Jump to ContentJump to Main Navigation
Understanding street-level bureaucracy$

Peter Hupe, Peter Hupe, Michael Hill, and Aurèlien Buffat

Print publication date: 2015

Print ISBN-13: 9781447313267

Published to University Press Scholarship Online: January 2016

DOI: 10.1332/policypress/9781447313267.001.0001

Show Summary Details
Page of
Subscriber: null; date: 13 December 2017

Bureaucratic, market or professional control? A theory on the relation between street-level task characteristics and the feasibility of control mechanisms

Bureaucratic, market or professional control? A theory on the relation between street-level task characteristics and the feasibility of control mechanisms

Chapter:
(p.205) Twelve Bureaucratic, market or professional control? A theory on the relation between street-level task characteristics and the feasibility of control mechanisms
Source:
Understanding street-level bureaucracy
Author(s):

Duco Bannink

Frédérique Six

Eelco van Wijk

Publisher:
Policy Press
DOI:10.1332/policypress/9781447313267.003.0012

Abstract and Keywords

Street-level bureaucrats do their work in different contexts and these contexts pose different control challenges. This study distinguishes two context variables: complexity and ambiguity. When ambiguity and complexity are simultaneously high, managers face a double control challenge - achieving both the expertise to address high complexity and the alignment to overcome high ambiguity. Traditional control mechanisms, enforcement, incentives or competence control, fall short. Using evidence from a wide cross-section of street-level tasks (42 cases), this study provides support for the claim that the double control challenge in contexts of simultaneously high complexity and high ambiguity is important in street-level bureaucracies and that it is difficult to devise a response. Usually one of the challenges is simply ignored. Where a response to both control challenges is devised simultaneously, tensions occur because some of the mechanisms conflict with each other. We found very few organisations that have found appropriate solutions.

Keywords:   control, complexity, ambiguity, double control challenge

Introduction

Street-level bureaucracy is two things: it is an effort to have policy implementation conform to general and abstract rules; and it is an effort to apply rules to specific and concrete cases. This central assumption informs our analysis of the managerial control of street-level action. The first assumes control of action through general regulation; the second assumes adjustment of action to specific case conditions. These are necessarily incongruent functions. Bridging the distance between general and abstract rules and specific and concrete cases requires action that cannot be fully defined by either. This is why, as Lipsky (2010) phrased it, street-level bureaucrats are policymakers. Street-level bureaucrats do not simply implement given rules in cases that can be fully understood on the basis of these rules, but instead translate rules into client-level decisions, building upon information (not fully defined in the rule) on clients’ conditions and upon expertise (also not fully defined) on client treatment.

There is an abundant literature on managerial and bureaucratic control over street-level bureaucrats. Meyers and Vorsanger (2003, p 246) argue that ‘the questions of whether, and how, policy making principals control the discretion of their implementing agents have dominated much of the empirical research’. Winter (2003) asks whether ‘bureaucrats [are] servants or masters, and to what extent … bureaucrats [can] be controlled by their political superiors’ and argues that there is ‘differentiated and limited control’ over street-level bureaucracies that varies according to the extent of information (p.206) asymmetry. Meyers, Riccucci and Lurie (2001, p 165) argue that the achievement of ‘goal congruence’ between political control actors and street-level bureaucrats might increase ‘implementation fidelity and policy achievement’. The main thrust of the insights in this literature is that control over street-level action is limited and indirect.

These arguments conceptualise the issue of street-level control as placed on a single dimension of increased control versus increased street-level discretion. We disagree with this conceptualisation. Our central assumption is that street-level bureaucracy does not have one manifest function, but two: the control of policy implementation given existing regulations; and the translation of policy rules to clients given the street-level information on client conditions and expertise on client treatment. Brodkin (2011, p 254, emphasis in original) argued that street-level organisations ‘not only do policy work, but are manifestly responsible for making policy work’. Discretion is accepted as ‘an inherent – at times even necessary – feature of implementation’ (Brodkin, 2011, p 254). Through control systems, managers seek ‘to align employee capabilities, activities, and performance with organisational goals and aspirations’ (Sitkin et al, 2010, p 3). They do not minimise the application of employee capabilities in order to achieve maximum implementation fidelity. There is an autonomous contribution that the street-level bureaucrat makes to street-level bureaucracy; this is the second function of informing the street-level bureaucratic system with client information and treatment expertise. Also, the political decision-maker makes an autonomous contribution to the street-level system: the control of street-level bureaucrats’ action. These functions are both necessary functions of a street-level bureaucratic system. In Brodkin’s phrasing, street-level bureaucrats do the work and make it work.

We transferred this insight to a discussion on control mechanisms in street-level bureaucracy. Street-level bureaucrats do their work in different contexts (Hupe and Hill, 2007). These contexts determine street-level task characteristics and pose the different control challenges that managers need to deal with. The two context variables critical for determining task characteristics and the corresponding nature of the control challenge are complexity and ambiguity. Social problem complexity is about the uncertainty of the factual estimation of social problems. Social problem ambiguity is about the uncertainty of the normative evaluation of social problems and solutions.

In this chapter, we propose a theory about the relationship between two variables: street-level task characteristics (the levels of task complexity and ambiguity) and the control mechanisms (enforcement, (p.207) incentives and competence control) that may be applied. When both the ambiguity and complexity of tasks are low, the control challenge entails how to achieve conformity. Managers may achieve this by means of enforcement: process control or standardisation of work processes (Mintzberg, 1989). When ambiguity is high but complexity is low, the control challenge is how to achieve the alignment of preferences and interests. Managers may control street-level bureaucrats by means of incentives: output controls (Bouckaert and Halligan, 2008) or the standardisation of outputs or results. So, incentive-based government is one of the main governance arrangements suitable for the organisation of street-level action, where tasks are highly ambiguous but complex to an only limited extent. When complexity is high but ambiguity is low, the challenge is one of expertise, and managers may control street-level bureaucrats, who act more or less autonomously as experts, by competence control: autonomous task execution and mutual adjustment of action, with expertise normally being guaranteed through competence controls or the standardisation of skills (see Table 12.1).

Table 12.1: Control challenges in relation to the complexity and ambiguity of social problems

Ambiguity: heterogeneity of preferences or interests

Low

High

Complexity: factual uncertainty of risks

Low

Conformity

Alignment

High

Expertise

Double control challenge

In situations of simultaneous high ambiguity and high complexity, there is a double control challenge of simultaneously controlling the application of the necessary expertise and controlling to ensure that preferences and interests are sufficiently aligned. This interaction of complexity and ambiguity, we argue, is the normal state of affairs for broad segments of street-level bureaucracy. This poses an important problem since extant theories on governing or controlling SLBs, or public professionals for that matter, provide adequate explanations and guidance for the first three situations but not for the bottom-right-hand situation of the double control challenge (eg Terpstra and Havinga, 2001; Bannink et al, 2006).

The dimensions of our matrix are almost, but not completely, similar to the dimensions in Matland’s (1995) well-known ambiguity– conflict matrix. Matland makes a distinction between the dimension (p.208) of ambiguity, in which he includes complexity, and the dimension of conflict. We diverge from Matland’s conceptualisation because we try to convey that conflict may concern either factual estimations or normative evaluations, or both. The interaction of factual and normative uncertainty creates the double control challenge that is of central interest to the discussion in this chapter.

The central question we address in this chapter is a normative or, rather, prescriptive one and concerns this double control challenge: which governance mechanisms are functional responses to the double control challenge of combined task complexity and ambiguity? In this chapter, we present an empirical study made up of 42 case studies of task characteristics and governance mechanisms. The study supports our theoretical point of departure that tasks and governance mechanisms are related and that a double control challenge – where it is difficult to choose a governance mechanism – is, indeed, important in street-level bureaucracies.

We studied cases comprised of a set of tasks and governance mechanisms applied in a human service organisation. The tasks vary widely: judge’s rulings, community policing, inspecting chicken factories, providing care at home, re-employment services, caring for troubled youth in institutions, child protection services and so on. We show that instead of finding a new governance solution that integrates the requirements of ambiguity (alignment) and complexity (expertise), most of these organisations ‘flee’ to a partially adequate governance approach in which only one of these requirements is addressed.

Some cases, however, appear to make an effort to stay in the double control challenge box. They appear to seek a way to address the double control challenge. We study the following empirical question: what are the mechanisms that they make use of? The (not yet theoretically understood) roots of a solution to the double control challenge are empirically located in the smaller cases that we studied.

This leads us to an additional, conceptual, question that we briefly address at the end of our chapter: how may control systems be designed specifically to deal with the double control challenge? In the situations where the double control challenge is present, ambiguity needs to be overcome through interactive processes that lead to the alignment of preferences and interests, while, at the same time, managers and collaborating professionals need to trust the expertise of the street-level bureaucrats that they are in disagreement with. In a programmatic sense, we make one central claim: the double control challenge posed by task complexity and ambiguity is, indeed, important for the understanding of the governance of street-level bureaucracy but the relation between (p.209) task and governance is as yet not fully understood. Some authors have begun to explore possible solutions to the double control challenge but, so far, without adequate theoretical explanations (eg De Bruijn, 2007; Bouckaert and Halligan, 2008, Noordegraaf, 2011; Thomas and Hewitt, 2011). We look, among others, at research that integrates trust-building and control as complementary managerial activities (Long and Sitkin, 2006) and the role of intrinsic motivation (Deci and Ryan, 2000; Ryan and Deci, 2000; Weibel and Six, 2013).

Control challenges and governance mechanisms

We discuss how existing organisations that have to perform in contexts of simultaneous high complexity and ambiguity respond to the double control challenge in terms of the governance principles they apply. We apply an explicitly normative approach, in which we define the appropriate solution to the control challenge in relation to the task at hand. This does not lead to a single statement on appropriate organisation, however. Our normative position includes both functions: alignment of street-level bureaucrat action to regulations and input of autonomous street-level bureaucrat expertise. As outlined in Table 12.1, where complexity and ambiguity are limited, appropriate control is oriented to ‘conformity’. Where ambiguity is high but complexity is low, the control challenge concerns the ‘alignment’ of normative orientations. Where complexity is high but ambiguity is low, street-level bureaucrat ‘expertise’ in the task execution is sought. When complexity and ambiguity are simultaneously high, seeking expertise weakens alignment and seeking alignment weakens the input of local expertise. As such, the control challenges of alignment and expertise appear contradictory and create a double control challenge. Our normative position with respect to this double control challenge is that the ‘integration’ of alignment and expertise needs to be sought. Table 12.2, then, outlines through which governance mechanisms the various control challenges can be addressed.

We define three types of governance with their respective main governance mechanisms. In bureaucracy, enforcement is the main mechanism to steer the action of subordinates; this presupposes compliance by subordinates and the capacity of the controlling actor to define tasks. In managerialism, incentives function as the main mechanism to direct the action of service providers; this presupposes a conflict of interests (to be overcome by incentives) and the capacity of the controlling actor to define the required performance. In professionalism, competence control is the main mechanism in (p.210)

Table 12.2: Ability of governance mechanisms to deal with complexity and ambiguity

Ambiguity: heterogeneity of preferences and interests

Low

High

Complexity: factual uncertainty of risks and policy objectives

Low

Bureaucracy (enforcement)

Managerialism (incentives)

High

Professionalism (competence control)

professional governance to accommodate expertise and task; this presupposes a natural willingness by experts to engage in mutual adjustment and requires (but also allows) the controlling actor not to define tasks and performance.

Addressing the need for conformity: bureaucracy

When both complexity and ambiguity are low, conformity to the wishes and preferences of management may be achieved through governance based on bureaucratic rules and authority (cf Bradach and Eccles, 1989; Thompson et al, 1991; Terpstra and Havinga, 2001; Hupe and Hill, 2007). The aim here is full control of policy implementation. In such a hierarchical governance mechanism, management operates as the regulator of the actions at the street level (Kirkpatrick, 1999; Terpstra and Havinga, 2001; Keast et al, 2006; Bannink et al, 2006). Management defines the social problems that street-level bureaucrats respond to and the way in which the latter need to implement the desired response, that is, how they execute their task.

The Dutch General Pension Act is a standardised retirement arrangement in which central government defines that a pension benefit is provided when clients reach 65 years of age. The central (supra-organisational)-level regulation defines how age is determined (by checking personal records) and how the benefit is disbursed. Such a governance arrangement is only feasible under highly demanding conditions as it requires sufficient information and expertise at the control level and street-level compliance with the regulation and definition of the policy objectives, or – the other way around – a sufficiently simple task that allows the management-level definition of policy goals and implementation.

(p.211) Addressing the need for expertise: professionalism

Addressing high complexity and facilitating the accompanying need for local expertise is structurally achieved by granting policymaking autonomy and implementation discretion to street-level officials, be they called street-level bureaucrats (Lipsky, 2010) or (hybrid) public professionals (Noordegraaf, 2007). The terms are not interchangeable; they differ in the extent that officials have ‘ownership’ of their expertise. Our central assumption of street-level bureaucracy being ‘two things’, however, holds that street-level bureaucracy necessarily includes aspects of both. When ambiguity is low, this may be geared towards support of the street-level official fulfilling his or her professionally defined role obligations, with management possibly enforcing competence control, but without intervening in the actual work process. When ambiguity is low, preferences and interests are aligned, so that street-level bureaucrats are assumed to naturally comply with management objectives.

The public professional literature distinguishes ‘pure’ forms of professionalism from ‘hybrid’ ones (cf Noordegraaf, 2007, p 2011). In the pure, traditional form, professionals are autonomous from organisational management and are accountable to professional peers and the professional association. This form of organisation is chosen because professionals command relevant information and expertise about important issues at the street level that management does not have (Abbott, 1988; Freidson, 2001; Bannink et al, 2006; Keast et al, 2006). Lipsky (2010, p 3) defines SLBs as ‘public service workers who interact directly with citizens in the course of their jobs, and who have substantial discretion in the execution of their work’. The inevitable human aspect of public service delivery to citizens means that – although street-level bureaucrats are ‘bureaucrats’ and need to implement rules; they can only be ‘semi-professionals’ (see Chapter One) – managers intentionally give discretion to street-level bureaucrats so that they may tailor the service delivery to the specific needs of the citizen concerned by applying their expertise to the case.

In effect, however, such an arrangement places implementation under ‘weak control’ (Héritier and Eckert, 2008) and may empower street-level actors, possibly allowing them to give priority to street-level interests at the expense of management interest. However, the other way around, discretion may also be given because managers intend to shift the blame for lacking implementation effectiveness to the street level (cf Evans and Harris, 2004).

(p.212) Addressing the need for alignment: managerialism

High ambiguity is addressed by measures to align preferences and interests. When complexity is low, alignment may be imposed by management onto street-level bureaucrats as little resistance may be expected. Management may choose to impose this alignment via rules, that is, bureaucracy, but when there is a natural discrepancy between the interests and preferences, elaborate control systems will be needed to enforce alignment. Hence, New Public Management (NPM)-type reforms were introduced, which aim to improve the performance of public service delivery while still controlling and prescribing which interpretation of preference and interest is to prevail. In managerialism, preference/interest alignment is, in a sense, ‘bought’ by the prospect of rewards when the imposed performance targets are achieved.

Such a control arrangement is aimed at influencing the choice options of street-level bureaucrats. Although the street-level bureaucrat is given implementation discretion, a specified output, or outcome, component is measured and rewarded (Brodkin, 2011). The establishment of a specified performance indicator supports increased management control over street-level performance, while ‘performance’ is defined in such a way as to conform to management objectives. Related to Brodkin’s (2011, p 254) observations, we argue that NPM, at the same time, assumes and underestimates the extent of autonomous street-level bureaucrat input in policy implementation. It assumes street-level bureaucrat input in the sense that compliance is actively bought through incentive-based governance. It underestimates the extent of street-level bureaucrat input where unintended effects of incentive-based control mechanisms are evoked. For example, the objective that is implied in the performance indicator may evoke ‘shirking’ behaviour, or such arrangements may give rise to a so-called ‘performance paradox’ (Van Thiel and Leeuw, 2002), in which street-level bureaucrats comply with the limited output components measured and rewarded in the system, while broader objectives disappear out of sight and may not be realised.

Addressing the double control challenge

In ‘double challenge’ situations, both the application of local expertise on a case-by-case basis and the alignment of preferences and interests are asked for. Extant theories of governance mechanisms fail to address these situations properly. The three governance mechanisms – as ideal types – each address one of the three cells in Table 12.1, but not the bottom-right-hand one with the double control challenge (see (p.213) Table 12.2). Where social risk complexity rises, market governance faces the performance paradox (Van Thiel and Leeuw, 2002), with disconnects (Bouckaert and Halligan, 2008) and blurred (Pollitt and Bouckaert, 2004) or unexpected outcomes (Hood and Peters, 2004). Where market issues come to inform professionalism, processes of closure (Abbott, 1988; Ackroyd, 1996) and double closure (Ackroyd and Muzio, 2007) indicate the increasing orientation of professional actors to organisational or individual interests.

Some authors have begun to explore possible solutions to the double control challenge (eg De Bruijn, 2007; Bouckaert and Halligan, 2008; Noordegraaf, 2011; Thomas and Hewitt, 2011). Network governance mechanisms are proposed by Klijn and Koppenjan (1997). An implicit assumption seems to exist in the network literature that interaction supports the mutual adjustment of interests and eventually action. This implies, we argue, that conflict is underappreciated: a constellation of interests is postulated that allows discursive mechanisms to produce effective cooperation. Where interests enter the debate – and they necessarily do – the horizontal coordination of networks becomes hybrid coordination. The network literature, of course, acknowledges the need to combine discursive methods of coordination and hierarchical ones, but the problems that the interaction of these mechanisms engenders seem to be underestimated. Furthermore, when they are discussed, as for instance by Klijn et al (1995), no new mechanism (comparable to incentives, enforcement or competence control) is introduced. The only ‘new’ aspect of network governance is a different constellation of actors that are supposed to engage in inter-organisational cooperation or institutional design. The proposed methods described to manage these constellations are bureaucratic (enforcement) or managerialistic (incentives). When problems of conflicting norms and perceptions are addressed, they opt for a top-down approach called ‘reframing’ in network structuring (Klijn et al, 1995, p 449). In this sense, network governance is nothing new and is, in itself, not able to deal successfully with combined complexity and ambiguity at the street-level bureaucrat level.

In the performance management literature, authors have proposed ways to address situations where accurate or unambiguous performance indicators are hard to come by (eg De Bruijn, 2007; Bouckaert and Halligan, 2008). The general gist is that managers and street-level bureaucrats collectively agree on the design of the performance management system, jointly make sense of the measurements, and agree on the conclusions drawn.

(p.214) From an opposite angle, in the professionalism literature, Noordegraaf (2011, p 1365) proposed the concept of organised or ‘connective’ (Noordegraaf, 2013) professionalism since ‘various changes in and around professional services … call for organized responses. … This means that so-called “secondary” aspects of service – efficiency, communication, cooperation, safety, reputation management and so on – become of primary importance, also for professionals.’ This builds upon a distinction between primary and secondary aspects of the task by Jamous and Peloille (1970). However, Noordegraaf does not give clear guidance on how to structurally address this control challenge, that is, what controls are appropriate.

In sum, we argue that there is currently no adequate conceptualisation of how to best address the double control challenge of simultaneous input of local, case-specific expertise and the alignment of preferences and interests. In so far as research has begun to address this, answers appear to include ‘discursive contestation’ (Thomas and Hewitt, 2011), ‘dialogue’ (Noordegraaf and Abma, 2003) and ‘diplomacy’ (Rhodes, 1997). However, while the ideal-typical form of dialogue – conflict-free suspension of and inquiry into beliefs and subjective knowledge (Isaacs, 1999) – may be feasible for highly complex situations, it may be very difficult to achieve in highly ambiguous ones.

We argue that management needs to in some way ‘integrate’ the objectives of bureaucratic, managerial and professional governance arrangements. Theoretically, this is difficult to devise. What, exactly, would such activity include? We turn to the analysis of a large number of empirical cases in order to shed light on this issue. In the next section, we present the results of an empirical study that inventoried the street-level control arrangements that were devised to address the double control challenge.

Research method and findings

The model developed earlier is applied to 42 cases where street-level bureaucrats perform public tasks that combine medium to high complexity with medium to high ambiguity. The ‘case’ (Yin, 1994) is comprised of the set of a task and a governance mechanism applied in a human service organisation. We studied the complexity and ambiguity of tasks and the governance mechanism (enforcement, incentives, competence control) that was applied. These cases were analysed by master’s students in the context of a seminar, ‘Governance of public and third sector organisations’, taught at VU University Amsterdam. Students were required to select a street-level organisation in which (p.215) ‘personal-interactive’ (Mills and Margulies, 1980) ‘human services’ (Hasenfeld, 1983) were implemented. Within these confines, the selection of cases was a convenience sample. In Table 12.3, an overview of all cases is given. The tasks vary widely: judges’ rulings, community policing, inspecting chicken factories, providing care at home, re-employment services, child protection services and so on. street-level bureaucrats are civil servants employed by public organisations (eg judges, community police officers and inspectors), employed by notfor-profit organisations (eg teachers, nurses and youth care workers) or employed by commercial organisations who won contracts to perform public tasks (eg re-employment services and providing care at home). All tasks are performed in the Netherlands.

For each case, on average, two street-level bureaucrats plus their supervisor were interviewed and relevant documents were studied. Data were collected by pairs of public administration master’s students enrolled in the seminar during the academic years 2009/10 and 2010/11. We only included the studies with good grades and removed cases where the ambiguity and/or complexity were not high enough.

Data

We analysed the data in terms of relatively high or low complexity and ambiguity on SLB task level. Complexity increases where tasks are more personally interactive (Mills and Margulies, 1980) and where information gathering is linked to client interaction (Van der Veen, 1995). Ambiguity is operationalised in terms of the possibility of competing normative assumptions about ends and the degree to which clients or peers are able to influence these normative assumptions. Ambiguity also rises where there is a relative vagueness of the goals to be achieved. Where street-level bureaucrats meet civilians, there is a certain degree of negotiation about goals between the street-level bureaucrat and civilian.

In the earlier discussion, the control mechanisms were discussed as ‘ideal types’. In our research, we measured the different control mechanisms in terms of their relative dominance within one organisation. Professional control is measured by looking at the level of freedom and trust that street-level bureaucrats are given by their managers in performing their task. High levels of freedom correspond with high levels of professional control. Bureaucratic control is measured in terms of the presence of a hierarchal structure and the amount of prescribed modes of working. A strong hierarchy and strictly prescribed task result in high levels of bureaucratic control. Managerial (p.216)

Table 12.3: Relative dominance of control mechanisms (professional organisations)

Case number

Applied control mechanisms

Professional

Bureaucratic

Managerial

24

Dominant

High

Low

18

Dominant

High

Medium

67

Dominant

High

None

46

Dominant

High

None

50

Dominant

High

Low

15

Dominant

Low

Upcoming

17

Dominant

Low

High

22

Dominant

Low

Low

29

Dominant

Low

Low

68

Dominant

Low

Low

8

Dominant

Low

Low

44

Dominant

Low

Low

2

Dominant

Low

Low

41

Dominant

Low

Low

65

Dominant

Low

Low

9

Dominant

Low

Medium

11

Dominant

Low

Medium

38

Dominant

Low

None

32

Dominant

Low

Upcoming

36

Dominant

Low

None

42

Dominant

Low

Low

39

Dominant

Medium

Low

48

Dominant

Medium

Low

30

Dominant

Medium

Low

3

Dominant

Medium

Low

21

Dominant

Medium

Low

35

Dominant

Medium

Medium

37

Dominant

Medium

None

31

Dominant

Medium

None

27

Dominant

Medium

Low

40

Dominant

Medium

Low

1

Dominant

None

Low

16

Dominant

None

Low

(p.217) control is operationalised in terms of output control and accountability. Where the amount of clients to be helped is prescribed and controlled for, and where other performance measures such as client satisfaction are used as a method of control, managerialism is high. The level of relative dominance of each mechanism is based on the estimations of our respondents. Thus, ‘dominance’ means the dominance as experienced by the street-level bureaucrats. We scored the influence of the different mechanisms that they experienced for each individual case. This means that the ‘high’ in one case does not have to be exactly the same ‘high’ in another case, but that in that specific case, there is a relatively ‘high’ amount of one mechanism experienced by a street-level bureaucrat compared with the other mechanisms.

To operationalise adequacy, we have limited data to work with. It is only in the report of street-level bureaucrats and, in many cases, their managers that clues about adequacy can be found. We could not measure whether organisations perform optimally (ie solve or diminish the social dilemmas that they have been set up for). However, in the interviews, street-level bureaucrats and their managers brought up conflict in some cases between different demands originating from different control systems. There are also cases in which no complaints or conflicts were reported.

Analysis

What types and constellations of mechanisms have we found? At first, it should be no surprise that ideal types have not been found. Instead, all organisations studied are a mix of several mechanisms applied simultaneously. In Table 12.3, the organisations are ordered on the basis of the dominant mechanisms at the street-level bureaucrat level. The professional mode of control is dominant within most organisations. Furthermore, almost all professionally dominated organisations also have some form of bureaucratic and managerialistic control systems. The relative dominance of control mechanisms among professionally dominated organisations in our case studies is presented in Table 12.3.

We first focus on those organisations where professional control is dominant. Only five cases (18, 24, 46, 50, 67) combine professionalism with a high level of bureaucratic control. Two of these cases (18, 46) are organisations providing elderly care. Case 50 is a criminal court. The bureaucratic control in both the elderly care cases takes the shape of detailed instructions about how to treat the clients under prescribed circumstances. The tasks performed here are mostly day-to-day routine; therefore, complexity in these cases is relatively low, while ambiguity (p.218) is also relatively low but not absent in the sense that street-level bureaucrats have to make judgements about the well-being of their client and adapt their services to clients’ needs where possible. If we do not take into account the question of adequacy, it is not surprising that, within our logic of suitability of control mechanisms, lower levels of complexity in practice could be suitable for more hierarchical control. The last case (a criminal court) is an exception. Complexity and ambiguity are extremely high for the street-level bureaucrats in this case: judges. However, the high level of bureaucratic control stems from the fact that in their judgements, the law itself is the cornerstone of their work and, as such, controls the way in which they come to decisions in the courtroom.

The same can be said for case 24, the Immigration and Naturalisation Service, where street-level bureaucrats judge refugees in order to grant them residential status. Again, here, law (constantly changing) prescribes a big part of decision-making, but street-level bureaucrats are left almost completely free in their judgements. Control over their decisions is mainly via peer review and sometimes through random selection by the supervisors. In case 67, the level of complexity is relatively low (day-care for the mentally ill and disabled). The task at hand is not described in detail, but is performed within a strong hierarchical structure. High bureaucratic control in this case also stems from the nature of the task.

Furthermore, there are a few cases where dominant professionalism is accompanied by a high level of another form of control. Apart from the former cases, only case 17 combines professionalism with high levels of other control mechanisms. This case is a social work organisation that aims at helping adolescents (aged 13–23) in order to prevent issues in the public domain of so-called problem neighbourhoods. The street-level bureaucrats’ task here is patrolling the streets looking for adolescents who skip school, have drug-related issues and so on – a highly complex and ambiguous task. They show a high level of inter-organisational action, which derives from their street-level position as key player/case manager. It is their responsibility to redirect their clients to other organisations and to monitor their behaviour in everyday life. Therefore, a central task is to communicate with other specialised organisations (schools, addiction clinics, etc). This is combined with a high amount of managerialistic control. The combination of control mechanisms leads to tension. The interviewed street-level bureaucrats report relatively high pressure from management for accountability in terms of how many clients they have helped and in what way.

(p.219) Other cases where the professional control is substantially accompanied with other forms of control are those with a medium level of bureaucratic control. Cases among these are police officers working in strong hierarchical structures, elderly care homes, inspectorates and home-care organisations.

In sum, most of the control mechanisms found seem to primarily respond to the nature of the task at hand. Where tasks are highly complex, professional control prevails. Organisations do not aim to integrate the requirements of complexity and ambiguity, but, instead, seem quite apt in designing a control strategy around the nature of the task that they have to perform at the street level. The organisations where professionalism is not the dominant control system, on the other hand, appear to opt for a bureaucratic control mechanism that also does not integrate the requirements of complexity and ambiguity, but is unilaterally aimed at street-level bureaucrat conformity. Table 12.4 displays the level of ambiguity and complexity among these seven cases.

Table 12.4: Relative dominance of control mechanisms (non-professional organisations)

Case number

Characteristics of SLB tasks

Applied control mechanisms

Complexity

Ambiguity

Professional

Bureaucratic

Managerial

14

Medium

High

Low

Some tasks high

Medium

7

Low

Low

Low

Dominant

Low

34

Low

Low

Low

Dominant

Upcoming

49

Medium

Medium

Medium

Dominant

Low

25

Medium

Medium

Medium

Dominant

Low

33

Medium

Medium

Medium

Dominant

Low/upcoming

20

Medium

Medium

High

Dominant

Low

45

High

High

High

Low

None

4

High

High

High

Low

None

In these seven cases, complexity is relatively low to medium, as is ambiguity. It may hardly be surprising that these organisations are dominated by bureaucratic control systems because the level of street-level bureaucrat expertise required in the task at hand is relatively low. Given their task, these organisations are able to emphasise the objective of conformity. The one organisation (14) where ambiguity is high shows a relatively high level of managerial control, which is the (p.220) adequate response to a control challenge of ambiguity, when complexity is low. The street-level bureaucrats in this case are restaurant managers whose task it is to promote neighbourhood cohesion. The ambiguity is high because there is no clearly defined idea of when cohesion is achieved and which organisations should participate. It is by no means a complex task however. The governing organisation has described in detail what the restaurant should look like, what should be served at what price and so on. More than in the cases with high complexity and ambiguity (where a double control challenge applies for which a single response has not yet been found), in the low complexity cases, the relative dominance of different control systems can be explained using the nature of the task as a starting point.

The same can be said for the two cases that do score high on both ambiguity and complexity. These organisations show a dominant orientation to inter-organisational governance. Thestreet-level bureaucrats support multi-problem individuals. In case 4, judicial partners (police, public prosecutor), social work-oriented professionals (psychologists) and teachers develop a customised approach for every individual client. Trade-offs are negotiated (ie obligatory psychological treatment instead of judiciary punishment) between the different partners. Dialogue between professionals (with expertise guaranteed through competence control) is the dominant control mechanism, which stems from the nature of the task.

There is, however, one control mechanism that does not seem to stem from the demands of the task at hand: managerialistic influences. Cases 9, 11, 18 and 35 report medium levels of managerialistic control combined with high levels of professional control. Cases 15 and 32 report an increasing importance of managerial control systems in professional environments. In cases 17, 33 and 34, where bureaucratic control is dominant, managerialism is also a profound factor. Furthermore, more than half the cases (25) report low managerialistic control.

As we have explained, this would be logical if complexity were low. However, that is not the case. Instead, in all cases where managerial control is found, this seems a relatively new influence that has been imposed from outside the organisations. In most cases, this takes the form of output control. Street-level bureaucrats report administrative tasks demanded by managers, who themselves have to account for the output that their organisation produces. To understand why these mechanisms have been introduced (and, in some cases, are being introduced), we need to look beyond the nature of the task and consider the institutional context of these organisations. It seems that (p.221) the policy trend of managerialism is at the root of the existence of these mechanisms. Theories of sedimentation (Cooper et al, 1996) and institutional layering (Thelen, 2004) point our attention to the idea that institutional change is an incremental and slow process in which new elements are attached to existing institutions. This view is particularly useful for our analysis because these managerialistic influences ‘do not come alone’. With the heightened attention to accountability and output control, bureaucratic control also rises. For the measurement of output, measuring tasks in a detailed way is necessary. To measure street-level bureaucrats’ performance, detailed descriptions of tasks are needed. In almost all cases where managerial control systems were implemented, street-level bureaucrats report a rising standardisation of work processes, which can be characterised as bureaucratic control. It appears that where control systems are imposed on organisations by governing actors or political policy developments, conflicts between task-related mechanisms and externally imposed mechanisms arise. This implies, in other words, that these organisations appear unable to solve the double control challenge that is imposed upon them. While the complex nature of the task requires a strong position for street-level bureaucrat expertise (professionalism), the output control imposed upon the organisation is an indication of ambiguity, requiring the alignment of street-level bureaucrat policy implementation to policy goals (managerialism). Instead of integrating these requirements, control mechanisms are simply combined, which leads to tensions in the policy implementation system.

Discussion and conclusion

In this chapter, we addressed the following central research question: which governance mechanisms are functional responses to the double control challenge of combined task complexity and ambiguity? We posed a theory on the relation between the variable of task characteristics (complexity and ambiguity) and governance mechanism (enforcement, incentives, competence control). Programmatically, we argued that the relation between street-level bureaucrat tasks and governance is, as yet, not fully understood. In order to improve our understanding of this issue, we studied governance mechanisms in a large number of cases where task complexity and ambiguity were high.

High complexity of tasks and the task environment create a high degree of discretion for individual street-level bureaucrats and blocks the effectiveness of rule-based and, to some extent, incentive-based governance mechanisms. Where control incentives are strong, (p.222) however, political and organisational managers tend to apply rule-based or incentive-based control mechanisms anyway, even if task characteristics block the effectiveness of these controls. This normally leads to tension and relatively strong unintended effects with respect to case treatment and/or employee well-being. This indicates that the appropriateness of such methods is limited. In eight cases, street-level bureaucrats and their managers openly complained about the bad fit between different control mechanisms (see earlier). We observe a dilemma in which political and organisational leaders need to choose between allowing discretion in order to respond to task complexity and controlling task implementation through rule- or incentive-based governance mechanisms in order to retain control. In other words, we have observed that it is difficult to design a control mechanism that forms an appropriate response (defined here as a control mechanism that does not lead to tension) to the double control challenge of combined complexity and ambiguity. In some way, the control mechanism needs to integrate the requirements of complexity (expertise) and ambiguity (alignment).

Shielded professionalism

In one third (14) of the cases, managers have found a way to shield SLBs from the external pressures of upcoming managerialistic and bureaucratic systems. This solution does not imply the integration of control challenges, but, instead, is best described as ‘decoupling’ (Meyer and Rowan, 1977; see also Chapter Thirteen). Five of these cases are in (lower, higher or special) education. In these cases, managers explicitly report that their teachers should be free to teach and spend all the time they have on their primary task. The street-level bureaucrats in these cases are aware of this shielding and praise their managers for it. Problems with conflicting mechanisms, therefore, do not seep through to the street-level bureaucrat level. In case 11, the direct supervisor of the street-level bureaucrats even openly admits to manipulating output indices to create more freedom for the street-level bureaucrats.

Shielded professionalism strongly acknowledges the complexity challenge of governance arrangements, but underappreciates high ambiguity. This mode of control is aimed at the maximised utilisation of professional expertise through allowing professionals or street-level bureaucrats to infuse the work process with their own values and preferences. The demanded accountability and transparency, however, is not being fully provided. In these organisations, street-level bureaucrats (p.223) and managers seem to view control and professional autonomy as mutually exclusive alternatives.

Integration

A number of other cases also showed a rather ‘organic adaptation’ (Hernes, 2005) to conflicting demands. In these cases, management appeared able to devise a control mechanism that responds to the challenges of both complexity and ambiguity, without leading to tension in the implementation system. This is what we call the ‘integration’ of control challenges. In a number of small or start-up consultation firms, the owner of the firm operates both as the top managerial official and a co-worker, operating as a primus inter pares among the workforce. While we have observed that a response was devised to the combined control challenge of complexity and ambiguity (and so, in a functional sense, have observed that integration has apparently been achieved), we do not fully understand the structure of the mechanism that is applied. This suggests two important implications. In the first place, it actually seems to be possible to devise a mechanism that integrates control functions in response to the double governance challenge of combined complexity and ambiguity without leading to tension. However, and this is the second conclusion, our understanding of the structure of such a mechanism is still underdeveloped.

Although it appears difficult to define the structural characteristics of the governance mode that these managers apply, a number of central structural characteristics of the context in which it is produced can be identified. Management and co-workers simultaneously define the governance arrangement and its application is also strongly bound to the personal interaction between the management and street-level bureaucrats. This seems an echo of an already-old idea by Maynard-Moody et al (1990, p 844), who argued: ‘Moving beyond Lipsky’s thesis, our exploratory research suggests that delegating authority and including the perspectives of street-level workers in programmatic decisions is one realistic alternative to managerial control when the objective is to reduce the dangers of discretion.’

Lacking knowledge, alternative approaches?

Extant theories on how street-level bureaucrats should be controlled are found to be lacking in situations of the double control challenge. The double control challenge is about achieving both the expertise to address high complexity and the alignment to overcome high ambiguity. (p.224) We argued that bureaucracy, managerialism and professionalism do not properly address the double control challenge. The empirical results support our conceptual analysis. We have found support for our argument that the double control challenge of simultaneous high complexity and high ambiguity is important in street-level bureaucracies and that it is difficult to devise a response. In most cases, one of the challenges is simply ignored. Where a response to both control challenges is devised simultaneously, tensions occur because some of the mechanisms conflict with each other. We found very few organisations that have found appropriate solutions.

How may the expertise that street-level bureaucrats possess be applied in ways that align the interests and preferences of street-level bureaucrats and their managers? The assumption that control engenders dysfunctional effects also appears to be dominant in the professionalism literature; formal controls are assumed to crowd out intrinsic street-level bureaucrat motivation. When managers impose controls on professionals, professionals see this as a sign of distrust; when managers grant professionals autonomy, no controls may be imposed. The ‘Third Logic’ (Freidson, 2001) of professionalism is a logic of horizontal coordination of substance, without the application of hierarchical controls. Power (1997) even argued that the insertion of controls in governance relations evokes the further insertion of even more controls in order to control the application of the controls.

Formal controls generate resistance and dysfunctional effects when street-level bureaucrats perceive them as coercive (Adler and Borys, 1996), restrictive (Elias, 2009) or constraining self-determination (Weibel, 2010). When this happens, ambiguity is not overcome in a productive way as intrinsic street-level bureaucrat motivation is likely to be crowded out (Frey, 1997; Weibel, 2010). Controls may also be perceived by street-level bureaucrats in more constructive ways, enhancing intrinsic motivation and organisational value internalisation (Weibel, 2007, 2010). This happens when controls are perceived to be enabling (Adler and Borys, 1996), promotive (Elias, 2009) or supportive of self-determination (Deci and Ryan, 2000; Ryan and Deci, 2000; Weibel, 2007, 2010). The more street-level bureaucrats are able to participate in the development of the formal controls, the self-determination theory poses, the more their self-determination is enhanced.

This might actually resemble the structural form of organisational governance that the managers of the small firms discussed earlier apply. ‘Self-defined’ control may, on the one hand, force street-level bureaucrats to conform to organisational demands (responding to the (p.225) ambiguity problem) while, on the other, support their capacity and their preparedness to infuse the work process with their expertise (the complexity problem). This supports self-definition (related to SLBs’ need for autonomy), but also control (related to external demands). Self-defined control is also related to Noordegraaf’s (2013) concept of the ‘connective professional’. The emphasis De Bruijn (2007) and Maynard-Moody et al (1990) placed on the need for SLBs to be involved in programmatic decisions, such as the design and evaluation of performance, supports self-determination theory. This is, indeed, what we see in a few of the smaller organisations in our sample. Further research on the applicability of governance mechanisms related to self-determination theory and related approaches might shed more light on the structural form of these mechanisms. (p.226)