Leading frontline enforcers: how supervisors’ leadership style impacts inspectors’ enforcement style

ABSTRACT

This article analyzes the relation between leadership style of managers and the enforcement style of street-level bureaucrats. We also studied the influence of organizational culture. The analysis is based on a survey among 549 inspectors of the Food and Consumer Product Safety Authority (NVWA) in The Netherlands. Studying transactional leadership and servant leadership the findings show that contrary to the general assumptions in leadership literature the influence of leadership style on enforcement behaviour of inspectors is only very limited. Organization culture has more influence on how inspectors enforce rules in their interactions with inspectees.

KEYWORDS:

• Leadership styles
• street-level bureaucrats
• enforcement styles
• organizational culture
• inspectors

Introduction

Inspectors, just like police officers, nurses and social workers, are street-level bureaucrats who implement public policies during face-to-face encounters with inspectees such as citizens and various organizations (restaurants, schools, companies etc.) As street-level bureaucrats inspectors have considerable autonomy and discretion in their daily implementation practice (Lipsky 1980). The nature of inspection work, however, differs from that of some other street-level bureaucrats because they are often out in the field, visiting schools for days at the time, or being based, for example, in slaughter houses. Moreover, inspectors are powerful bureaucrats (Raaphorst 2018) because (1) they limit the freedom of inspectees by sanctioning non-compliant behaviour; (2) unlike encounters with more service-oriented bureaucrats such as social workers there is not an exit option for inspectees when they encounter an inspector and; (3) this interaction is often, from the inspectees-side, unwanted (Nielsen 2015; Sparrow 2000; May and Winter 2011).

Since inspectors have discretion to perform their tasks, they develop varying enforcement styles to conduct inspections and assess inspectees’ compliance (May and Winter 2011; Lo, Fryxell, and Van Rooij 2009; De Boer 2019). There is growing scholarship on how inspectors enforce and whether their enforcement style can help secure compliance (de Boer, 2019; Lo, Fryxell, and Van Rooij 2009; May and Winter 1999). Empirical studies have shown several situational and organizational characteristics that influence enforcement style of street-level bureaucrats like the disclosure of performance information (De Boer, Eshuis, and Klijn 2018), social pressures (Lo et al, 2009) the influence of peers (Maynard-Moody 2003; Raaphorst 2018) and hierarchical forms of control (Loyens 2013). There is, however, little research addressing how leadership style of their direct supervisor influence enforcement style (see Keulemans and Groeneveld 2019; Gassner and Gofen 2018 for notable exceptions).

This is surprising because supervisors are important in the work of inspectors because they hold inspectors to account (Thomann, Hupe, and Sager 2017) and are tasked with steering the daily work of inspectors (Loyens 2013). The literature about leadership is vast, as is the number of leadership styles and types that have been elaborated in the leadership literature (Van Wart 2012). The literature distinguishes many leadership qualities, both behaviours and traits, that are considered important (see Van Wart 2012; Zehndorfer 2014). Regardless, virtually all leadership literature emphasizes that leadership style is important for the way employees’ function at their daily work (see Bass 1985; Avolio et al, 1991; Zehndorfer 2014). This importance is echoed by leadership literature that focuses on the public sector (see Van Wart 2012; Tummers and Knies 2013; Ricard, Klijn, Lewis, and Ysa 2017; Keulemans and Groeneveld 2019). The main difference in the literature is in what type of activities are seen as most effective in supporting or directing employees and achieving better performance.

In this article, we study how enforcement styles of inspectors are influenced by the leadership style of their direct supervisor. Thus, our research question is: How are enforcement styles of inspectors influenced by the leadership style of their direct supervisor? To answer our research question we conducted a survey among inspectors (549 respondents) of the Netherlands Food and Consumer Product Safety Authority (NVWA). The NVWA is the largest inspectorate in The Netherlands and the inspectors are concerned with food and product safety and perform on the spot inspections of organizations such as restaurants and fish and meat industries.

Given the specific situation of our case, namely that the inspectors of the NVWA perform their task in a fairly legalized and bureaucratic context, we theoretically embedded our study in two leadership styles: the transactional leadership style and the servant leadership style. We elaborate on this choice further in the theoretical section. A second important factor influencing enforcement styles, that also influences leadership styles of supervisors, is organizational culture. Since inspectors in our research work in a fairly bureaucratic setting in the public sector we focused on two main dimensions of organizational culture, namely the bureaucratic and developmental dimensions.

Section 2 builds the theoretical argument about the relation between leadership style and enforcement style and the relation with organizational culture. Section 3 elaborates the research design. Section 4 discusses the results and we finish with conclusions and reflections.

Theoretical framework

Street-level bureaucrats, on the one hand, have discretionary freedom in the implementation of rules and regulation because rules do not match complex realities within which bureaucrats work (see Lipsky 1980; Maynard-Moody 2003). On the other hand, street-level bureaucrats are plagued by limited time and resources and need to find ways to equitably implement the rules for concrete cases (Lipsky 1980; Maynard-Moody 2003; Tummers and Knies 2013). As a response, street-level bureaucrats develop strategies for the decisions they must make about, for instance, enforcement, that are suitable to the specific situation of clients or ‘inspectees’ and the organizational constraints they face (Kagan 1994; May and Winter 1999; 2011). In other words, street-level bureaucrats’ make decisions within the boundaries of existing bureaucratic structures (Lipsky 1980). As Cohen (2018) states ‘organizational conditions have a significant effect on the parameters of [street level bureaucrats’] choices’ (p. 176).

Within the organizational structures the relationship between street-level bureaucrats and other individuals plays a prominent role in the executing or their tasks (e.g. De Boer and Eshuis 2018; Maynard-Moody 2003; Raaphorst 2018). Maynard-Moody (2003) emphasizes that bureaucrats ‘define their work and to a large extent themselves terms of relationships more than rules’ (p. 20). An important relationship is with their supervisors, even though street-level bureaucrats operate with a certain distance, granted by their autonomy and discretion, from their supervisors (Keulemans and Groeneveld 2019). Regardless, we know surprisingly little about how exactly supervisors matter to the way street-level bureaucrats implement public policies.

That is remarkable since both general leadership literature (see Bass 1985; Avolio et al, 1991; Zehndorfer 2014) and leadership literature about the public sector (Van Wart 2012; Javidan and Waldman 2003; Keulemans and Groeneveld 2019) emphasize the importance of leadership for the behaviour of employees. However, it may be that leadership works differently in the case of inspectors than other street-level bureaucrats. Although inspectors certainly show similar characteristics if we compare their situation with other street-level bureaucrats (e.g. implementation of central rules, contact with clients, discretion in their work) there are also some striking differences. First of all, many inspectors have regular contact with clients, which is similar to other street-level bureaucrats. However, for inspectors, the encounters mostly take place outside the inspectorate. In that sense inspectors as ‘regulation-focused’ civil servants, busy with enforcing rules, for example, on food and product quality, show strong resemblance with street-level bureaucrats like policemen, but less with ‘service-focused’ civil servants such as social security officers. Just as policemen, inspectors deliver obligations, translated in fines and sanctions rather than services (see Sparrow 2000). But the fact that inspectors operate out in the field in direct contact with inspectees could also result in a less prominent role of their organizational background and specifically leadership. This makes our research more interesting as we now can explore whether there is any differences between inspectors and what is known from the literature on other street-level bureaucrats where research already has shown the importance of leadership (see Javidan and Waldman 2003; Tummers and Knies 2013; Miao et al. 2014).

Enforcement styles of inspectors: a typology

The way inspectors behave and use their discretion during enforcement has been studied through the concept of enforcement style. Enforcement style is defined as ‘the character of the day-to-day interactions of inspectors when dealing with representatives of regulated entities’ (May and Wood 2003, 119). In earlier research enforcement style was conceptualized as being composed of formalism and coercion (see for instance May and Winter 1999, 2011). Formalism means that inspectors apply the rules strictly and coercion refers to the degree inspectors threaten with sanctions. Later research has shown street-level enforcement style is multi-dimensional. De Boer (2019) showed that street-level enforcement style of Dutch inspectors is composed of three dimensions. The first dimension was conceptualized as legal, and includes both the legal and formalism aspect. The second dimension was conceptualized as facilitation and captures the behaviour of inspectors to communicate to inspectees about the rules and how they are applied. The last dimension was conceptualized as accommodation. This dimension has a more cognitive character and refers to the way opinions of peers, are considered when the inspector conducts the inspection visit. Important peer groups are the own inspection team, other inspections teams, and inspectors’ supervisors. Earlier research shows that the opinion of the above-mentioned peer groups forms one dimension of inspectors’ enforcement style (see also De Boer 2019) Together these three dimension characterize an enforcement style of an inspector and the way he/she behaves during inspection visits. In this article we use these dimensions to analyse the enforcement style of inspectors (De Boer 2019).

Enforcement styles can be influenced by a wide variety of conditions. In this theoretical framework, we highlight two elements that are central in our research. First the leadership style of the superior of the inspector. Given the vast literature that has shown how important leadership is for the behaviour of employees (Bass 1985; Van Wart 2012; Zehndorfer 2014), including leadership literature about the public sector (Van Wart 2012; Tummers and Knies 2013), we can expect that leadership also impacts on behaviour of street-level bureaucrats (see also Gassner and Gofen 2018; Keulemans and Groeneveld 2019). However, it is also possible that inspectors, in particular, are quite different from other types of public officials because they are more out in the field (conducting relatively less desk-work), have direct contact with inspectees, and their work is strongly characterized by rules and regulations (see Nielsen 2015; Sparrow 2000; May & Winter 2011). After all, inspectors are members of a highly bureaucratic public organization that is strongly regulated and motivated by public rules (Lipsky 1980). For that reason, we also discuss as an additional variable organization culture.

Factor 1: leadership style of supervisor, transactional or servant leadership

There is no doubt in the vast leadership literature on one thing: leaders affect our lives, they are important in achieving performance of (public) organizations, and they crucially influence behaviourthe of employees in organizations (see Bass 1985; Van Wart 2012; Zehndorfer 2014). In the leadership literature of the past decades, the most used distinction is probably the one between transactional and transformational leadership (Avolio, Waldman and Yammarino, 1991; Van Wart 2012). In the perspective of transactional leadership, leadership is not only fairly top down but also transactional in nature. It focuses on the reward structure in the relation between leader and employee/follower (Van Wart 2012). Leaders motivate people in organizations by payments and other reward systems (Zehndorfer 2014). Monitoring performance of employees is then crucial to implement this form of leadership. Thus, in the transactional leadership perspective leaders/managers mostly rely on rational incentives. Literature argues that transactional leadership is most suited for what some authors call ‘normal performance’ (see Avolio, Waldman and Yammarino, 1991; Howell and Avolio 1993). Some argue that transactional leadership is something like the necessary basic job which has to be complemented by other forms of leadership. Usually the leadership literature contrast transactional leadership against transformational leadership where realizing change and innovation through stimulation, charismatic leadership and motivating employees are the core (see Avolio, Waldman and Yammarino, 1991; Howell and Avolio 1993; Van Wart 2012).

As mentioned, inspectors work in a bureaucratic structure, and their work, although different each day, is dominated by routine rather than change and innovation. And the supervisors whose leadership style we have studied are middle managers, not strategic top managers. This makes transformational leadership, which focuses on change and charismatic leadership less suitable for our study.¹

Given the individual character of the work of inspectors – they commonly do inspections alone, and they develop individual enforcement styles – it is important for their supervisors to guide them individually and apply an interpersonal leadership style. Thus, it is logical to study a more inter-personal perspective of leadership as complement of a transactional leadership style. The last decennia a wide range of theories has emerged that can be typified under the heading of inter-personal perspectives on leadership (see Van Wart 2012; Ricard, Klijn, Lewis, and Ysa 2017). Interpersonal leadership perspectives emphasize how leaders interact with their followers/employees, and consider that as the most crucial part of leadership. Interpersonal leadership styles, and this holds for all varieties that can be considered part of this group, focus on interacting with and facilitating of employees. Fairness and setting an example are also often emphasized (Van Wart 2012).

One of the many theories that fit in this inter-personal perspective on leadership is servant leadership, which builds on Greenleaf’s ideas of ‘the servant as leader’ (Greenleaf 1977). Van Dienrendonck (2011) in his overview article gives as main characteristics of a servant-oriented view on leadership:

• empowering and developing people;

• humility, i.e. the ability to put one’s own accomplishment in perspective;

• authenticity which means that leaders express themselves in ways that are consistent with their thoughts and feelings;

• providing direction and ensuring that people know what is expected;

• stewardship, i.e. taking responsibility for the organization as a whole and prioritizing service rather than self-interest (see also Miao et al. 2014).

In contrast with a transactional style of leadership that is more top-down oriented, servant leadership puts the leader less central and is more horizontally oriented. It also embraces the idea that leaders should serve as role models, thus relating to ideas about ethical leadership (see Van Wart 2013b). Miao et al. (2014) argue that servant leadership has close connections to a public sector ethos which emphasizes public values and intrinsic motivation to serve public values (which is also emphasized in the Public Service motivation literature) (Perry and Wise 1990; Moynihan and Pandey 2007). This also makes the perspective suitable for our purpose since we are dealing with public servants. And we also have identified two clearly different and well-identifiable leadership styles for our research purpose: transactional and servant leadership.

How would a transactional leadership style and a servant leadership style influence enforcement styles? Transactional leadership is aimed at getting results, and leading through performance indicators and rewards (Avolio et al, 1991; Howell and Avolio 1993). The most important performance for inspectors is compliance of inspectees. The more transactional leadership is, the more inspectors need to accommodate to performance indicators set by their leaders in order to get rewards and positive performance evaluations. This would imply a positive correlation between transactional leadership and the accommodative enforcement style (especially paying attention to their supervisor). However, if inspectors have to accommodate demands from their leader/manager, they may become less accommodative to wishes from other actors. This would imply a less accommodative enforcement style. It is difficult to predict the net effect of transactional leadership style on the accommodative dimension of enforcement style but we expect that the net effect is not significant (more attention to supervisors is compensated by less attention for others is our expectation).

Transactional leaders focus on performance indicators and rewards. In the case of inspectors, this means that they are largely judged on being on target in terms of inspectee compliance, and on following procedures. Therefore, one may expect that transactional leadership will influence inspectors towards a formal enforcement style and using coercion to hold inspectees in line. We expect a less facilitating enforcement style when inspectors perceive their superiors have a strong transactional leadership style. In general, a transactional style emphasizes clear (top down) instructions which will limit the space for inspectors to decide things on their own and will likely lead to less space to manoeuvre and a tendency of inspectors to use a less facilitative enforcement style.

With servant leadership, we expect different effects. Servant leadership would be aimed to support inspectors and create more space for their own judgements. Therefore, we expect that inspectors who perceive their supervisors to have a more servant leadership style, apply a higher level of facilitation in their enforcement style. Further, inspectors who get more space from their (servant) managers, have the option to rely less on strict coercion. Therefore, we expect a correlation of a more servant leadership style with a less strong emphasis on a legal enforcement style, although it will not be very strong given the tight regulation rules. We also expect a positive effect of perceived servant leadership on the accommodation dimension of the enforcement style. A servant leadership style is more horizontal and directly aimed at helping the inspector, thus we would expect that a servant leadership style encourages inspectors to pay more attention to their superiors and their fellow inspectors.

The arguments above lead to the following hypotheses:

1.a. when inspectors perceive the leadership style of their supervisor as more transactional leadership, they will show a more legal enforcement style

1.b. when inspectors perceive the leadership style of their supervisor as more transactional leadership, they will show a less facilitating enforcement style

1.c. When inspectors perceive the leadership style of their supervisor as more transactional leadership this will not show any relation with the accommodation dimension of the enforcement style of the inspector

2.a. When inspectors perceive the leadership style of their supervisor as more servant leadership, they will show a less legal enforcement style

2.b. When inspectors perceive the leadership style of their supervisor as more servant leadership, they will show a more facilitating enforcement style

2.c. when inspectors perceive the leadership style of their supervisor as more servant leadership, they will show a more accommodative enforcement style

Factor 2: organizational culture

It is widely recognized that organizational culture, as the basic assumptions and beliefs shared by members of the organization, influences the behaviour of members of the organization (see Schein 2010). Organization culture can manifest itself in the (implicit) rules and norms of an organization but also in artefacts, myths, stories and rituals (see Schein 2010). As such organizational culture can influence both the leadership style of an organization like an inspection agency, and the enforcement style of inspectors (see e.g. Cohen 2018). Organizational culture is however a complex concept and many different conceptualizations and measurements exist (Moynihan and Pandey 2007). To research the influence of organizational culture on both leadership style and enforcement we use two of the four categorizations from the work by Quinn and Kimberly (1984), who make a distinction between group, developmental, hierarchical and rational cultures.

In this research, we draw on the hierarchical and development types of culture. The first emphasizes rules and bureaucratic procedures while a developmental culture emphasizes an entrepreneurial attitude and learning (see also Schein 2010). These two dimensions are particularly relevant and suitable for our case, the Dutch Food and Drug agency (NVWA). The NVWA as a large public inspection organization, is arguably more a state agency than a citizens agency because it is oriented at enforcing rules. It is thus heavily determined both by public rules about food and product safety and procedural rules on how they have to organize enforcement. Thus, inspectors function in a bureaucratic organization that is strongly regulated by rules and procedures, and has a clear hierarchical structure. Thus, we expect higher scores for the hierarchical culture, than for the developmental dimension which we included to measure how entrepreneurial the culture is according to inspectors and whether this influences their enforcement style.

So, one can expect that in an organizational culture that is perceived as strongly bureaucratic, where rules have to be followed and also rule-based control of employees is emphasized, inspectors will apply a more legal enforcement style, and will be less inclined to use a facilitation style (see also Hood, Rothstein, and Baldwin 2001; Destler 2017). In a hierarchical culture, one can also expect that inspectors have a less accommodative style since they are primarily motivated by following the rules not by accommodating others (that is supervisors or peers). When inspectors perceive the organizational culture to be more developmental, we expect the opposite because a development culture encourages inspectors to make their own judgements but also discuss enforcement with their peers and superiors in a more open way.

This leads to the following hypotheses:

3.a. when inspectors perceive the organizational culture of their organization as more hierarchical, they will show a more legal enforcement style

3.b. when inspectors perceive the organizational culture of their organization as more hierarchical, they will show a less facilitating enforcement style

3.c. When inspectors perceive the organizational culture of their organization as more hierarchical, they will show a less accommodative enforcement style

4.a. When inspectors perceive the organizational culture to be more developmental, they will show a less legal enforcement style

4.b. When inspectors perceive the organizational culture of their organization to be more developmental, they will show a more facilitating enforcement style

4.c. when inspectors perceive the organizational culture of their organization to be more developmental, they will show a more accommodative enforcement style

One can also expect that a more hierarchical culture is positively related to a more transactional leadership style since a hierarchical culture focuses on rules and clear demarcations which have to be followed. On the other hand, we expect a developmental culture be positively correlated to a servant leadership style. After all a developmental culture does emphasize risk-taking and individual initiative more. That fits with a leadership style that focuses on individual employees and encourages them to develop themselves (Van Dienrendonck 2011).

5a. When inspectors perceive a more hierarchical culture of their organization, they will also perceive a more transactional leadership style

5.b. When inspectors perceive a more developmental culture of their organization, they will also perceive a more servant leadership style

Methods

The data for this research were collected through an online survey among all inspectors of the Netherlands Food and Consumer Product Safety Authority (NVWA) (n = 1153) in the autumn of 2018. 757 inspectors responded to our survey invitation. The response rate was, thus, 65.7%. 208 of those respondents were excluded from the analysis because they did not complete the questionnaire. The final sample used for the analysis was 549 respondents. With regard to gender, 70.8% is male, 28.0% is female and 1.2% other. On average, inspectors are 51.1 (SD = 10.7) years old and have 18.7 (SD = 12.0) years of work experience.

These sample characteristics mimic the characteristics of the overall population of inspectors. The gender and average age of the total population only deviated slightly from our sample characteristics. With regard to gender, the total population includes 66.9% men and 33.1% women. The average age of the total population is 50.8. The average years of work experience of the total population is 13.4 which is substantially lower than in our sample (18.7 years). This should be taken into account when interpreting our findings.

Measurements

Enforcement style

With the conceptualizing and measurement of enforcement style, we followed the conceptualization of De Boer (2019). This measurement was developed and validated drawing on a similar group of respondents (De Boer 2019) and has been empirically applied to explain enforcement style (De Boer and Eshuis 2018; De Boer, Eshuis, and Klijn 2018). Thus, we measured enforcement style through three dimensions: (1) legal; (2) facilitation; and (3) accommodation. Each dimension (see Appendix 1) was measured with four or five items on a 10-point scale (1 = never, 10 = always). Some examples of the items are: ‘During inspections, I focus on transferring my professional knowledge to inspectees’ (i.e. facilitation). Or ‘During inspections, I consider the opinion of my team about enforcing’ (i.e. accommodation).

Leadership style

Servant leadership was measured through items developed by Ehrhart (2004). Those items have been widely used by other researchers in Public Administration (e.g. Miao et al. 2014). We adapted the statements slightly so that they fit the situation (for example, replacing the word ‘subordinate’ by ‘inspector’ to make the statement more relevant for the respondents). The original scale consists of 14 items, but our preparatory conversations in the organizations and interviews with inspectors, had pointed out that we had to limit the length of our survey or otherwise the response rate would drop seriously. Therefore, we did not use all 14 items, but rather the 5 items that were strongest related to the core idea of servant leadership and emphasize support of employees (thus omitting items that emphasize consensus, or the involvement in community service). The measurement is reliable with ω = .96.

The items for transactional leadership were taken from the classical items developed by authors like Bass and Avolio (the MLQ leadership questionnaire see Avolio, Bass, and Jung 1999; Den Hartog, Van Muijen, and Koopman 1997) on transactional leadership. Again we had the problem that the original scale contains many items (and different aspects like an active and a passive or laisses fair component). We chose the statements in the transaction leadership scale that focus on continuous reward and active management (see Den Hartog, Van Muijen, and Koopman 1997) as these were most closely to the core idea of transactional leadership.

All statements that were used are presented in the appendix. We tested the items with experts in the NVWA to ensure that they reflected the reality of our respondents. Some items were slightly amended. An example of the rewording is the transactional statement ‘keeps track of mistakes’ which we rephrased in ‘Keeps track of your mistakes that do not fit to standard protocols for inspectors’. All statements were translated from English to Dutch. In translation and rephrasing, we tried to stay closely to the original wording and meaning. The measurement for transactional leadership is reliable with ω = .91.

Organizational culture

To measure the two dimensions (hierarchical culture and developmental culture) we used the dimensions developed by Quinn and Kimberly (1984) that were converted into survey questions by Zammuto and Krakower (1991, see also Moynihan and Pandey, 2007). They developed two items for hierarchical culture (1. my agency is a very formalized and structured place and 2. Hierarchical procedures generally govern what people do) and two for developmental culture (1. My agency is a very dynamic and entrepreneurial place and 2. People are willing to stick their necks out and take risks). We tested the scores of the four items with a factor analysis. This analysis clearly showed two factors (hierarchical and development culture). However, they did not form reliable measurements, which is common with latent constructs with only two items (with ω = .43 for both hierarchical and developmental culture). Therefore, we used 1 item of each measure which best fitted our conceptualization (see appendix). We measured how inspectors perceive organizational culture, which is relevant since different people may experience the culture in an organization differently. We measured in how far inspectors experience the culture in the NVWA as bureaucratic and as developmental. This measurement is at the organizational-level and not unit-level since inspectors often operate in small teams. We are not interested in comparing these organizational units, but are interested in the impact of organizational culture at large and, therefore, we measure this variable at the NVWA-level.

Common method variance

Common method variance was combatted by following common best practices in our field concerning the design of the questionnaire, such as: testing the questionnaire with experts, separating core variables on different pages, ensuring organizational support and continuedly highlighting the importance of participating to respondents (see George and Pandey 2017; Podsakoff, MacKenzie, and Podsakoff 2012). On top of that, two statistical remedies were executed. First, a marker variable was used (Lindell and Whitney 2001). The marker (i.e. n of hours spent using media platforms) was loaded to all latent variables. While the model fit (χ2 = 621.822, df = 274, p = .000) was good with CFI = .916, TLI = .900, RSMEA = .069 and SRMR = .062. ANOVA indicated that including the marker did not significantly improve our model (χ2 difference = 6.1931, df = 5, p = .288). More importantly, the marker was not significantly related to any of our variables (with p-values ranging between .164 and .969). Second, a common factor analysis was estimated (Podsakoff et al. 2003). This means that all individual items were loaded on one latent construct. The fit of the model (χ2 = 2289.426, df = 275, p = .000) is very poor with CFI = .649, TLI = .617, RSMEA = .123, PCLOSE = .000 and SRMR = .154. These statistical remedies show that common source bias is unlikely to impact our results.

Results

Table 1 presents the correlations between the main variables. The correlations between several independent variables (i.e. developmental- and hierarchical culture) and dependent variables (i.e. the three dimensions of street-level bureaucrats’ enforcement style) are statistically significant. Hierarchical culture correlates positively with the accommodation (.16) dimension of bureaucrats’ enforcement style. Developmental culture correlates positively with the facilitation (.13) and accommodation (.19) dimension.

Table 1. Means, standard deviations and correlations

		M	SD	1	2	3	4	5	6	7	8	9	10
1	Facilitation	7.80	1.26	1
2	Accommodation	6.08	1.65	0.29***(0.000)	1
3	Legal	7.81	1.12	0.46*** (0.000)	0.18*** (0.000)	1
4	Developmental culture	5.46	2.32	0.07 (0.131)	0.19*** (0.000)	0.13** (0.003)	1
5	Hierarchical culture	5.97	2.24	0.06 (0.166)	0.16*** (0.000)	0.03 (0.479)	0.11** (0.011)	1
6	Transactional leadership	5.38	2.08	0.09* (0.052)	0.38*** (0.000)	0.10** (0.025)	0.31*** (0.000)	0.13** (0.004)	1
7	Servant leadership	6.08	2.21	0.07 (0.107)	0.36*** (0.000)	0.11** (0.017)	0.36*** (0.000)	0.16*** (0.000)	0.84*** (0.000)	1
8	Job engagement	7.23	2.00	0.16*** (0.000)	0.14** (0.002)	0.31*** (0.000)	0.39*** (0.000)	0.19*** (0.000)	0.22*** (0.000)	0.23*** (0.000)	1
9	Age	51.1	10.7	0.01 (0.825)	0.01 (0.765)	0.01 (0.875)	−0.07 (0.117)	−0.10** (0.021)	0.02 (0.737)	0.02 (0.696)	−0.04 (0.854)	1
10	Gender (1 = male)	0.71	0.46	0.02 (0.716)	0.11** (0.010)	0.00 (0.772)	0.04 (0.955)	0.15*** (0.000)	0.04 (0.325)	0.00 (0.976)	0.04 (0.358)	0.38*** (0.000)	1

Note: P-values between brackets

Concerning our two mediation variables – transactional and servant leadership style – Table 1 also reveals several significant correlations. Transactional leadership correlates positively with all dimensions of our dependent variable (with .09, .38 and .10 for facilitation, accommodation and legal enforcement style, respectively). The servant leadership style only correlates positively with the accommodation enforcement style of street-level bureaucrats (.16). Last, both leadership styles correlate positively with developmental culture (.31 and .36, respectively). Both leadership styles also correlate with hierarchical culture (.13 and .16, respectively).

Three control variables, job engagement, age and gender (1 = male) were also correlated with our main variables. Table 1 shows that neither of them are related to our leadership or enforcement style variables in a statistically significant manner. They were, therefore, not included in our model.

The hypothesized relations were further investigated using Structural Equation Modelling in the statistical programme R. The packages ‘lavaan’ (Rosseel 2011), ‘psych’ (Revelle 2014), and ‘semTools’ (Pornprasertmanit et al. 2013) were used to estimate our parameters. Our data deviate slightly from multivariate normality assumptions and, therefore, the Satorra-Bentler correction was used. The model (χ2 = 713.34, df = 256, p = .000) fit is good with CFI = .924, TLI = .911, RSMEA = .068 and SRMR = .059). Table 2 shows the direct, indirect and total effects of our model. Figure 1 visualizes our model.

Figure 1. SEM results visualization.

Table 2. SEM results of direct and indirect effects on enforcement style

	z	St.SE	St.B	p-value
Direct effects
Of hierarchical culture on
Legal enforcement dimension	2.416	0.043	0.182	0.016**
Facilitation enforcement dimension	3.836	0.050	0.284	0.000***
Accommodation enforcement dimension	2.090	0.037	0.118	0.037**
Transactional leadership	−0.886	0.048	−0.045	0.376
Servant leadership	−0.398	0.055	−0.018	0.690
Of developmental culture on
Legal enforcement dimension	2.015	0.028	0.119	0.044**
Facilitation enforcement dimension	1.827	0.032	0.104	0.068*
Accommodation enforcement dimension	1.721	0.032	0.101	0.085*
Transactional leadership	6.585	0.040	0.333	0.000***
Servant leadership	7.214	0.050	0.357	0.000***
Of transactional leadership on
Legal enforcement dimension	0.772	0.123	0.159	0.440
Facilitation enforcement dimension	1.316	0.109	−0.118	0.188
Accommodation enforcement dimension	3.103	0.108	0.490	0.002**
Of servant leadership on
Legal enforcement dimension	−0.544	0.092	−0.107	0.587
Facilitation enforcement dimension	−0.810	0.082	−0.118	0.418
Accommodation enforcement dimension	−0.080	0.086	−0.012	0.936
Indirect effect
Of developmental on accommodation via transactional	2.802	0.031	0.163	0.005**
Total effect
Of developmental on accommodation via transactional	3.159	0.052	0.281	0.002**

Hypothesis 1 predicts that a transactional leadership style increases a legal enforcement style (H1a), decreases a facilitation enforcement style (H1b) and will not affect the accommodation enforcement style (H1c). Table 2 shows that the standardized coefficients for neither the legal nor the facilitation enforcement style are statistically significant (with p = .440 and p = .188, respectively). Opposed to our expectations, a transactional leadership style does increase an accommodation enforcement style (z = 3.103, st.B = .490, SE = 108, p = .002). In other words, the more inspectors perceive their supervisors to be transactional, the more accommodative their enforcement style is during face-to-face encounters with inspectees. Hypotheses 1a, 1b and 1 c are all rejected.

The second set of hypotheses predicted that a servant leadership style would result in a less legal enforcement style (H2a), a more facilitation enforcement style (H2b) and a less accommodation enforcement style (H2c). Table 2 reveals that none of the standardized coefficients are statistically significant. To put it differently, when inspectors perceive their supervisors to have a more servant leadership style this does not affect their enforcement style during inspection visits. Hypotheses 2a, 2b and 2 c are all rejected.

The third and fourth set of hypotheses address the direct effects of organizational culture on inspectors’ enforcement style. The third hypotheses predicted that a more hierarchical culture leads to a more legal enforcement style (H3a), less facilitation (H3b) and more accommodation (H3c). Table 2 shows that the standardized coefficients for all the expected direct effects are statistically significant. More specifically, when inspectors perceive their organizational culture to be hierarchical, they become more legal (z = 2.416, st.B = .182, SE = .043, p = .016). Hypothesis 3a is confirmed. It is also found that when there is a more hierarchical culture, inspectors become accommodative in their enforcement style (z = 2.090, st.B = .118, SE = .037, p = .037). Hypothesis 3 c is also confirmed. Opposed to our predictions, when the experienced hierarchical culture rises, so does the inspectors’ facilitation (z = 3.836, st.B = .284, SE = .050, p = .000). In other words, while we expected that a hierarchical culture would be negatively related to inspectors’ facilitation, it is positively related. Hypothesis 3b is, therefore, rejected.

The fourth set of hypotheses expected that a more developmental culture would result in a less legal (H4a), more facilitation (H4b) and more accommodation (H4c) enforcement style. Table 2 shows that only one standardized coefficient is statistically significant at the .05-level. It is found that when inspectors’ perceive the organizational culture to be more developmental, they also become more legal in their enforcement style (z = 2.015, st.B = .119, SE = .028, p = .044). This relationship is the opposite of what we predicted. Therefore, hypotheses 4a, 4b and 4 c are all rejected.

The final set of hypotheses predicted the direct effect of organizational culture on leadership style. It was hypothesized that a more hierarchical culture would increase the perceived transactional leadership style (H5a) and a more developmental culture would increase the perceived servant leadership style (H5b). Table 2 shows that only the effect of developmental culture on servant leadership is statistically significant (z = 6.585, st.B = 357, SE = .050, p = .000). Therefore, hypothesis 5a is rejected and 5b confirmed.

In addition to the direct effects, our model also included indirect effects. As discussed earlier, for developmental culture (H5b) is the direct effect of our independent variable significantly related to our mediating variable: transactional leadership (parameter a). Likewise, the direct effect of transactional leadership on accommodation enforcement style (H1c) is also statistically significant (parameter b). Therefore, in our model, we explored the direct effects of these two parameters. Table 2 shows that the indirect effect is, indeed, statistically significant (z = 2.802, st.B = .161, SE = .031, p = .005). The total effect is also statistically significant (z = 3.159, st.B = .281, SE = .052, p = .002). To put it differently, when inspectors’ perceive their organization to be more developmental in culture, they perceive their supervisors to be more transactional which, in turn, results in a more accommodative enforcement style when they encounter inspectees (see Appendix 2 for full overview of all indirect and total effects). Table 3 shows an overview of our confirmed and rejected hypotheses.

Table 3. Overview hypotheses

Hypothesis	Confirmed	Rejected
1.a. when inspectors perceive the leadership style of their supervisor as more transactional leadership, they will show a more legal enforcement style		X
1.b. When inspectors perceive the leadership style of their supervisor as more transactional leadership, they will show a less facilitating enforcement style		X
1.c. When inspectors perceive the leadership style of their supervisor as more transactional leadership this will not show any relation with the accommodation dimension of the enforcement style of the inspector		X opposite found
2.a. When inspectors perceive the leadership style of their supervisor as more servant leadership, they will show a less legal enforcement style		X
2.b. When inspectors perceive the leadership style of their supervisor as more servant leadership, they will show a more facilitating enforcement style		X
2.c. When inspectors perceive the leadership style of their supervisor as more servant leadership, they will show a more accommodative enforcement style		X
3.a. When inspectors perceive the organizational culture of their organization as more hierarchical, they will show a more legal enforcement style	X
3.b. When inspectors perceive the organizational culture of their organization as more hierarchical, they will show a less facilitating enforcement style		X opposite found
3.c. When inspectors perceive the organizational culture of their organization as more hierarchical, they will show a less accommodative enforcement style	X
4.a. When inspectors perceive the organizational culture to be more developmental, they will show a less legal enforcement style		X
4.b. When inspectors perceive the organizational culture of their organization to be more developmental, they will show a more facilitating enforcement style		X
4.c. When inspectors perceive the organizational culture of their organization to be more developmental, they will show a more accommodative enforcement style	X
5a. When inspectors perceive a more hierarchical culture of their organization, they will also perceive a more transactional leadership style		X
5.b. When inspectors perceive a more developmental culture of their organization, they will also perceive a more servant leadership style	X

Conclusions and reflections

This paper focuses on the effects of leadership styles on enforcement styles of inspectors. Our research question was: ‘How are enforcement styles of inspectors influenced by the leadership style of their direct supervisor’. Given the literature on leadership that emphasizes the importance of leadership for the behaviour of employees (see Bass 1985; Avolio et al, 1991; Van Wart 2012) we expected that leadership styles would affect enforcement styles of inspectors. Especially because the importance of leadership has been argued regularly also for the public sector (see Van Wart 2013b; Javidan and Waldman 2003).

However, our findings show only very limited influence of leadership styles (whether servant or transactional leadership style) on inspectors’ enforcement styles. The only significant relation is between transactional leadership and an accommodative enforcement style. So, inspectors tend to pay more attention to the opinion of others when their direct supervisor uses a more transactional style of leadership. This is contrary to what we expected, namely no relationship between transactional leadership and an accommodative enforcement style because in cases of strongly transactional leadership the greater attention of inspectors to supervisors would be compensated by less accommodation to others. But the overall conclusion is that there is very little effect of leadership on enforcement style, which is remarkable and begs further theoretical reflection.

Organizational culture has more effect on enforcement styles of inspectors than leadership. We find significant positive effect of hierarchical culture on all three enforcement styles and significant positive effect of developmental culture on legal and accommodative style. The results indicate that organizational culture is more important than leadership style in shaping inspector’s enforcement styles.

Street-level bureaucrats: leadership less important?

The very limited influence of leadership on enforcement style is remarkable, and asks for reflection because a vast literature on public leadership emphasizes its importance (Van Wart 2013a; Tummers and Knies 2013). This has been confirmed in a wide range of empirical research (Avolio, Bass, and Jung 1999; Tummers and Knies 2013; Miao et al. 2014) although some authors find that leadership is less important in a public context (see Javidan and Waldman 2003).

Of course our results are not a definitive rejection of the leadership matters hypothesis in the case of street-level bureaucrats. Recently other research, even done in the same country (The Netherlands) did find a relation between leadership and attitude towards clients (see Keulemans and Groeneveld 2019). But this research was done under tax inspectors who stay mostly in their own organization. Thus, their situation is different from our sample that concerns inspectors who do their enforcement job on the spot. This points out that the actual situation of the street-level bureaucrat makes a lot of difference if it comes to the influence of leadership. It nuances the general conclusion in much leadership literature that leadership matters. Our research points at a classical answer on this: it depends!

In our case, there may be several reasons why leadership does not have much effect. First of all, inspectors are strongly constrained by the rules they must follow in their inspection and the hierarchical/hierarchical organization culture they are in (see Raaphorst 2018 for similar argumentation). Thus, leadership may be less important than the tight regulative structure and bureaucratic culture of the inspection as an organization. So although inspectors have individual discretion in their job their enforcement styles is still dominated more by the (formal) regulation that is in place, than by the leadership style of their direct supervisor. The fact that the scores for the legal style of enforcement are rather high can be seen as a support for this assumption.

Another explanation may be that the work of inspectors as street-level bureaucrats is a unique setting, and its characteristics (such as the type of task, the characteristics of the encounter with the inspectee) may be more important than the leadership style of the supervisor in shaping enforcement style (see also Broekema, Van Kleef, and Steen 2017). For instance, the inspectors in our study specifically are often ‘out’ in the field and spend the majority of their time alone (or with a colleague) on the road or visiting inspectees for inspections. In other words, these inspectors may have less contact with a supervisor than many other inspectors (like the inspectors in the tax agency researched in the Keulemans and Groeneveld research). It is possible that inspectors develop their enforcement style while working in the field, and the situation they encounter there, outside their organization, is more important to shape their enforcement style than the contact with their supervisor. Future research could explore whether the amount of time spent physically at the office explains why for some inspectors leadership does and for other inspectors it does not matter.

Limitations of the research

Of course, this research has several limitations. First of all, we did not use all the items of the two leadership styles (see our discussion about the items in the method section) which potentially harms the validity and reliability of our measurement. However, our reliability analysis does not indicate problems. Further, we had only limited items for a complex concept like organizational culture which may have influenced the outcomes. A second limitation was that unlike, for instance, Keulemans and Groeneveld (2019) we only have measured inspectors’ perceptions of leadership style, and we were not able to connect the leadership style to specific leaders. We, therefore, cannot control for potential differences between supervisors. Third, we measured both our independent- and dependent variable at the same time and among the same respondents raising concerns of common source bias. As extensively discussed in our method sections, there is no indication that this bias has affected our results. Nevertheless, this limitation should be taken into account when interpreting the findings.

Regardless, we think these results are valuable for the ongoing discussion about street-level bureaucrats and what determines their behaviour. Our findings do suggest that leadership might be a less important factor than much of the literature suggests. We hope that this urges Public Administration scholars to open up the debate about the influence of leadership styles for the case of street-level bureaucrats.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Dutch Food and Consumer Product Safety Authority (NVWA).

Notes on contributors

Erik Hans Klijn

Erik Hans Klijn is professor of public administration at the department of Public Administration and sociology, Erasmus school of Social and behavioral studies at the Erasmus university Rotterdam.

Noortje de Boer

Noortje de Boer is assistant professor at the school of Governance, Utrecht University.

Jasper Eshuis

Jasper Eshuis is associate professor at the department od Public Administration and Sociology, Erasmus School of Social and Behavioral Studies, erasmus University Rotterdam.

Notes

1. The fact that transformational leadership has been criticized heavily the last years for flaws in the conceptualization and research (see for a harsh criticism Knippenberg and Sitkin 2013) is another reason to focus on transactional and servant leadership

Appendix 1.

SURVEY ITEMS FOR THE VARIABLES

Factor loadings

Enforcement style

Legal style (ω =.80)

•During inspections, I focus on:

•Implementing the intervention policy by following the letter of the law0.56

•Enforcing in an unambiguous way0.76

•Making strict agreements with [inspectees]0.71

•Executing the inspection as complete as possible0.69

•Upholding high standards for compliance with rules and regulations for [inspectees]0.56

Facilitation style (ω =.80)

During inspections, I focus on:

•Transferring my professional knowledge to [inspectees]0.73

•Giving indications to [inspectees] on how to improve compliance0.84

•Being as helpful as possible to clients0.76

•Considering the circumstances of [inspectees]0.43

Accommodation style (ω =.77)

During inspections, I consider:

•The opinions of inspectors from my team about enforcing0.70

•The opinions of inspectors from other teams about enforcing0.75

•The opinion of my team leader about enforcing0.58

•The opinions of directors/head inspectors about enforcing0.66

Leadership style

Servant leadership (ω =.96)

My supervisor:

•Spends time to form quality relationships with inspectors0.94

•Creates a sense of community among inspectors0.94

•Her/His decisions are influenced by inspectors’ input0.83

•Makes the personnel development of inspectors a priority0.86

•Works hard at finding ways to help inspectors be the best they can be0.94

Transactional leadership (ω =.91)

My supervisor:

•Tells me what to do to be rewarded0.75

•Tracks your mistakes and monitors performance for error needing correction0.87

•Recognizes your achievements0.83

•Keeps track of your mistakes that do not fit the standard protocols for inspectors0.79

•Rewards my achievement0.85

Organizational culture

Hierarchical culture

•Bureaucratic procedures generally govern what people do

Developmental culture

•The NVWA is a very dynamic and entrepreneurial place

Appendix 2.

ALL INDIRECT AND TOTAL EFFECTS

SEM results indirect effects

Table

	z	St.SE	St.B	p-value
Indirect effects
Of hierarchical on legal via transactional	−0.562	0.007	−.007	0.574
Of developmental on legal via servant	−0.542	0.033	−0.038	0.588
Of developmental on legal via transactional	0.766	0.032	0.053	0.444
Of hierarchical on legal via servant	0.318	0.003	0.003	0.750
Of hierarchical on facilitation via transactional	−0.755	0.008	−0.009	0.450
Of developmental on facilitation via servant	−0.808	0.029	−0.042	0.419
Of developmental on facilitation via transactional	1.293	0.029	0.067	0.196
Of hierarchical on facilitation via servant	0.351	0.004	0.002	0.726
Of hierarchical on accommodation via transactional	−0.853	0.017	−0.022	0.394
Of developmental on accommodation via servant	−0.080	0.031	−0.005	0.936
Of developmental on accommodation via transactional	2.802	0.031	0.163	0.005**
Of hierarchical on accommodation via servant	0.078	0.002	0.000	0.938

SEM results total effects

Table

	z	St.SE	St.B	p-value
Total effects (direct + indirect)
Of hierarchical on legal via transactional	2.283	0.043	0.175	0.022**
Of developmental on legal via servant	0.921	0.041	0.081	0.357
Of developmental on legal via transactional	2.426	0.053	0.235	0.015**
Of hierarchical on legal via servant	2.022	0.028	0.121	0.043**
Of hierarchical on facilitation via transactional	3.683	0.051	0.275	0.000***
Of developmental on facilitation via servant	0.809	0.043	0.062	0.418
Of developmental on facilitation via transactional	3.588	0.064	0.351	0.000***
Of hierarchical on facilitation via servant	1.861	0.032	0.106	0.063*
Of hierarchical on accommodation via transactional	1.443	0.044	0.097	0.149
Of developmental on accommodation via servant	1.224	0.042	0.096	0.221
Of developmental on accommodation via transactional	3.159	0.052	0.281	0.002**
Of hierarchical on accommodation via servant	1.717	0.032	0.101	0.086*