returned by civil servants’ positive continuance commitment:

Hypothesis 1: A closed bureaucratic structure is positively associated with senior public sector managers’ continuance commitment.

HR activities in a closed system and normative commitment

Secondly, we expect that closed systems are also positively associated with normative commitment. Closed systems include lifelong tenure, and a high degree of internal homogeneity in employees’ educational background and professional skills. Typically, once public officials enter the public sector, they remain in the public sector for the rest of their career. Job mobility between public and private sectors is limited. On the other hand, open systems usually include more job mobility between public and private organizations, and more diversity in employees’ educational and career background.

Allen and Meyer (1990) have proposed that employees may develop a sense of moral obligation toward the organization through organizational socialization. Organizational socialization refers to the process by which individuals learn what they believe most others in the organization will actually do (Van Vugt and Hart 2004). Through this process, employees acquire knowledge about the adjustment to new jobs, roles, workgroups, and the culture of the organization in order to participate better as an organizational member (Haueter, Macan, and Winter 2003; Saks and Ashforth 1997; Cohen and Veled-Hecht 2010). We argue that closed systems are more likely to produce normative commitment than open systems do, for a number of reasons, through the organizational socialization process. First, unlike open systems, closed systems tend to have a high degree of internal homogeneity among employees, which stems from similar educational background and professional skills. Such similarity in educational and career backgrounds engenders and fosters collegiality or esprit de corps (Barberis 2011; Peters 2010). Through the socialization process, public officials in closed systems develop such esprit de corps, which leads to a normative obligation to remain in their organization. Second, closed systems tend to need a high degree of loyalty to the organization, which comes from the high degree of political loyalty expected in a traditional bureaucratic system. Therefore, through the socialization process, civil servants may produce a sense of moral obligation to the organization in closed systems.

Employing social exchange theory, we argue that if civil servants perceive this loyalty norm and a high degree of internal homogeneity in closed system they are more likely to reciprocate with a strong sense of moral obligation to the organization as a positive normative commitment. Thus, our second hypothesis is:

Hypothesis 2: A closed bureaucratic structure is positively associated with senior public sector managers’ normative commitment.

HR activities in a closed system and affective commitment

Finally, we expect that closed systems are also associated with affective commitment. Meyer and Allen (1987) have pointed out that work experiences such as organizational support or perceptions of justice provide the strongest evidence for affective commitment. Those experiences fulfil employees’ psychological needs and make them feel comfortable within the organization and competent in their work-role. We argue that the organizational support received and perceptions of justice made by public officials in closed systems are likely to produce more affective commitment via social exchange.

Formalized systems with formal rules and regulated career recruitment and promotion systems are significant components of closed systems. For instance, theoretically, in highly formalized systems, little flexibility exists in determining how a decision is made or what outcomes are due in a given situation; procedures and rewards are dictated by the rules (Schminke, Ambrose, and Cropanzano 2000, 96). We expect that this should contribute to an individual’s confidence that he/she is being treated the same as others in similar situations, as the rules are well documented and well known (ibid). Additionally, bureaucratic structures and processes with formal rules such as meritocratic recruitment are especially meant to support professionals’ autonomy (Diefenbach and Sillince 2011, 1522–1523). Thus, we can assume that employees in closed systems should have more psychological support in the organization than employees in open systems. Thus, using a social exchange framework, employees who perceive a high level of organizational support in a closed system are more likely to feel an obligation to ‘repay’ the organization in terms of affective commitment (Eisenberger et al. 1986). Thus, we expect that creating HR activities in closed systems (i.e., hierarchical bureaucratic structures with formal rules and regulated career recruitment and promotion systems) would enhance civil servants’ perceptions of being valued equitably, supported, and cared for by reciprocating with positive affective commitment. Thus, we propose the following hypothesis:

Hypothesis 3: A closed bureaucratic structure is positively associated with senior public sector managers’ affective commitment.

Data and methods

Little comparative research has been done in the study of public administration and bureaucracy (Dahlström, Lapuente, and Teorell 2012b; Eglene and Dawes 2006; Fitzpatrick et al. 2011). One reason for the scarcity has been the lack of systematic data on both bureaucratic structures and civil servants’ attitudes. This study aims to bridge this gap in the literature with two unique cross-national data sets. The first one is the COCOPS Executive Survey on Public Sector Reform in Europe (Hammerschmid 2015), which contains the survey answers of 9,333 senior public sector executives from 21 European countries. The second data set is the QoG Expert Survey Dataset II (Dahlström et al. 2015a), which captures characteristics of national bureaucratic structures constructed from the opinions of over 1,200 country experts. In this study, we combine these two data sets. Independently, both the COCOPS survey and the QoG Expert Survey data have been used in many academic publications.⁴ The empirical novelty of this study is combining these two data sets. Using these data sets is appropriate for our research question in several ways. First, the COCOPS survey was conducted to collect data on public officials’ value preferences and work attitudes, including their organizational commitment. Thus, its purpose fits our research interest. Second, the survey was ‘a response to the observed lack of rigorous quantitative comparative research into public administration reforms in Europe’ (Jeannot, Van de Walle, and Hammerschmid 2018, 5). Thus, it serves our purpose of advancing the field of empirical comparative public management. Thirdly, the QoG Expert Survey provides crucial information on a variety of bureaucratic structures across countries. Finally, by combining two different data sets, we can avoid a common source bias issue (George and Pandey 2017; Jakobsen and Jensen 2015) because our dependent and independent variables are not from the same data set.

Between 2012 and 2015 the COCOPS project administered its Executive Survey, in order to quantitatively assess the impact of NPM-style reforms in European countries (Hammerschmid, Oprisor, and Štimac 2013). The cross-national survey captured the experiences and perceptions of public-sector executives with respect to the current status of management, coordination and administrative reforms, as well as to gauge the effect of NPM-style reforms on performance, and the impact of the financial crisis. Designed by an international team of public administration researchers, it provides a comprehensive census of both central government ministries and agencies as well as state and regional administrations in the target countries while avoiding random sampling and response bias issues.⁵ Twenty-one countries were surveyed, targeting 36,892 senior-level managers; the countries included were: Austria, Belgium, Croatia, Denmark, Estonia, Finland, France, Germany, Hungary, Iceland, Ireland, Italy, Lithuania, the Netherlands, Norway, Poland, Portugal, Serbia, Spain, Sweden, and the United Kingdom. However, due to missing data, we exclude Poland, leaving 20 countries in our data set. After data cleaning, there were 9,333 responses reflecting a 25.3% response rate. Compared to similar online surveys, the response rate is relatively high. However, the sample cannot be considered as representative of the distribution of top-level officials within and among ministries and agencies. Nonetheless, as Bezes and Jeannot (2018, 7) argue, ‘the diversity of our responses is, on average, [satisfying], both for the distribution between central administrations in ministries and agencies and for policy sectors.’ In addition, the data set provides crucial source of information on understudied cross-national variation in senior officials’ attitudes and preferences. Therefore, while being aware of the above limitation, it is preferable to advance the field with currently available data rather than waiting for ideal data set.

The QoG Expert Survey offers quantitative data for the study of Weberian bureaucracy, which has hitherto been neglected by empirical analysis (Dahlström, Lapuente, and Teorell 2010). Based on the innovative work of mapping bureaucratic structures in 35 less-developed countries conducted by Peter Evans and James Rauch (Rauch and Evans 2000; Evans and Rauch 1999), the first survey was taken by researchers at the QoG Institute in 2008–2012, culminating in their first data set (Teorell, Dahlström, and Dahlberg 2011). Their second survey – the Expert Survey II – was undertaken in 2014 and received 1,294 responses from experts in 159 countries. The survey elicited experts’ perceptions of their country’s public bureaucracy, including views on bureaucratic structures such as recruitment and career systems, policies concerning replacement and compensation, procedures for policy-making and implementation, gender representation, and the level of transparency. We used information from the second QoG Expert Survey regarding employment systems in the current research.

Dependent variable

This study utilizes an organizational commitment variable created as an index from the COCOPS survey items that aims to measure public managers’ three forms of organizational commitment. Our dependent variable depends on the respondents agreement or disagreement with the following seven statements: 1) ‘I feel valued for the work I do;’ 2) ‘I would recommend it as a good place to work;’ 3) ‘I really feel as if this organization’s problems are my own;’ 4) ‘I would be very happy to spend the rest of my career with this organization;’ 5) ‘It would be very hard for me to leave my organization right now, even if I wanted to;’ 6) ‘I was taught to believe in the value of remaining loyal to one organization;’ 7) ‘Things were better in the days when people stayed with one organization for most of their career.’ Following Caillier (2013), Meyer and Allen (1991), and Moldogaziev and Silvia (2015), we measured three types of commitment, namely affective commitment from items 1–2, continuance commitment from 4, 5, 6, and 7, and normative commitment from 3. Affective commitment is a mean value of answers for survey items 1–2 (Cronbach’s alpha = 0.75), and continuance commitment are the mean values of items 4–7 (Cronbach’s alpha = 0.69). As for the normative commitment, we use the original scale (ordinal from 1 to 7) of the answer for item 3. Since none of these commitment variables take values lower than 0 or greater than 7, we treat them as truncated dependent variables.

Independent variables

We use the QoG Expert Survey Dataset II to capture how closed the employment system is. This is an aggregate measure, and this variable has been used in previous research (Dahlström, Lapuente, and Teorell 2012a; Dahlström and Lapuente 2017). Following the existing literature, this index is created based on a principal component analysis (Cronbach’s alpha = 0.75) of the following three questions: (1) ‘Public sector employees are hired via a formal examination system;’ (2) ‘Once one is recruited as a public sector employee, one remains a public sector employee for the rest of one’s career;’ and (3) ‘The terms of employment for public sector employees are regulated by special laws that do not apply to private sector employees.’ The higher the value, the more isolated, or ‘closed,’ public employees are from the practices of the private sector. Figure 1 shows the variation in the closed/open nature among our sample of 20 European countries. Countries such as France, Belgium, Italy, Croatia, Austria, Germany, and Spain have relatively closed systems, in which entry into a public service career is restricted through formalized exams, public employees enjoy lifetime tenure protection, and special labour regulations are applied to public sector employees. On the other hand, bureaucrats in countries such as Sweden, Finland, United Kingdom, Denmark, Estonia, and the Netherlands work in more open civil service systems.

Figure 1. Variations in the degree of closedness in civil service systems. Source: Created from the QoG expert survey dataset II (Dahlström et al. 2015a)

Control variables

We also control for other factors that are expected to influence organizational commitment. The relatively small number of countries (N = 20) does not allow us to include a large number of country-level controls. Therefore, we limit the number of controls to significant factors that may affect our dependent variables and test them in different models. Control variables, which are expected to influence dependent variables, are selected based on previous literature. The models include GDP per capita (2011–15 mean values, logged) as a country level control from the QoG Standard Dataset (Teorell et al. 2017). Results of previous studies suggest that levels of economic development are negatively associated with affective and normative commitment (Fischer and Mansell 2009). Previous research has also identified country culture as a significant predictor for commitment (Fischer and Mansell 2009; Meyer et al. 2012; Randall 1993). Therefore, we use national cultural factors from Hofsted’s dimension of cultural values, namely power distance and individualism-collectivism (Hofstede, Hofstede, and Minkov 2010) as additional country-level control variables. Following previous studies (Steyrer, Schiffinger, and Lang 2008; Avolio et al. 2004; Ang, Van Dyne, and Begley 2003; Mathieu and Zajac 1990), we control for the following individual-level factors that may affect commitment: gender, organizational type (ministry or agency/other), organizational size, respondent’s current position, age, years of experience in current organization, private sector experience, educational level, degree of job autonomy, organizational social capital, job satisfaction, and organizational goal clarity. All of these variables are collected or created from the COCOPS survey dataset. Table 1A in the Appendix presents descriptive statistics of all variables in the analysis. We conducted collinearity diagnostics using Variance Inflation Factors (VIF) based on our main models. Mean values of VIF are less than 1.49 in all main models (model 2 in Table 1). The highest individual VIF score for individual variables is 1.77 (age). These results suggest that the models do not have serious multicollinearity issues. Table 2A in the Appendix reports the correlation matrix.

Table 1. Multilevel models measuring senior public managers’ organizational commitment.

	DV = Continuance Commitment		DV = Normative Commitment		DV = Affective Commitment
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Independent variable
Closed bureaucracies		0.09***		0.18***		0.02
		(0.03)		(0.05)		(0.02)
Individual level controls
Female managers^a	−0.13***	−0.13***	−0.09*	−0.09*	−0.00	−0.00
	(0.03)	(0.03)	(0.05)	(0.05)	(0.03)	(0.03)
Organizational type^b =Agency or Other	0.15***	0.15***	0.15***	0.15***	−0.00	0.00
	(0.04)	(0.04)	(0.06)	(0.06)	(0.03)	(0.03)
Organizational size^c=100–999	0.01	0.02	−0.00	0.00	0.06*	0.06*
	(0.04)	(0.04)	(0.07)	(0.07)	(0.03)	(0.03)
Organizational size =over 1000	0.07	0.08*	−0.09	−0.08	0.07*	0.07*
	(0.05)	(0.05)	(0.08)	(0.08)	(0.04)	(0.04)
Respondent’s position^d	−0.00	−0.01	0.25***	0.24***	0.12***	0.12***
=Second hierarchical level in organization	(0.04)	(0.04)	(0.06)	(0.06)	(0.03)	(0.03)
=Top hierarchical level in organization	−0.06	−0.07	0.46***	0.45***	0.09**	0.09**
	(0.05)	(0.05)	(0.08)	(0.08)	(0.04)	(0.04)
Age^e=46–55	0.20***	0.21***	0.18***	0.18***	−0.04	−0.04
	(0.04)	(0.04)	(0.06)	(0.06)	(0.03)	(0.03)
=56 or older	0.49***	0.49***	0.25***	0.26***	−0.04	−0.04
	(0.04)	(0.04)	(0.07)	(0.07)	(0.04)	(0.04)
Working years at current organization^f	0.29***	0.29***	−0.02	−0.03	−0.03	−0.03
=5–20 years	(0.04)	(0.04)	(0.06)	(0.06)	(0.03)	(0.03)
=More than 20 years	0.74***	0.74***	0.06	0.05	0.01	0.01
	(0.05)	(0.05)	(0.08)	(0.08)	(0.04)	(0.04)
Private sector experience^g	−0.02	−0.01	0.14**	0.15**	−0.00	−0.01
=Less than 5 years	(0.04)	(0.04)	(0.06)	(0.06)	(0.03)	(0.03)
=More than 5 years	−0.09*	−0.08*	0.19**	0.19**	−0.03	−0.03
	(0.05)	(0.05)	(0.07)	(0.07)	(0.04)	(0.04)
Educational level^h = MA level	−0.12***	−0.13***	−0.05	−0.06	−0.01	−0.01
	(0.05)	(0.05)	(0.07)	(0.07)	(0.04)	(0.04)
=PhD/Doctoral level	−0.20***	−0.20***	−0.23**	−0.23**	0.01	0.01
	(0.06)	(0.06)	(0.10)	(0.10)	(0.05)	(0.05)
Degree of job autonomy	−0.00	−0.00	0.01**	0.01***	0.01***	0.01***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Organizational social capital	0.01***	0.01***	0.02***	0.02***	0.03***	0.03***
	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Job satisfaction	0.24***	0.24***	0.30***	0.30***	0.62***	0.62***
	(0.01)	(0.01)	(0.02)	(0.02)	(0.01)	(0.01)
Organizational goal clarity	0.04***	0.04***	0.04***	0.04***	0.06***	0.06***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Coutry level control
GDP/capita (ln)		−0.85***		−1.18***		0.07
		(0.10)		(0.20)		(0.07)
Variance of random intercept at the country level	0.45***	0.08***	1.12***	0.31***	0.05***	0.04***
	(0.14)	(0.03)	(0.36)	(0.10)	(0.02)	(0.01)
Constant	1.40***	8.88***			−0.66***	−1.78**
	(0.19)	(1.12)			(0.11)	(0.84)
Number of individuals (level1)	5,486	5,486	5,469	5,469	5,487	5,487
Number of countries (level 2)	20	20	20	20	20	20

Standard errors in parentheses.

*** p < 0.01, ** p < 0.05, * p < 0.1

a: ref. = male executives, b: ref. = Ministry, c: ref. < 100, d: ref. = third hierarchical level, e: ref. = 45 or less, f: ref. = less than 5 years, g: ref. = none, h: ref. = BA level

Table 1A. Descriptive statistics.

	Obs	Mean	Std.Dev.	Min	Max
Dependent variables
Continuance commitment	5,486	4.15	1.36	1	7
Normative commitment	5,468	4.79	1.76	1	7
Affective commitment	5,486	5.23	1.35	1	7
Independent variable
Closed bureaucracy	5,486	15.16	2.53	10.84	18.82
Individual level controls
Female managers	5,486	1.38	0.48	1	2
Organizational type	5,486	0.53	0.50	0	1
Organizational size	5,486	1.04	0.70	0	2
Respondent’s position	5,486	0.90	0.75	0	2
Age	5,486	1.02	0.77	0	2
Working years at current organization	5,486	0.91	0.72	0	2
Private sector experience	5,486	0.95	0.69	0	2
Educational level	5,486	0.94	0.57	0	2
Degree of job autonomy	5,486	29.68	10.84	8	56
Organizational social capital	5,486	43.46	10.69	9	63
Job satisfaction	5,486	5.67	1.28	1	7
Organizational goal clarity	5,486	10.94	2.77	2	14
Country level control
GDP/capita	5,486	39,060.38	20,762.71	5,974.79	95,307.31

*Sample is based on model 2 for continuance commitment in Table 1.

Table 2A. Correlation Matrix.

		1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
1	Continuance commitment	1
2	Normative commitment	0.41	1
3	Affective commitment	0.24	0.20	1
4	Closed bureaucracy	0.06	0.15	−0.13	1
5	Female managers	−0.02	−0.03	−0.02	−0.10	1
6	Organizational type	0.08	−0.01	0.06	−0.35	0.07	1
7	Organizational size	−0.03	−0.05	0.01	−0.01	−0.06	−0.03	1
8	Respondent’s position	0.01	0.14	0.12	0.02	−0.11	0.10	−0.05	1
9	Age	0.15	0.00	0.08	0.02	−0.10	0.01	0.00	0.15	1
10	Working years at current organization	0.25	−0.02	0.01	0.01	−0.02	0.03	0.09	−0.09	0.29	1
11	Private sector experience	−0.07	−0.02	0.01	−0.14	−0.01	0.12	−0.04	0.02	0.01	−0.15	1
12	Educational level	−0.05	−0.07	0.00	0.09	−0.05	0.02	−0.02	0.07	0.05	0.01	−0.05	1
13	Degree of job autonomy	−0.10	−0.02	0.32	−0.22	−0.10	0.13	0.07	0.36	0.08	−0.08	0.08	0.10	1
14	Organizational social capital	0.17	0.12	0.55	−0.18	−0.06	0.05	−0.08	0.07	0.09	0.05	0.04	0.01	0.28	1
15	Job satisfaction	0.26	0.21	0.72	−0.13	0.01	0.04	0.00	0.12	0.07	0.02	0.00	−0.03	0.26	0.48	1
16	Organizational goal clarity	0.19	0.13	0.44	−0.19	0.01	0.20	−0.05	0.09	0.07	0.03	0.04	−0.04	0.24	0.47	0.39	1
17	GDP/capita	−0.34	−0.36	0.15	−0.05	−0.10	−0.06	0.12	−0.06	0.17	0.02	0.08	0.07	0.29	0.18	0.09	0.04	1

*Sample is based on model 2 for continuance commitment in Table 1.

Empirical strategy

Our dataset has a hierarchical structure, with public sector managers (level 1) nested in country-level factors (level 2), and thus multilevel analysis seems to be an appropriate method (Jones 2008). We assume that intercepts of individual-level variables can vary across countries due to the country-level factors, therefore a random intercept model is applied. Since the main dependent variables are censored continuous variables of mean values of survey items ranging from 1 to 7, we employ multilevel mixed-effects tobit regression models. As for the normative commitment variable, we utilize multilevel-ordered logit models as the variable is in ordinal form from 1 to 7. The first model includes only individual-level independent and control variables. The second model adds a country-level control variable and the independent variable. In our robustness check models (models 3–8), we use a different set of country-level controls, including power distance and individualism (from Hofsted’s dimension of cultural values), competitive salary in the public sector, administrative burden, corruption perception index (public officials/civil servants), and polity score.

Descriptive results: cross-national variation in organizational commitment

We present cross-national variation in levels and forms of commitment prior to the main regression results. Figure 2 shows the variation in country mean values of all three types of commitment across our sample countries. It is notable that countries ranked high or low in terms of commitment values differ depending on types of commitment. For instance, Hungary, Serbia, Lithuania, Portugal, Belgium, Italy, and Croatia score high in continuance commitment. On the other hand, the Netherlands, Denmark, Sweden, Norway, and UK score low. Country orders in the normative commitment show a similar trend. Affective commitment seems to have a negative relationship with continuance and normative commitment (correlation coefficients between CC and AC = −0.58 with p = 0.008 and CC and NC = −0.59 with p = 0.007). Norway, Sweden, Finland, Denmark, and Iceland show higher affective commitment, while Croatia, Hungary, Estonia, and Spain exhibit lower commitment. These results suggest that our sample exhibits cross-national variation in levels and forms of commitment.

Figure 2. Country comparison of mean values of organizational commitment by commitment type. Source: Created from the COCOPS Executive Survey on Public Sector Reform in Europe (Hammerschmid 2015)

Having shown the variation, we now turn to results of the multilevel analysis. Results of the multilevel tobit regression models are reported in Table 1 for the three types of commitment. Results suggest that closed bureaucracies are positively and statistically significantly associated with continuance commitment and normative commitment. However, the link between closed bureaucracy and affective commitment is positive but not statistically significant. In closed systems, senior public officials show higher level of continuance and normative commitment than those in more open systems, holding other factors fixed. Thus, senior public officials in more closed systems show more commitment to their organizations based on the side benefits and costs of leaving and feelings of obligation to remain with the organization than those in more open systems. This confirms our hypotheses 1 and 2, but not hypothesis 3. It is notable that results show that all forms of commitment are not associated with the closed systems, only continuance and normative commitment. In fact, this result is consistent with existing literature that attitudes to continue membership of an organization for economic benefits or out of norm-based obligations may not necessarily coincide with higher affective commitment or loyalty towards the organization (Chordiya, Sabharwal, and Goodman 2017; Solinger, Van Olffen, and Roe 2008; Stazyk, Pandey, and Wright 2011). Results of our robustness check models (see Table 2) also show the same results (positive impact of closed systems on normative and continuance commitment, but not affective commitment).

Table 2. Results from multilevel model estimates: using additional country-level controls.

	Continuance Commitment						Normative Commitment
	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Independent variable
Closed bureaucracies	0.11***	0.09***	0.09***	0.09***	0.13***	0.10***	0.14**	0.19***	0.17***	0.22***	0.15**	0.17***
	(0.03)	(0.03)	(0.03)	(0.03)	(0.03)	(0.03)	(0.06)	(0.05)	(0.05)	(0.06)	(0.07)	(0.06)
Country level controls
GDP/capita (ln)	−0.99***	−0.88***	−0.85***	−0.85***	−1.04***	−0.94***	−0.87***	−1.38***	−1.19***	−1.36***	−1.19***	−1.13***
	(0.13)	(0.13)	(0.10)	(0.12)	(0.14)	(0.11)	(0.27)	(0.24)	(0.19)	(0.21)	(0.29)	(0.23)
Culture: power distance	−0.01						0.02
	(0.01)						(0.01)
Culture: individualism		0.00						0.01
		(0.01)						(0.01)
Public sector salary			−0.03						0.21
			(0.09)						(0.17)
Administrative burden				0.03						0.53**
				(0.15)						(0.27)
Corruption perception:					−0.41*						−0.11
public officials/civil servants					(0.22)						(0.45)
Polity score						0.21*						−0.04
						(0.11)						(0.24)
Variance of random intercept	0.07***	0.08***	0.08***	0.08***	0.07***	0.07***	0.29***	0.29***	0.29***	0.27***	0.29***	0.32***
at the country level	(0.02)	(0.03)	(0.03)	(0.03)	(0.02)	(0.02)	(0.10)	(0.10)	(0.10)	(0.09)	(0.10)	(0.11)
Constant	10.48***	9.05***	8.94***	8.63***	11.60***	7.62***
	(1.48)	(1.21)	(1.13)	(1.48)	(1.86)	(1.22)
Number of individuals (level1)	5,343	5,343	5,486	5,343	5,146	5,343	5,326	5,326	5,469	5,326	5,130	5,326
Number of countries (level 2)	19	19	20	19	19	19	19	19	20	19	19	19
	Affective Commitment
	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
Independent variable
Closed bureaucracies	0.00	0.02	0.03	0.01	0.04	0.02
	(0.02)	(0.02)	(0.02)	(0.02)	(0.03)	(0.02)
Country level controls
GDP/capita (ln)	0.22**	0.16*	0.08	0.15*	0.03	0.09
	(0.09)	(0.09)	(0.07)	(0.08)	(0.11)	(0.08)
Culture: power distance	0.01**
	(0.00)
Culture: individualism		−0.00
		(0.00)
Public sector salary			−0.11*
			(0.06)
Administrative burden				−0.15
				(0.10)
Corruption perception: public officials/civil servants					−0.05 (0.17)
Polity score						−0.01
						(0.09)
Variance of random intercept at the country level	0.03***	0.04***	0.04***	0.04***	0.04***	0.04***
	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Constant	−3.30***	−2.21***	−1.56**	−0.89	−1.41	−1.75*
	(1.03)	(0.84)	(0.79)	(1.01)	(1.42)	(0.97)
Number of individuals (level1)	5,344	5,344	5,487	5,344	5,147	5,344
Number of countries (level 2)	19	19	20	19	19	19

Standard errors in parentheses.

*** p < 0.01, ** p < 0.05, * p < 0.1

Individual controls are not reported.

Results of analysis also offer other significant results that deserve attention. Being a female manager, compared to a male manager, is negatively associated with continuance commitment (p < 0.01) and normative commitment (p < 0.1). Compared to male officials, female officials are likely to exhibit lower continuance and normative commitment. However, we do not find any statistically significant link between gender and affective commitment. A previous meta-analysis using all correlations indicated that women tend to be more committed than men, while it illustrated a slightly stronger relationship between women and affective commitment (Mathieu and Zajac 1990). Our present study shows no statistically significant association between women and affective commitment. Organizational type also affects the degree of commitment. Compared to senior officials working for a ministry, those in an agency or other types of organization tend to show more commitment to their organization in the form of continuance and normative commitment (p < 0.01), holding other factors constant. However, the level of affective commitment is not statistically associated with types of organizations. Age also has positive link with continuance and normative commitment (p < 0.01). Previous meta-analysis shows that age was significantly more related to affective commitment than to normative commitment (Mathieu and Zajac 1990). In our study, older officials tend to have more commitment based on side benefits or normative obligation to remain in the organization than younger officials. The positive relationship between working years at the current organization and continuance commitment is also found (p < 0.01). The longer officials work for the same organization, the more continuance commitment they are likely to exhibit. This is consistent with previous studies as organizational tenure tended to be more positively related to normative commitment (Mathieu and Zajac 1990). In our study, years spent in the same organization are likely to yield greater side benefits, such as a pension plan, and develop greater continuance and normative commitment. With respect to private sector experience of managers, those managers having less than 5 years of private sector experience or more than 5 years of private sector experience exhibit higher normative commitment than those without private sector experience (p < 0.01).⁶ Education level is negatively associated with continuance and normative commitment. Compared to senior officials who completed a bachelors-level degree, those who completed master-level education show lower continuance commitment (p < 0.01). Those who completed PhD-level education exhibit lower continuance (p < 0.01) and normative commitment (p < 0.05). Therefore, education negatively affects these forms of commitment. This inverse relationship result is consistent with previous meta-analysis results (Mathieu and Zajac 1990). In this research, we found that more educated civil servants would have higher expectations that the organization may be unable to meet and have a greater number of job options, and are less likely to become entrenched in any one position or organization (Mowday, Porter, and Steers 1982; Mathieu and Zajac 1990; Meyer and Parfyonova 2010). Organizational factors such as organizational social capital and goal clarity have a positive association with all forms of commitment (p < 0.01). Also, job autonomy is positively related to normative and affective commitment (p < 0.01) and job satisfaction has a positive link with all forms of commitment (p < 0.01).

Visualizing predicted probabilities help to interpret the results of the multilevel-ordered logit model. Figure 3 illustrates the predicted probabilities for outcomes 1 (strongly disagree)-7(strongly agree) concerning normative commitment (model 2 in Table 1). As seen from the figure, the probability of selecting outcome 7 or 6, which shows greater normative commitment, increases as the degree to which bureaucracies are closed increases. The probability of selecting outcome 7 is above 30% when the closedness value is higher than 17, which is around Spain’s closedness value (17.04). The probability of picking low levels of normative commitment such as outcomes 1–3 are relatively higher when the closedness bureaucracy variable is low (i.e. open civil service systems). On the other hand, these probabilities drop as the degree of closedness increases. In other words, senior public officials in more closed systems are less likely to show lower normative commitment than those in more open systems. This gives empirical support for H2. Figure 4 shows predicted values of continuance commitment as the degree of closed service systems varies. Recall that continuance commitment is a censored continuous variable, that is created from separate survey items, ranging from 1 to 7. These values are calculated based on model 2 in Table 1. Holding other factors at mean, a higher degree of bureaucratic closedness increases senior public managers’ continuance commitment. Senior public officials working in more closed civil service systems are more likely to have higher continuance commitment than those working in more open systems.

Figure 3. Predicted normative commitment by degree of closed bureaucracy*.

*Samples are based on model 2 for normative commitment in Table 1.

Figure 4. Predicted continuance commitment by degree of closed bureaucracy^*.

*Samples are based on model 2 for continuance commitment in Table 1.

In sum, the results of our multilevel models show that characteristics of employment systems of civil servants matter for levels of some types of organizational commitment even after controlling for several significant country-level factors such as culture, administrative conditions, and political factors. There are strong empirical positive associations between closed systems and higher levels of continuance and normative commitment. However, we do not find any empirical link between closed systems and affective commitment.

We ran robustness check models with additional control variables. Those additional variables include country cultures (power distance and individualism from Hofsted’s dimension of cultural values), competitive salary in the public sector, administrative burden, corruption perception index (public officials/civil servants), and polity score. The coefficients of closed bureaucracy are robust across models with different country controls as seen in Table 2. A closed bureaucracy has a positive empirical association with continuance commitment (p < 0.01) and normative commitment (p < 0.01 or p < 0.05), holding other variables at mean. Affective commitment is not statistically significantly associated with closed bureaucracy. These results demonstrate the robustness of our results.

Discussion and conclusions

Bureaucratic structure plays a large role in shaping public policies, their implementation, and the related socioeconomic outcomes. The effect of the different characteristics of civil service systems on employee work attitudes is an important empirical question. This study presents evidence on this broad question using cross-national comparative data sets. We have argued that closed civil service systems are associated with higher levels of continuance, normative, and affective commitment via social exchange. Senior public managers in a more closed bureaucracy should show higher levels of continuance commitment due to the perceived costs associated with leaving and the number of alternative options outside the public sector. Public officials in closed systems should also have greater normative commitment because they feel that organizations expect their loyalty and because of the high degree of internal homogeneity. Finally, closed systems should be also associated with greater affective commitment because of higher perceived organizational support, justice, and equal treatment. After controlling for significant individual and country level factors, results of multilevel analysis suggest that only continuance and normative, but not affective commitment, are associated with the type of civil service system.

Results of this study suggest the importance of looking at individual bureaucrats for broader outcomes. As results of this study show, civil servant work morale varies not only among individuals but also among countries with different bureaucratic structures. Civil servants in closed systems are more committed to their organization than those in more open systems. However, the types of commitment are not the same. Bureaucrats in closed systems have higher continuance and normative commitment than those in the open systems. What does this result imply? Although our current studies cannot directly connect these commitment differences with macro-level outcomes, future study should explore consequences of organizational commitment in more macro-outcomes. Results reported in previous good governance studies suggest that the Weberian bureaucratic structure is one of the primary predictors for favourable country-level outcomes. How does organizational commitment of civil servants play a role in these outcomes? As Meyer and Parfyonova (2010) argue, benefits of organizational commitment are not equal among different types of commitment. Empirical results of some studies suggest a positive link between affective commitment and perceived performance and quality of work (Park and Rainey 2007), organizational ethical climate (Erben and Güneşer 2008), whistle-blowing attitudes when combined with transformational leadership (Caillier 2015), and innovative attitudes (Jafri 2010; Xerri and Brunetto 2013). On the other hand, previous studies have not yet found a strong link between continuance or normative commitment and good outcomes such as innovativeness, ethical behaviour, and acceptance of organizational change. Future studies should undertake to discover how different levels and types of commitment of bureaucrats are connected to country-level differences such as innovativeness, government effectiveness, and corruption level.

Results of this study also shed light on potential positive aspects of closed civil service systems. Previous studies have shown that closed civil service systems are not linked with favourable outcomes such as government effectiveness, lower levels of corruption, and innovative administration (Dahlström and Lapuente 2017). In addition, reform efforts, particularly those associated with the New Public Management approach, have further cast doubt on the effectiveness of the traditional civil service system, pushing to reduce the distinction between public and private careers and to make public organizations more like private ones. However, results of our study show that civil servants in more closed systems have higher levels of commitment to their organization than those in more open systems. As we discussed, organizational commitment is an important determinant of individual performance and an organizations’ success and well-being. Therefore, it is reasonable to deduce from our study that having a sense of working under hierarchical bureaucratic structures with formal rules, tenure protection, seniority, regulated career recruitment and promotion systems would strengthen civil servant’s commitment to the organization (especially, higher continuance and normative commitment as suggested by the current study) which, in turn, could lead to greater job performance and government effectiveness. Thus, while the Weberian model of bureaucracy has been challenged by the NPM reform efforts, it is still important to reassess positive aspects of the classic model in the NPM and post-NPM era.

There are, of course, limitations associated with our study. First, due to the unavailability of data, our study does not consider intra-country variation. Even within the same country, hiring and promotion practices may differ across government organizations. Although we include several organization-related variables such as organizational size, organizational level of social capital, and organizational goal clarity in our model, the COCOPS data set does not allow us to implement three-level multilevel analysis (country-organization-individual). Thus, future studies should be conducted to consider such intra-country variations and organizational-level differences as data becomes available. Secondly, our study focuses on top-level executives, not street-level or mid-level bureaucrats. The association between closed systems and commitment that we have identified might be different for lower level civil servants. Therefore, we stress that results of this study cannot be generalizable for different levels of bureaucrats. Thirdly, survey response rate of the COCOPS survey and country selection are also a limitation of this research. Our samples only include European countries. Even though these European countries still have variations as shown in the literature on the public administration tradition (Painter and Peters 2010), obviously bureaucratic structures that are grounded in other traditions such as East Asia, Latin America, or Middle Eastern countries are not included in our analysis. We do not claim generalization from our findings.

Relatedly, we advise that future research should examine the relevance of potential mediator or moderator variables on the relation between civil service systems and organizational commitment. Specifically, one may wonder whether the moderating effects of an employment system of a national bureaucracy and organizational commitment would have been even stronger if the data which was collected had included information such as professional identification, lack of employment alternatives, and organizational socialization. Hence, future research might explore additional contextual factors in these relationships. Finally, future research should consider sector differences in employees’ commitment. We built our theoretical framework based on social exchange theory and organizational commitment, which are grounded in the literature of private sector management. However, as previous research suggests, employee organizational commitment differs between private and public employees depending on the organizational context and environment (Baldwin 1990; Boyne 2002; Fletcher and Williams 1996; Markovits et al. 2010; Odom, Randy Boxx, and Dunn 1990; Rainey 2014). Employees in the public and private sectors experience different working conditions and employment relationships. Therefore, theories built on the private sector management research may need some modification in future research. Despite these limitations, this study shows how senior public managers’ commitment and their work morale can be associated with bureaucratic structure. As Van de Walle et al. (2016) argue, scholars are still in the early stages of data collection efforts for comparative bureaucratic research. Future research should undertake these abovementioned tasks as the data becomes available.

Acknowledgments

We would also like to thank Carl Dahlström and Victor Lapuente at the Quality of Government Institute for their valuable and helpful comments on our earlier drafts of our manuscript. We also received helpful suggestions from other members of the Quality of Government Institute and panel discussants and participants at the annual conferences of American Political Association, Midwest Political Science Association, European Consortium for Political Research, and International Research Society for Public Management. We also thank three anonymous reviewers for their comments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was conducted as a part of the research project ‘Out of Control or Over Controlled? Incentives, Audits and New Public Management’ (Riksbankens Jubileumsfond [the Swedish Foundation for Humanities and Social Sciences], Grant No. SGO14-1147:1). We gratefully acknowledge their financial support.

Notes on contributors

Kohei Suzuki

Kohei Suzuki is an assistant professor at Institute of Public Administration, Leiden University, the Netherlands. He obtained his Ph.D. in Public Policy from the School of Public and Environmental Affairs at Indiana University, Bloomington. He mainly studies administrative reform and bureaucratic structures with a focus on cross-national settings and Japanese local municipalities. Please see his website for his recent work.

Hyunkang Hur

Hyunkang Hur is an assistant professor at the Department of Public Administration and Health Management, Indiana University, Kokomo. He obtained his Ph.D. in Public Affairs from the School of Public and Environmental Affairs at Indiana University, Bloomington. His research focuses on human resource management and organizational commitment of civil servants.

Notes

1. We appreciate an anonymous reviewer for pointing this out.

2. Previous meta-analytic reviews demonstrate that affective and normative commitment are positively correlated with job performance and organizational citizenship behaviour, while continuance commitment is negatively correlated with job performance and unrelated to organizational citizenship behaviour (Meyer et al. 2002). Additionally, in terms of turnover, absenteeism, and stress and work-family conflict consequences, the correlations between the three commitment scales and turnover were all negative, while only affective commitment was found to be negatively related to absenteeism; however, continuance commitment is positively correlated with absenteeism, stress, and work-family conflict (Meyer et al. 2002). Thus, results of previous meta-analyses suggest that affective commitment is most strongly tied with job performance and other employee outcomes followed by normative and continuance commitment in turn. In particular, continuance commitment has often been found to be unrelated or negatively related to those employee behaviour outcomes (Meyer et al. 2002).

3. A study by Andrews, Hansen, and Huxley (2018) examines organizational commitment across European countries; however, their focus is effects of private sector experience on commitment, not the relationship between bureaucratic structures and commitment.

4. For the COCOPS survey, see, for example, Andrews (2017), Andrews, Beynon, and Aoife (2019), Bezes and Jeannot (2018), Hammerschmid, Van de Walle, and Stimac (2013), Hammerschmid et al. (2016), Ongaro, Ferré, and Fattore (2015), Raudla et al. (2015), Raudla et al. (2017), Van der Voet and Steven (2015). For the QoG Expert Survey, see Boräng, Nistotskaya, and Xezonakis (2017), Charron, Dahlström, and Lapuente (2016), Cho et al. (2013), Cornell (2014), Cornell and Grimes (2015), Dahlström and Lapuente (2017), Fernández-Carro and Lapuente-Giné (2016), Gustavson and Aksel (2016), Kopecký et al. (2016), Nistotskaya and Cingolani (2016), Sundell (2014), Schuster (2016), Suzuki and Demircioglu (2018), Van de Walle et al. (2016), Versteeg and Ginsburg (2017), Van de Walle, Steijn, and Jilke (2015).

5. Within central government ministries, the top two administrative levels are included in the target. For central government agencies, the survey targets the first two executive levels. State-owned enterprises and audit courts are excluded. Appropriate regional and state government ministries and agencies are included in order to maximize the number of senior executives reached. However, local government bodies and local service delivery organizations are not included (Hammerschmid, Oprisor, and Štimac 2013).

6. This result is not comparable with the results of Andrews, Hansen, and Huxley (2018) because their work operationalizes the ratio of private sector experience to public sector experience and also includes different control variables.

Appendix