31
1 How are Arm’s Length Relationships Instrumentalised in Europe? A Three Country Comparison Katherine Tonkiss, Sandra van Thiel, Koen Verhoest and Chris Skelcher 1 Paper to be presented at the ECPR General Conference, 4 th -6 th September 2014, Glasgow When governments delegate functions to agencies, they seek to manage this relationship so that the risks of opportunism are controlled while the benefits of agency autonomy are not undermined. ‘Agency theory’ approaches seek to provide an account of relationships between governments (principals) and their agencies (agents). Critics have asserted, however, that agency theory over-simplifies principal-agent relationships by assuming that the principal is singular and unitary, while studies have shown the existence of multiple principals with relationships to a single agent. The public sector literature on multiple principals is not well established and tends to focus on the US political system. Less is known about how a multiple principal model might transfer into other political contexts, such as the parliamentary democracies of Western Europe. In this paper, we present a comparative analysis of a particular subset of multiple principal relationships in three European countries. By taking a comparative perspective, we are able to illustrate how one particular aspect of context – that of political culture – affects the shape of multiple principal relationships. Introduction When governments delegate functions to agencies, they seek to manage this relationship so that the risks of opportunism are controlled while the benefits of agency autonomy are not undermined. They relinquish direct control over the delivery of key policy objectives to agencies, on the basis that this should embody a credible commitment to those objectives (Bertelli 2006a; 2006b), but often struggle with exactly how to develop a relationship with those agencies to ensure that they do meet these policy objectives in an efficient and effective way. This problem is typically thought of as a principal-agent relationship, where the principal must decide how best to manage the relationship with the agent in order to ensure the effective delivery of policy objectives. ‘Agency theory’ approaches seek to provide an account of relationships between governments (principals) and their agencies (agents). They assume that the interests of the principal and the agent will diverge, and so the principal will implement measures 1 Dr. Katherine Tonkiss, Institute of Local Government Studies, University of Birmingham, UK ([email protected]); Professor Sandra van Thiel, Department of Public Administration, Radboud University, the Netherlands; Professor Koen Verhoest, Department of Political Science, Antwerp University, Belgium; Professor Chris Skelcher, Institute of Local Government Studies, University of Birmingham, UK.

How are Arm’s Length Relationships Instrumentalised in Europe… · 2014-09-01 · How are Arm’s Length Relationships Instrumentalised in Europe? A ... which principals seek to

Embed Size (px)

Citation preview

1

How are Arm’s Length Relationships Instrumentalised in Europe? A

Three Country Comparison

Katherine Tonkiss, Sandra van Thiel, Koen Verhoest and Chris Skelcher1

Paper to be presented at the ECPR General Conference, 4th-6th September 2014, Glasgow

When governments delegate functions to agencies, they seek to manage this

relationship so that the risks of opportunism are controlled while the benefits

of agency autonomy are not undermined. ‘Agency theory’ approaches seek to

provide an account of relationships between governments (principals) and

their agencies (agents). Critics have asserted, however, that agency theory

over-simplifies principal-agent relationships by assuming that the principal is

singular and unitary, while studies have shown the existence of multiple

principals with relationships to a single agent. The public sector literature on

multiple principals is not well established and tends to focus on the US

political system. Less is known about how a multiple principal model might

transfer into other political contexts, such as the parliamentary democracies

of Western Europe. In this paper, we present a comparative analysis of a

particular subset of multiple principal relationships in three European

countries. By taking a comparative perspective, we are able to illustrate how

one particular aspect of context – that of political culture – affects the shape of

multiple principal relationships.

Introduction

When governments delegate functions to agencies, they seek to manage this

relationship so that the risks of opportunism are controlled while the benefits of agency

autonomy are not undermined. They relinquish direct control over the delivery of key

policy objectives to agencies, on the basis that this should embody a credible

commitment to those objectives (Bertelli 2006a; 2006b), but often struggle with exactly

how to develop a relationship with those agencies to ensure that they do meet these

policy objectives in an efficient and effective way. This problem is typically thought of as

a principal-agent relationship, where the principal must decide how best to manage the

relationship with the agent in order to ensure the effective delivery of policy objectives.

‘Agency theory’ approaches seek to provide an account of relationships between

governments (principals) and their agencies (agents). They assume that the interests of

the principal and the agent will diverge, and so the principal will implement measures

1 Dr. Katherine Tonkiss, Institute of Local Government Studies, University of Birmingham, UK

([email protected]); Professor Sandra van Thiel, Department of Public Administration, Radboud University, the Netherlands; Professor Koen Verhoest, Department of Political Science, Antwerp University, Belgium; Professor Chris Skelcher, Institute of Local Government Studies, University of Birmingham, UK.

2

to reduce the risk of opportunism on the part of the agent (Caers et al. 2006; Shapiro

2005; Strom 2000). Three key issues have, however, been raised by critics of agency

theory. The first is that the approach overlooks the role of context, rules and norms

which may shape the behaviour of principals and agents (Bruce et al. 2005; Davila and

Elvira 2012; Huber 2000; Thatcher and Stone Sweet 2002), the argument being that

different contexts may affect behaviour in different ways and therefore it is more

difficult to make generalisations about the behaviour of principals and agents in

different contexts. For example, ‘stewardship’ theory highlights how some agents act as

stewards, attempting to build trust with the principal and working in the general

interest rather than for self-interest (cf. Schillemans 2012). The second is that, because

research into principal-agent relationships was developed in economic theory, this

approach has only limited applicability to the study of public administration where

principals and agents act in different ways as a result of political and bureaucratic

structures (Miller 2005; Waterman and Meier 1998).

The third issue is that agency theory over-simplifies principal-agent relationships by

assuming that the principal is singular and unitary. Studies have shown the existence of

multiple principals with relationships to a single agent (Holmstrom and Migrom 1991;

Miller 2005; Moe, 1984; 1987; Skelcher 2007; Waterman and Meier 1998). While there

is an extensive literature on multiple principals in the private sector, the public sector

literature is less well established and tends to focus on the US political system (Kiser

1999; Miller 2005) or transnational integration in the EU (Bergman 2000; Egeberg

2006). Far less is known about, and far fewer empirical studies exist on, how a multiple

principal model might transfer into other political contexts, such as the parliamentary

democracies of Western Europe.

In this paper, we address all three of these issues through a comparative analysis of a

particular subset of multiple principal relationships in three European countries. We

term this the ‘complex principal’ model, a particularly under-explored type of multiple

principal relationship which exists when a number of different principal-agent

relationships occur within one over-arching relationship. This is particularly evident in

the case of the governments delegating functions to arm’s length agencies, because

agencies commonly have relationships with many different ministries (and units within

ministries) which are instrumentalised in different ways. By taking a comparative

perspective, we are able to examine how one particular aspect of context – that of

political culture – affects the shape of the complex principal model.

Through this comparative analysis, we explore (1) the relevance of the complex

principal model in three different political cultures, (2) the types of complex principal

model evident in each case study, and (3) how and why the relationships(s) between

the principal and its agencies are instrumentalised in each context. The three

countries forming our dataset are the UK, the Flanders region of Belgium and the

Netherlands. We adopt these countries for our comparison because they each represent

very different political cultures and so offer suitable conditions for comparison; and

3

because we hold a substantial data archive on arm’s length relationships in these

countries.

Our intention with the comparative analysis is not to deliver a generalizable theory of

arm’s length relationships, but rather to identify and explain different forms of arm’s

length relationships in terms of the prevailing cultures of government. This intra-

European perspective offers new insights that help us to refine the conceptualisation

and theorisation of these relationships, which currently are largely derived from the US

context.

We first review the literature on the agency problem and multiple principals, and then

justify our comparative research design and the focus on the case of arm’s length

agencies. Next, we present an empirical account of the composition of the complex

principal in each country, the institutionalised relationships with arm’s length agencies,

and the political and governmental cultures that motivate the behaviour of this system.

Finally, we draw conclusions regarding the relationships between complex principles

and agencies and how this can inform further study.

The agency problem and multiple principals

Theorising the relationship between governments and their arm’s length agencies has

been dominated by agency theory (James & Van Thiel 2011). This approach seeks to

provide an explanatory account of how such relationships are characterised. Agency

theory assumes that the interests of the principal (government) and agents (agencies)

will diverge, and this may give rise to goal conflict.2 It is then assumed that the agent

will act in accordance with their own interests, thereby not acting to achieve the

interests of the principal (Caers et al. 2006; Eisenhardt 1989; Shapiro 2005). Therefore,

the principal will seek to control the agent through ex ante contractual agreements and

ex post bonding, monitoring and incentives (Bertelli 2006b; Bhatta 2003).

According to agency theory, the main problem faced by principals in delegating to

agents is that of information asymmetry: that the principal may now have all of the

information about the agent at its disposal. Information asymmetry is problematized in

two key ways. The first is known as moral hazard, and this occurs when the agent

withholds information about its potential performance which impacts negatively on the

interests of the principal; and the second is adverse selection, which occurs when the

agent withholds information about its potential performance at the time that the

contractual agreement between each party is under consideration (Dixit 2002).

Information is therefore viewed as a commodity and information systems are critical in

order for the principal to control the risk of agency opportunism (Eisenhardt 1989).

2 We note that this ‘may’ give rise to goal conflict with reference to the stewardship literature, which has

suggested that goal conflict may not always exist, and that agents may act in the collective interest (Caers et al. 2006; Davis et al. 1997; Schillemans 2012).

4

There are three issues which are raised recurrently by critics of agency theory. The first

is that the approach overlooks the role of context, rules and norms which may shape

behaviour, thereby reducing the power of agency theory as an explanation of principal-

agent relationships (Bergman 2000; Bruce et al. 2005; Davila and Elvira 2012; Huber

2000; Kiser 1999; Koppell 2003; Pollitt et al. 2001; Skelcher 2010; Thatcher and Stone

Sweet 2002). Agency theory attempts to predict the behaviour of actors in given

situations, but often this neglects the effect that different environments will have on

that behaviour, making generalizable predictions less likely. In this paper, we address

the issue of context by engaging in a comparative study which draws attention to the

ways in which different political cultures can affect the relationship between principals

and agents.

The second issue is that, because the majority of principal-agent models were

developed in economic theory, this approach has only limited applicability to the study

of public administration where principals and agents act in different ways as a result of

political and bureaucratic structures and incentives. For example, research has ground

that, away from the markets and competition which characterise economic models, it is

bureaucratic processes rather than incentives which are the most common means by

which principals seek to control agents in the public sector, that negotiation between

principal and agent is more common in such contexts than in private sector

organisations, and that goal conflict is less likely to exist between principals and agents

in the public sector (Miller 2005; Waterman and Meier 1998). Through its comparative

study, this paper offers further detailed insights on principal-agent relationships within

studies of public administration.

The third issue is that agency theory over-simplifies principal-agent relationships, in

particular by assuming a singular, unitary principal. Studies highlight that, in reality,

there are often multiple principals; and that principals and agents may have multiple,

overlapping relationships as a result of the delegation of multiple tasks (Holmstrom and

Migrom 1991; Miller 2005; Moe 1984; 1987; Skelcher 2007; Waterman and Meier

1998). The multiple principal model has been highlighted in particular by the private

sector corporate governance literature, where multiple principal situations may arise as

a result of the divergent interests of multiple shareholders, or may arise from complex

organisational arrangements such as mergers, multinational corporations and strategic

alliances (Child and Rodrigues 2003; Davila and Elvira 2012; Ingley et al. 2012; Kiser

1999).

The multiple principal literature on the public sector is less well established, although

the idea has been considered particularly in relation to the US political system, where

the separation of powers has been shown to give rise to multiple principals – for

example, between the White House and the Senate (Kiser 1999; Miller 2005). It has also

been applied to a limited extent to the European Union, with national governments as

agents of multiple principals in the transnational space (Bergman 2000; Egeberg 2006;

5

Egeberg and Trondal 2009).3 Far less is known about, and far fewer empirical studies

exist on, how the notion of multiple principals might translate to other political

contexts, such as the parliamentary democracies of Western Europe.

Furthermore, little attention has been paid to a particular subset of multiple principal

relationships, which might be termed the ‘complex principal’ model. Bureaucracies are

complex and it is common for an agent to have many relationships with different

ministries (and units within ministries) of the same government, instrumentalised in

many different ways. In such a complex principal model, there are multiple principal-

agent relationships within one overarching principal-agent relationship between

government and agent (Fig. 1; see also Waterman and Meier 1998). This paper aims to

address this gap by exploring the existence of complex principals as an under-

researched type of multiple principal. Unlike other studies of multiple principals which

have tended to examine goal divergence among principals and how this affects their

relationship with agents, our focus is on modelling how relationships between complex

principals and agents are constructed and instrumentalised. Our comparative

perspective enables us to understand more about how context – and specifically

political culture – affects the shape of these complex principal relationships.

Figure 1. Complex Principal Model

Comparative Research Design

The value of a comparative research design is to expose to empirical challenge the

prevailing way of conceptualising and theorising a phenomenon. In our case,

comparison enables us to introduce in an explicit manner the ‘political and

governmental culture’ variable into the study of the relationships between political and

administrative principals and arm’s-length agencies. Much of the literature on this

phenomenon is based implicitly or explicitly on the US case. The countries selected for

the study are the Flanders region of Belgium, the Netherlands and the UK. In Belgium,

we focus specifically on Flanders. This resolves two potential problems in studying this

3 See also Kassim and Menon (2003) on the limitations of EU-focused multiple principal research.

Sub-principal Sub-principal Sub-principal Sub-principal

Agent

6

federal country: the complexity of exploring different sets of agencies with different

relationships to the federal level of government, and to the Flanders, Walloon and

Brussels regional governments (see Verhoest et al. 2011). Culturally, the differences

between these regional systems are profound, and much more than would be found, for

example, between states in the US. In the UK, our focus is on agencies for which the UK

government is responsible; we do not discuss those that are within the competence of

the devolved administrations in Northern Ireland, Scotland and Wales, again in order to

reduce the complexity of differing political and governmental cultures.

Blondel offers a note of caution about the process of explanation in comparative

research:

So much of political life is viewed as shaped by the particular characteristics of

each country that cross-national endeavours are sometimes regarded with

suspicion as likely to lead to misleading interpretations. The point has some

validity, so long as it is a warning only: we must be careful when drawing

conclusions on the basis of comparisons.' (Blondel 2005: 186)

His point is that we should not be over-concerned with cultural specificity, and thus

detract our attention from the regularities that cross political cultures. We respond to

this point by isolating points of similarity across the cases and a point of difference, in

order to explain the object off enquiry – the instrumentalisation of relationships

between complex principals and arm’s length agencies.

First, the three countries exhibit points of similarity with regard to relevant features of

the structure of government. They are parliamentary democracies in which the

government is created by the majority party or coalition, and where departments are

headed by ministers who sit within the parliament. Further, all three countries make

extensive use of arm’s length agencies, of which there are multiple types undertaking

advisory, executive, regulatory and quasi-judicial functions (Van Thiel 2012). We focus

on those corporate entities (i.e. bodies that have their own legal status) that are created

by government to undertake executive tasks on its behalf and that (in theory at least)

have considerable discretion within the policy and accountability framework overseen

by their parent ministry. In Belgium we include the group of public law agencies that

have their own governing board, and are only under ministerial supervision which is

constrained by their enacting statuses. In the Netherlands we will focus on the so-called

ZBOs (in Dutch: zelfstandige bestuursorganen), and in the UK we include executive non-

departmental public bodies (ENDPBs). This gives a comparable set of agencies in terms

of their legal status, function and degree of delegation from their ministry.

Secondly, each country has a different political/governmental culture. In Lijphart’s

(2012) analysis, the UK is a classic Westminster majoritarian democracy, while Belgium

and the Netherlands are consensus democracies. However although the Flanders region

has linguistic and social cultural similarities with the Netherlands, their political

cultures differ. Flanders has a considerable diversity of political parties leading to

7

instability in the governing coalition. The Netherlands is positioned somewhat

between the UK and Flanders as a result of its ‘polder model’ weakening in recent

decades, although Kickert (2003), Hendriks and Toonen (2000) and others argue that

the legacy of pragmatic consensus politics is still evident. These differences are also

evident in Hofstede and Hofstede’s (date) analysis of national cultures (table 1). We

draw particular attention to four dimensions (indicated in bold). These reveal

considerable differences between Belgium and the UK, with the Netherlands being

closer to Belgium on some variables and in others closer to the UK.

Table 1: Case study country characteristics

Country Power distance

Individualism Masculinity Uncertainty avoidance

Welfare system

Ministerial accountability

Belgium 65 75 54 94 Social democratic

Collectively

Netherlands 38 80 14 53 Social democratic

Individually and collectively

UK 35 89 66 35 Liberal Individually and collectively

Source: Verhoest et al. 2012, ch 1, in Verhoest et al. Original data derived from various

sources.

The data presented in our country analysis is drawn from the large data archives held

by the authors in relation to their own country. Different data sources, including

documents, interviews and quantitative studies, were used for each of the countries. We

outline data sources for each country in each of our case study sections below.

Analysing Relationships in Three Countries

Arm’s Length Relationships in Flanders (Belgium)4

As mentioned previously, we choose to elaborate upon the relations between agencies

and principals within the Flemish public sector, rather than focusing on the federal level

or attempting to offer a general description that applies to common practices at both

the federal level and the different regions. The motivations for this are three-fold. First,

in Flanders the relations between agencies and their principals to a large extent show a

more pronounced manifestation of patterns than in the two other regions and at the

federal level. However, due to past reforms, agency governance is to some extent more

streamlined and more institutionalised, and less ad hoc compared to the other levels.

4 For the Flanders case, we drew on six group interviews with top civil servants and high-level members of

political staff of the Prime Minister and of the Ministry of Public Governance, as well as nine individual interviews, conducted in 2013. We also used data from a 2013 COBRA survey, which surveyed all agency heads about their relations with their minister and government and had a 75% response rate.

8

Moreover, relations between government and agencies have been given much more

attention than in other regions and at the federal level both by the administration and

by academics, resulting in much better knowledge about practices.

Before we move on to describing and categorising relations and instruments, we point

to four specific characteristics which make the Belgian/Flemish case distinctive from

the two other countries considered in this paper. First, at all governmental levels there

is a longstanding tradition of ministers having a large political staff, called a ‘cabinet de

ministre’, which formulates and controls policy-making and even policy implementation

by the administration on behalf of the minister. These ‘cabinets de ministre’ are not

uncommon in Latin countries (such as France and Italy) and their role should not be

underestimated as it goes much further than guarding the political line and priorities of

the minister. In fact, these ‘cabinets de ministre’ have been claimed by several observers

to act as a parallel administration, taking over much of the coordinating and controlling

roles which ministerial departments often have in other systems. Despite deliberate

attempts to slim down these ‘cabinets de ministre’ in Flanders, they still play a very

strong role.

A second feature is the fairly high level of politicisation of top civil servant positions

(both secretary generals of departments and chief executives of agencies). There is a

continuing preference of governments to appoint top civil servants on the basis of

political affiliation (next to other criteria such as merit, leadership skills and

experience) with the political parties in power, creating trust between these top civil

servants and ministers of the same political family. However, this practice often leads to

problematic relations between agency chief executives and subsequent ministers with

different political affiliations, or between agencies and departments where top civil

servants represent different political families (Brans et al. 2005). The consequence is, in

many cases, an institutionalised distrust between ministers and senior management in

department and agencies which encourages the persistence of large ‘cabinets de

ministre’ where each minister must micro-control his administration.

A third important feature to understand the principals-agency relations in Flanders is

the institutionalisation of a non-hierarchical relationship between departments and

agencies in the same policy field. At all levels in Belgium, the so-called parent

departments do not have a clear and generally accepted role in the oversight of

agencies, leading to a longstanding struggle between departments and agencies for

more influence. This is related to the issue of political appointments; the system of

‘commissaire de gouvernement’, who have no link to the parent departments; a lack of

expertise within these parent departments; and the preference of agencies to deal

directly with ministers, and of ‘cabinets de ministre’ to deal directly with agencies,

bypassing the departments (Rommel and Christiaens 2009). In the Flemish agency

reform of 2006, the relationship between departments and agencies was even

stipulated as non-hierarchical. Although departments were assigned by the BB-decree

to have some role in supervising the agencies, this role is highly contested. Overall, the

9

relationship between agencies and the oversight authority is mainly oriented towards

the minister and his political staff.

A fourth and last specific feature is that the cabinet acts as a collegial body, making the

full cabinet responsible for decisions taken by individual ministers. This means that the

extent of delegation from the government towards individual ministers is rather small,

and most decisions have to be formally approved by the cabinet before taking effect.

Moreover, it also implies that individual ministers have limited leeway in designing

their own instruments and methods to deal with agencies, even more so because due to

the legalistic Napoleonic tradition most instruments and methods to deal with agencies

are described in and prescribed by law. Contrary to the Netherlands, differences

between ministerial portfolios with respect to principal-agency relationships are small,

and mainly relate to how ministers deal with the legally prescribed oversight

instruments. However, in Flanders, and in Belgium more generally, differences between

formally prescribed controls and actual controls can be substantial (Verhoest et al.

2004).

Figure 2 shows the different relationships between political principals and the public

law agencies with a governing board. Similarly to the two other countries, agency chief

executives are accountable to their governing board, which in principle sets out the

strategic objectives of the agency. In practice however, agency chief executives have a

large influence over this process. Recently, due to more emphasis being placed on good

governance practices, the supervisor role of agency governing boards has moved closer

to international standards, for example having a separate sub-commission for internal

audit.

10

The Cabinet Appointments of agency CE’s / general policy objectives / overall budget

Minister of public Governance (and his political staff) Coordination of

government-wide management issues

Information gathering

Parliament Legislation creating/reforming (all

bodies) Annual reporting (all bodies)

Budget setting

Public Law Agency

Parliamentary Committees

Inquiries and ad hoc hearings

Portfolio minister (and his political staff)

Policy notes and policy briefs Negotiation and monitoring

performance contracts and annual business plans

Yearly objective-setting and evaluation of agency CE

Court of Office

Financial auditing and VFM studies

Board

Strategic objectives

Budget oversight

Annual report

Public Law Agency chair

Overall performance

Chief Executive

Accounting Officer Operational objectives

Budget oversight

Minister of Finance (and his political

staff) Budget negotiations

Annual budgets Spending controls

Figure 2: Arm’s Length Relationships in Belgium/Flanders

Prime Minister (and his political staff) Coordination and

monitoring of transversal policies and cross-cutting

programs

Regulators (few bodies)

(within public sector or

linked to markets)

Financial performance

Policy delivery

Outcomes

Public Governance Department

Line Department

Prime Ministers’

Department

Finance Department

‘Commissaires de

Gouvernement’

and tutelage

11

We now turn to describing the relationships between the agency and its political

principals. As will be elaborated, the agencies are in direct contact with several

ministers (or at least their ‘cabinets de ministre’ and services), but let us first focus on

the relationship with the portfolio minister under which the agency resides. The

portfolio minister steers and controls the public law agencies under his or her remit in

four main ways.

First, agencies are obliged to contribute to the policy objectives which the minister

formulates in his long-term policy note (for the duration of the cabinet period: 5 years)

and the yearly updated policy letter, which are presented before parliament. The formal

instrument by which the organisational objectives of an agency are aligned with the

objectives of its portfolio minister is the long term performance contract, in which the

agency and minister negotiate which activities and outputs the agency has to deliver

and under which conditions (e.g. regarding autonomy). These performance contracts

can be discussed in parliament before being signed by the different actors: the minister

on behalf of government and the chair of the governing board (or agency chief

executive) on behalf of the agency. This potential for parliamentary involvement in

shaping an agency’s contract is quite unlike the position in the Netherlands and UK,

where it is a matter for the responsible minister. Every year the agency has to translate

the strategic objectives of the performance contract and the ministerial demands in the

policy letters into annual business plans, which are increasingly submitted for silent

approval by the minister. Agencies report about progress towards the objectives stated

in the performance contracts by making biannual or annual reports to the minister and

government. Only in rare cases do the departments residing under the portfolio

minister have an actual role in this monitoring procedure by providing an independent

evaluation. Mostly, only the cabinets de ministre and the minister are involved in the

procedure, without much real effort put into actual monitoring and evaluation of these

contractual documents.

While the performance contract and related documents can be seen as instruments for

ex post result-oriented control, there are still some quite traditional instruments in

place. In practice, performance contracts and annual business plans are only used to a

limited extent by political principals to steer agencies. Most ministers impose new tasks

or assignments on agencies by formulating ad hoc requests outside the existing

framework and procedures related to the performance contracts and business plans,

when the political need arises to do so (such as when a crisis occurs or new issues of

high political salience arise). For agency chief executives, this practice of ad hoc

demands make planning within their organisations a very challenging task. In the UK

case, in contrast, ad hoc demands are less prevalent; the greater stability of

governments in the UK compared with the Flanders/Belgian case puts less pressure on

ministers to intervene in agencies as a matter of course, although they do when crises

emerge.

12

Although the agency reform in 2006 proclaimed a full shift to ex post result control, an

important remaining ex ante control instrument is the traditional general tutelage

which was inspired by the French system, implying that all decisions of governing

boards are screened ex ante for conformity with the law, budget and ‘the general

interest’ by control agents (the so-called ‘commissaires de gouvernement’ as members

of the governing board) of the portfolio minister and the Minister of Finance. Based on

this screening the involved ministers can suspend decisions of the agency before they

take effect. These ‘commissaires’ were abolished by the 2006 reform, but re-installed

two years later in most public law agencies with a governing board. The ‘commissaires’

provide the portfolio minister and Minister of Finance with a direct view of decisions

taken and provide them the instruments to intervene if necessary. This traditional

control of agencies by the involved ministers is quite strongly oriented towards the

legality and budgetary conformity of decisions, and much less attention goes to the

content and strategic orientation of agencies. As such, this system of tutelage reduces

the managerial autonomy of agencies and is at odds with the basic philosophy of results

control and managerial autonomy implied by performance contracts (Verhoest et al.

2012). This legal orientation has some parallels with the position in the Netherlands,

but much less with the UK where although legal authority is important this is

moderated by a pragmatic approach to legal interpretation and to constitutional

matters in general.

A third group of instruments are related to structural control, that is: control through

the appointment, steering and evaluation of the governing board and agency chief

executives. Governing boards are composed of governmental representatives, and non-

governmental stakeholders, in different compositions. The government representatives

are selected by the portfolio minister based on their expertise and political affiliation,

and approved by cabinet. However, in general the Flemish governmental

representatives do not get detailed instructions from the minister on how to behave in

the agency board and only would consult their minister rarely in important decisions.

This overt political selection and cabinet approval of board members contrasts with the

UK, where partisan influence is more covert and moderated by an independent

recruitment and short-listing process, and the decision is in the gift of the minister.

More importantly is the appointment, steering and evaluation of agency chief executives

which gives the cabinet and portfolio minister a direct means of influencing the chief

executives, even bypassing the governing board in some cases. As noted above, political

affiliation, besides merit and management skills, still matters when it comes to the

appointment of agency chief executives, quite unlike the position in the UK where

neutral competence is still regarded as the core principle for such appointment.

Moreover, yearly the minister – on behalf of the cabinet – discusses the annual

objectives with the individual agency chief executives and an assessment procedure

provides for the evaluation.

A last set of instruments are related to formal steering of meetings and informal

contacts. Ministers can have (bi-)monthly formal steering meetings with all of the

13

agency chief executives and secretary-generals in the policy council, although this

platform functions only in a minority of the policy portfolios. Also, informal contacts

between ministers – or their political staff – and agencies can be quite intense,

depending on the extent to which the agencies deal with politically salient issues or are

heavily involved in policy-making.

However, the portfolio minister is clearly not the only political principal, as the agencies

have direct control and steering relations with other ministers and their political staff

too. The Minister of Finance is very prominent in this respect, holding the key to

budgetary resources. Annual agency budgets are mainly negotiated between the

Minister of Finance and his staff and the agency management, with the staff of the

portfolio minister not always involved. Also for the periodic budget controls the

Minister of Finance is in direct contact with the agencies. Before the agency reform in

2006, the influence of the Minister of Finance on agencies was very extensive, but even

after the NPM-inspired reforms, influence is still considerable and again increasing due

to budgetary scarcity.

A recent COBRA survey (Verhoest et al. 2013) asked agency chief executives about the

relative importance of the abovementioned steering and control instruments used by

the cabinet and their ministers. The instruments by which they reported to be steered

most significantly were – in order of decreasing importance – control by the budget, by

the policy notes and policy letters of the portfolio minister, by informal contacts with

their portfolio minister, by ad hoc requests outside the performance contact, and

control by the objective-setting and evaluation of the individual chief executives by

government. Note that most of these instruments are quite traditional. For most types of

agencies the result-oriented performance contracts and business plans were considered

as less important steering instruments in terms of their usage by cabinet and ministers.

However, chief executives of public law agencies ranked these result-oriented

contractual instruments second in importance, pointing at a clear difference between

public law agencies with a governing board, and agencies without such a board:

agencies with a governing board apparently experience their minister to adhere more

closely to what was agreed in the performance contracts compared to agencies without

such a board.

When asked in the same survey (Verhoest et al. 2013) about the extent to which the

cabinet and the different ministers influence the strategic orientation of the agency, a

vast majority of the public law agencies with a governing board perceive the portfolio

minister and his staff as well as the cabinet as a whole to have an influence to a high or

very high extent. But also the Minister of Finance and his staff is influence to a large or

very large extent for less than half of these agencies. Yet another minister has a strong

influence according one third of the agencies, namely the Minister of Public Governance

to which we will turn now. Please note that the departments are rarely considered to

14

have a strong influence on the strategic orientation of agencies, reflecting their

contested very weak position in agency governance. 5

As said, there are also some direct contacts between agencies and other ministers and

their services. The Minister of Public Governance who has some role in coordinating

organisational and human resources policies in the Flemish public sector interacts with

agencies and directly requests information from them in order to monitor these issues

at a government-wide level. However, due to a lack of authority this information

exchange proves to be very challenging. Next to that, the Prime Minister and his services

coordinate transversal policy issues (like sustainable development and equality issues)

through government-wide programmes and by formulating generic objectives in

performance contracts. In this role, these actors try to steer and monitor agencies, but

with varying degrees of success. Most coordination across policy domains still happens

at the political level (Verhoest et al. 2013).

A further issue is the parliamentary control over agencies, which parliamentarians and

agency chief executives perceive to be weak. Although the Court of Audit is increasingly

involved in the performance auditing of agencies, and despite the increasing load of

strategic plans and performance information (albeit unstructured and unsystematic),

parliamentarians complain of a lack of information. There is a need for consolidated

financial, budgetary and performance information for the whole government

(departments and agencies), as this information is now given through different

channels. Most of the available information is still on activity level, and information on

outcomes is generally lacking. Moreover, in general parliamentarians are quite unaware

of available control instruments and information, and instead use detailed

parliamentary questions to interrogate the responsible minister on very operational

issues.

The account given above shows that formal instruments are not always used to their full

extent in practice. However, when the tasks or activities of an agency become highly

politically salient, due to a crisis or shifting political priorities, both the kind and focus of

control instruments used by ministers and parliaments are affected. Even more so than

in normal circumstances, the emphasis is then on more traditional instruments like ad

hoc requests, informal contacts by ministers, as well as parliamentary questions (or

inquiries), with a focus on very detailed information on followed procedures or

operational issues. In such times, more NPM-style instruments like result-oriented

steering and monitoring through performance contracts and business plans, or the use

of such information by parliament, is only of secondary importance. For most politicians

these contractual instruments do not allow them enough flexibility to respond to

political crises, urgent needs or demands, or with the policy-oriented information they

need (about policy effectiveness, for example). Hence, most ministers prefer to deal

5 The perception of relative influence of the different principals is quite consistent across the different types of

agencies.

15

with their agencies with more informal instruments, especially in times of high political

saliency.

Arm’s Length Relationships in the Netherlands6

For the case on the Netherlands we focus on a particular type of agency called ZBOs.

Under Dutch administrative law, ZBOs are independent bodies with their own statute

and legal personality. This allows them to use their own financial accounting system

(most prefer an accrual account system or a variation thereof), their own labour

conditions (about half of the ZBOs do not adhere to the civil service regulations) and

organise their own internal management and operating routines. It is important to note

that ZBOs can be either hived off ministerial units that are granted autonomy, or hived

in privated organisations upon which legal authority is bestowed to carry out a public

task. Therefore, ZBOs can be public law but also private law agencies (the latter

constitute about 40% nowadays of the total population).

The first important feature of the Netherlands governmental system is that it is a

Rechtstaat system in which legislation plays an important role. Each ZBO has a legal

basis (law). As of 2007 there is also a Charter Law on ZBOs, which stipulates a number

of requirements regarding the rights of the minister of the parent department to obtain

information, give instructions, and recent reports on the ZBO’s performance. Also, the

Charter Law prescribes that ZBOs should preferably be public law based, use civil

service regulations as the basis for labour conditions, and there are deadlines and other

prescriptions about when and how ZBOs should account for their performance to the

parent ministry. Finally, while the Charter Law does not prohibit the establishment of a

non-executive board (two tier system) it does hinder the establishment thereof by

granting all overseer rights to the minister of the parent department. It is allowed to

diverge from the Charter Law, but that has to be approved by the minister of the Home

Office, who coordinates ZBO-policies, and by parliament that has to approve the

founding law of each ZBO. While each NDPB in the UK is also based in an Act of

Parliamemnt setting out its governance, responsibilities, and so on, this law-based

generic regulation of ZBO governance is quite unlike the position in the UK, with its

reliance on ‘guidance’ issued by the Cabinet Office and the Treasury.

A second feature of the Dutch politico-administrative system concerns its emphasis on

individual ministerial accountability (Andeweg and Irwin 2005). As in the UK, but in

contrast to Flanders/Belgium, the Dutch state is characterised by a high degree of

decentralisation. This does not only concern delegation of competencies to local

governments, but also affects the central government apparatus. Ministries have a large

degree of autonomy when it comes to, for example, personnel selection and

6 Data on the Dutch case originates from different sources: a survey in 2006 among 89 ZBOs, case studies of

four ministries and six ZBOs (twelve interviews and document analysis), prior research by the Dutch author (see references, mostly in Dutch).

16

appointment, internal functioning and organisational structure (Van der Meer and

Raadschelders 1999). Efforts have been made by almost all cabinets in the past two

decades to create more uniformity, for example by installing a Senior Public Service (in

Dutch: Algemene Bestuursdienst), by creating one logo, and by centralising facility

management in shared service centres. However, there are still many differences

between ministries which, as we will see below, also affect how ministries deal with

ZBOs, a feature shared with the UK case.

The emphasis on the individual accountability of ministers has also made the political

debate on ZBOs to revolve almost exclusively around legal issues and particularly the

level of control or ‘grip’ a minister has on ZBOs in his policy domain (Schillemans and

Van Thiel 2009). Technically speaking, the ministerial accountability for ZBOs is limited.

Although nobody knows the exact delimitations, it is generally assumed that the

minister is still or only accountable for: the policy that is implemented by a ZBO, the

decision to charge the implementation to a ZBO, and the supervision of that ZBO.

However, in practice we have seen several incidents where – instigated by media

attention and parliamentary pressure – ministers have intervened in daily affairs or

executive decisions by ZBO management (cf. Pollitt 2005, on poor parenting). Debates

about ministerial ‘grip’ also emerge in the UK in period of political crisis, but unlike the

Netherlands these do not revolve around questions of legal authority; they are much

more about the personal competence of the minister as a political executive.

We now consider the relationship between ministeries and ZBOs. First, we will

describe the preferred model of ministries to manage their relationship with ZBOs (cf.

Van Thiel and Pollitt 2007). This model was originally developed for a different type of

agency, but has become more common now, although not all ministries use it. The

model assumes that a parent ministry has three roles vis-à-vis ZBOs: the role of

sponsor, owner and overseer. The sponsor is the actor that determines the tasks of

ZBOs, i.e. the policy that has to be implemented. The sponsor (principal) draws up the

contract with the ZBO (agent) about performance to be produced and at what price. The

owner is responsible for the continuity of the ZBO and therefore involved in financial

decisions regarding budgets, tariffs, and so on. These roles may therefore have

conflicting interests; while the sponsor wants to pay the lowest price possible, the

owner will be interested in ensuring durability. The overseer has to assess both

whether the agent has achieved the projected objectives (policy effects) and whether

this was done in the most efficient way.

In the application of this model, the roles need to be identified explicitly and attributed

to units within the ministry. This is often laid down in a document (known under

different labels such as steering protocol or ‘vision on regulation’, in Dutch:

toezichtsvisie). Opinions differ on whether the roles should be attributed to different

units or can be attributed to the same unit(s). In practice, ministries deal with this in

different ways (Van Thiel and Hendriks, in review). For example, the Ministry of

Infrastructure and Environment has assigned the sponsorship role to one policy unit

17

that is responsible for the policy that a ZBO implements. The ownership role is charged

to the Secretary General, with administrative support from a unit that is linked to the

financial directorate. Finally, oversight is carried out by different actors: the audit unit

within the ministry, independent regulators like inspectorates, and some ZBOs have a

non-executive board (two tier system). Other ministries like the Ministry of Social

Affairs & Employment have charged the sponsor role to multiple units, while a

specialised unit is set up to maintain and monitor all contacts between ZBOs and the

ministry (a so-called interface unit). This ministry has abolished non-executive boards

and supervises ZBOs through an internal unit and a semi-independent inspectorate.

Other ministries have chosen either not to use the model at all (Justice & Safety) or

charge both the sponsorship and ownership role to the same policy unit, in combination

with a specialised interface unit in a different directorate (Health, Welfare & Sports).

Such discretion is not available to ministers in the UK; all NDPBs are expected to have a

board to exercise strategic decision-making, oversight and accountability on behalf of

the minister.

Ministries do not only differ in the way they organise their relationship with ZBOs, but

also in the instruments they use. For example, some ministries (Social Affairs) have a

much more ‘intense’ relationship with frequent meetings and high demands for detailed

information to be provided by the ZBO – also on matters regarding daily operations.

Other ministries keep much more distance, like the Ministry of Infrastructure which has

retained the non-executive board in most of their ZBOs contrary to cabinet policy lines.

We have also found differences regarding the development of performance indicators,

ranging from a ministry (Justice) imposing a set of indicators top down, to ZBOs making

proposals for indicators themselves (Infrastucture), sometimes even in the absence of

interest in performance agreements by the ministry (Health).

Figure 3 summarises the principal-agent relationships of ZBOs according to this model,

but the reader should be aware that there are large differences between ministries and

sometimes also between ZBOs within the same policy domain.

18

Figure 3: Arm’s length relationships in the Netherlands

ZBOs have no – or are expected not to have – direct relations with parliament. There

have been recommendations to allow ZBO chief executives to come to parliament to

inform MPs about what their organisation does (see e.g. the results from the

parliamentary inquiry by the Dutch Senate into privatisation and agencification, POC

2012) but these have been rejected by the cabinet because ‘it is ministers’ responsibility

to be held accountable’. This reinforces the singular line of accountability in the Dutch

system, in contrast to the multiple accountabilities in the UK (where chief executives are

called to account before parliamentary committees) and the more partisan and

collective model in Flanders. Therefore, in the Netherlands, the parent ministry is the

primary relationship for ZBOs. Consequently, there are no formal contacts with other

ministries either, unless ZBOs are working for multiple ministries (who are then also

sponsors). Even ministries like the Home Office and the Ministry of Finance that have

coordinating tasks for ZBOs, like the Charter Law and senior appointments (Home

Office) and the State Budget (Finance), also have to go through the parent ministry if

they want to contact ZBOs.

There is a range of instruments used in ministry-ZBO relations. Again, ministries differ

in which instruments they use or how they label them. For example, one of the newer

19

instruments that have been developed in the wake of agencification is the so-called

feasibility test. This is a test of feasibility of new policy proposals, to assess under which

conditions and at what cost they can be implemented, and which effects they might

achieve. In some ministry-ZBO relationships, feasibility tests are a standard means to

test the feasibility of new policies ex ante (e.g. Ministry of Infrastructure, Ministry of

Education) while in other ministries (like Justice) they are only used after the policy has

already been decided upon (ex post). A survey in 2006 (n = 89 ZBOs) showed that about

27% of the ZBOs used this instrument frequently but 45% seldom or never.

A number of instruments mentioned in Figure 4 have already been explained, like the

relationship model which is laid down in the steering protocol. Furthermore, the above

also described that the two coordinating ministries can only exercise control through

the parent department: the Home Office deals with the application of the Charter Law,

and is involved in appointments of top officials through the Senior Civil Service (they

have to co-sign appointments for the highest functions in some agencies). The Ministry

of Finance has issued guidelines for all ministries on providing information for the

annual budget and annual account procedures of the whole central government. Not all

ZBO budgets are included in the state budget, as mentioned before, but when they are,

information is provided by ZBOs via the parent ministry.

Sponsor relationships, whether with the parent ministry or otherwise, make use of

instruments like contracts to formalise agreements on output and associated costs. The

reader should note that these are not as legally binding as contracts in the private

sector, therefore they are known under ‘softer’ labels like ‘letter of subsidy’, covenant or

management agreement, as in the UK. These documents may also contain performance

indicators on which the ZBO has to report. Sponsors also can determine what agencies

do by means of the policy and associated guidelines. However, in most cases the

agencies have to ‘translate’ policies into operating routines themselves – as they have

the expert knowledge. None of these contractual arrangements is publicized.

The ownership role is instrumentalised with financial instruments mostly, like setting

or approving a proposal to set certain tariffs, approving budgets and annual accounts,

and making agreements on capital reserves (allowing agencies to have a reserve, or

capping reserves, or approving investment decisions). Many ZBOs use long-term budget

plans.

Most ZBOs have both formal and informal contacts with their parent department,

particularly the sponsor(s) and owner, but again in different frequencies depending on

the ‘style’ of the ministry. A survey in 2006 among 89 ZBOs showed that contacts with

sponsors are much more frequent than with the owner (cf. Table 3). Contacts are also

much more frequent at lower levels of both organisations. The frequency of contacts as

well as the frequency of reporting about performance and finances differs between

ministries and between ZBOs.

20

Top level Operational level Formal

contacts Informal contacts

Formal contacts

Informal contacts

Finance Policy Finance Policy Finance Policy Finance Policy Monthly - 10 15 36 15 50 34 63* 3 Months 20 27 31 28 38 21 18 15 Twice a year

24 20 23 15 15 12 21 12

Once a year

41 27 13 4 10 - 3 -

*Once a month or more frequent

Table 3. Frequency of contacts between ZBOs and parent ministries (2006, N=89).

All scores in percentages.

Oversight can be carried out by a multitude of actors: ministerial units, independent

regulators, the Netherlands Court of Audit, but also by the non-executive board (if

present), through peer review, or the legally mandated 5-year evaluations. In the UK,

such evaluations of agencies are conducted on the basis of administrative rather than

legal rules, again reflecting the contrast with the Dutch Rechtstaat system. To start with

the latter: although legally mandated, only 30% of the ZBOs have actually been

evaluated with regular intervals (POC 2012). This instrument is therefore not used as

frequently as it should be. Moreover, most reports were not made publicly available and

have not been used in the political debate (cf. Van Thiel et al. in review). Other reports

from audits or thematic studies by the court of audit (on aspects that concern the public

sector as a whole) have been more frequent, but with varying impact. For example, a

study into commercial activities of agencies (ARK 2012) received some attention from

MPs but did not lead to decisions. However, when a ZBO is already under fire because of

other issues (underperformance, managerial misconduct, governance issues), reports

will be used in the parliamentary debate (Van Thiel et al. in review).

Different overseers will pay attention to different aspects of ZBO performance, for

example financial performance versus policy delivery results. Moreover, different

overseers have different instruments and competencies. For example, independent

regulators will intervene in cases of underperformance (sanctions, capping prices, etc.),

while the Court of Audit will recommend the minister to intervene (fire alarm function),

and the non-executive board can decide to dismiss the executive managers.

Finally, instruments like peer review, governance codes and horizontal accountability to

stakeholders belong to the new instruments that have been developed by agencies

themselves. A group of ZBOs has formed an association to improve accountability for

ZBO performance to the general public and interested stakeholders. To that end, they

21

have formulated a governance code and they use a peer review procedure.7 The results

of these procedures are however hardly recognised and used by the overseers, the

parent ministry and parliament.

In sum, the primary relationship for ZBOs is with the parent ministry. Contacts with

other ministries only exist when the ZBO is carrying out tasks for those ministries as

well. As a side note, we would like to point to the absence of forms of coordination

between the parent ministry and other sponsor ministries. In view of the current

strategy of ZBO mergers, this lack of coordination will soon become problematic. First

indications of such multiple-principal problems have already been uncovered in the

case of the merger of the Food Safety Agency (a different type of agency, see report by

ARK 2013). This merger did not render the expected financial benefits at all.

How ministries shape their primary relationships with ZBOs can be very different. It is

therefore difficult to characterise the relationship as hierarchical or otherwise. Two

recent reports have criticised the state of development of ZBO-ministry relationships

(POC 2012; De Leeuw 2013) but the cabinet does not seem to want to force uniformity

onto the ministries. Because of the large autonomy that ministries have with regard to

this topic, it does not seem likely that more uniformity could be imposed anyway. The

coordinating force of the Home Office – officially in charge of ZBO policy – is too weak as

well.8 Several ministries are in fact allying with their ZBOs against harmonising

tendencies like the Charter Law, for example in the case of retaining a non-executive

board within ZBOs (Van Ammers and Van Thiel 2012). Traditions like the high degree of

decentralisation and the emphasis on individual ministerial accountability in the Dutch

politico-administrative system will probably remain strong determinants of ZBO

policies for a long time.

Arm’s Length Relationships in the UK9

The Westminster system of government has significantly shaped relationships between

government and its agencies in the UK, and there are three particularly salient features

to be drawn out. Firstly, arm’s length relationships in the UK are informed by an

overarching notion of parliamentary oversight and scrutiny of government business.

The adversarial nature of the political system in the UK’s majoritarian and

predominantly bi-partisan legislature has traditionally meant that there is a central role

7 For more information, see the website (www.publiekverantwoorden.nl, in Dutch).

8 The Ministry of Finance may have more success with its plans to harmonize guidelines on financial

management, oversight and reporting for the whole of the public sector, i.e. including ZBOs. Furthermore, they are the driving force behind the cabinet’s agenda to create more shared service centres and to centralize procurement for all public and semi-public organizations. However, these plans are still work in progress, so we cannot draw any conclusions yet. 9 For the UK case, we drew on over one hundred interviews with civil servants, politicians, and agency chairs

and chief executives. We also analysed a range of relevant documents such as publicly available select committee and think tank reports together with documents about arm’s length relationships provided to us by civil servant contacts. Furthermore, we attended a range of workshops and meetings within government, and held a roundtable with chairs of agencies in collaboration with the Public Chairs’ Forum.

22

for Parliament in holding government to account; however, its ability to do this has

come under recent scrutiny as a result of a perceived lack of public confidence in the

work of Parliament. As a result, the ‘Wright report’ (Public Administration Select

Committee (PASC) 2009) set out a series of recommendations aimed at strengthening

this role, working on the principle that Parliament should set its own agenda rather

than it being dominated by that of the government of the day. Adoption of these

recommendations resulted in a more significant role for backbench MPs and a far

greater role for select committees in scrutinising legislation and holding government to

account (PASC 2013).

Secondly, there is a parallel system of accountability to Parliament in respect of agency

performance. Individual ministers are accountable to Parliament for the exercise of

their functions and the actions of agencies for which they are responsible. This is a

deeply embedded principle in the UK governmental culture, reflecting the considerable

individual discretion available to ministers acting on behalf of Parliament. But in

addition agency chairs and chief executives of ENDPBs are also accountable to

Parliament by virtue of the legislative authority under which their agency operates and

the public money they spend. This accountability has routine aspects, for example

annual reports and accounts are prepared for Parliament. It also has episodic aspects,

for example a select committee may decide to investigate the performance of an agency

and call its chair and chief executive to give evidence. As a result, individual ministerial

accountability to Parliament through hierarchical relationships is apparent in the UK, as

it is in the Netherlands, but in contrast the UK also has direct accountability of agency

leadership to parliament.

Thirdly, the UK context is characterised by the involvement of the core executive in

managing arm’s length relationships (cf. Dommett and Flinders 2014), and in particular

the Cabinet Office which is responsible for cross-government initiatives. This role has

been strengthened as part of the 2010-2015 Coalition Government’s Public Bodies

Reform Programme which has reduced the number of NDPBs while also improving

accountability and governance arrangements for those that remain (cf. Dommett et al.

2014). The Cabinet Office, under the political direction of its minister (who remained in

post over the whole period), was responsible for leading this reform, coordinating the

work of the individual ministries in reviewing the agencies for which they were

responsible. The role of the Cabinet Office in controlling agencies has also been

significantly strengthened as a result of the introduction of a range of new instruments,

as will be described.

A detailed overview of arm’s length relationships in the UK is provided in Figure 4. This

illustration demonstrates the variety and complexity of relationships and instruments

present in a single government-agency relationship. The lowest level relationship in the

hierarchy is between the NDPB and its governance board. While the chief executive, as

23

the accounting officer10, is responsible for financial and operational matters and reports

directly to parliament, the agency’s relationship with the governance board is

instrumentalised through financial and performance monitoring to enable strategic

oversight. The chair of the governance board also typically has an informal but direct

relationship with the Secretary of State (the lead minister for the policy area), and in

some cases junior ministers, on matters of policy and strategy.

The NDPB’s relationship with the sponsoring department is instrumentalised through a

framework document, which provides the aims, objectives and functions of the NDPB

and the roles of the minister, sponsoring department and NDPB (Cabinet Office 2011).

NDPBs also report on their performance to the department and have regular formal and

informal contact through a ‘sponsor’ or sponsorship team which is responsible for day-

to-day liaison with the body. Sponsorship relationships between departments and

agencies have often been described as weak and lacking appropriate rigour to facilitate

effective information sharing and oversight, and the Coalition Government has

attempted to strengthen these arrangements by developing more expertise in

sponsorship. However, there is variation in the quality of sponsorship arrangements

between departments, and attempts by the Cabinet Office to develop a standard cross-

government model of sponsorship have failed to elicit support from departmental

sponsorship teams (Flinders and Tonkiss forthcoming).

The department also implements so-called ‘level 1’ spending controls which are part of

a wider spending controls framework implemented by the government in 2010

(Cabinet Office 2014). These ex ante controls set limits on how much money can be

spent prior to seeking further approval, and departments set limits on how much

money agencies can spend without seeking their approval. There are also then level 2

controls, for which Cabinet Office approval must be gained, and level 3 controls which

require Cabinet Office and Treasury approval. This spending controls framework is the

main way in which the relationship between the agency and the Cabinet Office and

Treasury is instrumentalised. The Cabinet Office also has oversight of the triennial

reviews process, which is a review of each NDPB implemented on a three-yearly basis,

and which considers whether a body should continue to exist and if so whether its

governance arrangements are adequate. These reviews are implemented by the

sponsoring department, but the Cabinet Office manages the process and the Minister for

the Cabinet Office must approve the triennial report before it is finalised.

The Secretary of State has responsibility for setting policy objectives and this is typically

operationalised through the sponsoring department, and also through an informal

relationship with the chair of the NDPB. The Secretary of State is also consulted by the

department in the recruitment process for members of the governance board. While

these are not political appointments, as in the Flanders case, the Secretary of State is

10

Agency budgets are delegated from departmental budgets. As such the Permanent Secretary of the relevant department is also indirectly an Accounting Officer for the NDPB, and reports to parliament on how the departmental budget allocation has been spent.

24

able to influence who is selected for the appointment and the Minister for the Cabinet

Office also has a role in this process, taking an active interest in specific appointments as

well as in overarching issues such as increasing diversity in public appointments.

In addition to the accountability of ministers, chairs and accounting officers to

parliament for the performance of the NDPB, agencies also have relationships with

departmental select committees which may undertake inquiries into the work of

particular NDPBs or into specific policies which significantly affect NDPBs, and may

undertake pre-appointment hearings with candidates selected to act as chairs of

agencies. The Public Accounts Committee, another parliamentary committee, also has a

critical role in monitoring the expenditure of NDPBs, and the National Audit Office, as

external auditor, produces the annual reports for the NDPB and reports these to

parliament.

In summary, it is apparent that in the UK, there is a significant role for parliament as a

principal in arm’s length relationships, and that while individual ministers are

accountable to parliament for the success of agencies in meeting policy objectives, in

recent years the role of the core executive in managing the performance of agencies,

particularly in terms of expenditure, has become more critical. The account presented

here is very much a snapshot of ‘normal’ principal-agent relationships in the UK; the

importance assigned to different principals and instruments can vary according to the

salience of the agency. For example, when there is a crisis which involves the agency,

the Secretary of State is a particularly critical actor and typically engages in greater

steering of the agency through contact with the Chair and Chief Executive, as well as

with the departmental sponsorship team. The role of parliament is also critical during

and after a crisis, when the minister and the chair/chief executive of the agency may be

scrutinised and held to account for the crisis. This was evident, for example, in the

aftermath of the 2014 flooding crisis which affected large areas of Southern England.

The Secretary of State for Environment, Food and Rural Affairs together with the

leadership of the Environment Agency were scrutinised via EFRA, the Environment,

Food and Rural Affairs Committee (EFRA 2014). Sponsoring departments also attempt

to pre-empt crises through ‘risk-based’ sponsorship models, under which closer

oversight is devoted to agencies which are failing to perform, are having financial

problems, and/or are of high political salience.

25

Cabinet Office Spending controls

Sponsorship development Triennial reviews oversight

NDPB reform/creation

HM Treasury Public expenditure Spending controls

Major projects

Parliament Legislation creating/reforming

(some bodies) Annual reporting (some bodies)

Budget setting

Department

Framework agreement

Policy guidance

Spending controls

Triennial reviews

Performance monitoring

Sponsorship relationship

EA reform/creation

Public body (depends on type)

Departmental Select Committees

Inquiries

Pre-appointment hearings

Public Accounts Committee

Inquiries

Secretary of State Policy objectives Budget setting Appointments

National Audit Office Auditing and VFM

studies

Board Strategic objectives

Budget oversight

Annual report

Public body chair Overall performance

Chief Executive Accounting Officer

Operational objectives Budget oversight

Permanent Secretary Accounting Officer

Policy advice

Figure 4: Arm’s Length Relationships

in the UK

Minister Public bodies reform

Regulators (some bodies) Financial performance

Policy delivery Outcomes

26

Understanding principal-agency relationships in comparative perspective

The value of adopting a cross-national comparison when considering the multiple

principle problem is that it exposes the similarities and variations in the way the

relationships between the institutions of government and arm’s length agencies are

instrumentalised. Our argument is that while there are broad similarities –

performance contracts, accountability to ministers and through them to Parliament,

ministerial involvement in senior management/board appointments – there are also

differences both in the way these mechanisms operate and also in the nature of the

mechanisms themselves. In fact, as Smullen (2010) has demonstrated, it is essential to

interrogate the meanings that imbue the language of public management. ‘Performance

contracts’, ‘ministerial accountability’ and so on take on different meanings in different

contexts; one cannot assume a correspondence across political and governmental

cultures, even ignoring the linguistic translation necessary where such terms are not

anglicised.

While the complex principal is evident in each of our cases, there are differences in the

models due to political culture. Flanders has a more legalistic approach, greater

involvement of the cabinet in decisions on agencies, and a high level of partisanship

expressed through parallel control structures from cabinets de ministre and

government commissioners in the board. In addition, agencies are brought into the

partisan environment by virtue of the political influence over appointments, the

absence of a clear role for departments in managing agencies, and the potential for

performance contracts to be debated in Parliament. In this environment, micro-control

by ministers is observable, in a way that is not necessary in the UK context. The Dutch

decentralized approach leads to variety; the desire for more uniformity by some is

stifled by the pluralism of others. Thus despite having a Rechtsstaat system of public

administration, and a Charter Law that regulates ministries and ZBOs, there is

considerable variety in the way the relationships are managed and how such agencies

are organised and governed. What is noticeable, however, is that in comparison with

Flanders and the UK the accountability of Dutch agencies is more strongly to the

minister; the minister is accountable to Parliament and proposals to introduce parallel

accountability for agency heads – as in the UK - have been rejected. In the UK the

combination of high levels of individual ministerial authority and a unified civil service

(the Whitehall model) allows a hierarchical approach to ministry-agency relationships,

as exemplified in the enhanced coordinating and directing role of the Cabinet Office.

This however is within a context in which Parliament is reasserting its authority over

the executive, there is considerable delegation to NDPB boards (recruitment to which

has an element of non-partisanship), and there are parallel accountabilities to

Parliament of minister and agency. Further, the Westminster model is one in which the

law sets a context for action, rather than regulating it as in Flanders and the

Netherlands; instead, the over-riding principle is one of pragmatism.

27

Our analysis shows that the principal is more complex than the literature suggests.

There is complexity in the actors who are institutionalised into the relationship with

agencies – ministers, ministries, cabinet de minister, parliaments, audit offices,

departments at the centre of government, parliamentary committees, cabinets, and so

on. But these are broadly speaking features of all three country cases. There is

complexity in the instruments through which these relationships take place – policy

statements, performance contracts, protocols, budgets, annual reports, Parliamentary

debates, ministerial reviews, and so on. Again, these are found in all three country

cases. But perhaps most importantly there is complexity in the interaction of these

within the prevailing political and governmental culture. For example, in the

Netherlands and UK most relationships between the agencies, the Treasury and the

Cabinet Office run through the sponsoring ministry, indicating that these actors do not

have direct contacts with agencies themselves. This may man that sponsoring

departments are able to streamline potentially conflicting control signals from these

actors, and as such could reduce the multiple principal problem to a great extent. In

Flanders, all these actors have separate and predominantly direct relations with the

agencies, rendering the multiple principal problem more prominent.

As our research develops it will be important to establish the extent to which the

interaction between actors and instruments is affected by crises and other moments

where agencies become highly salient politically. It may be that our initial analysis

shows what happens in normal political times. In other words, we are articulating of

the basic regularities in the system of agency governance. But instrumentalisation is

affected by political salience as well as political culture – some instruments become

more significant than others at certain points in time. Thus in exceptional political

times the interactions between actors and instruments may change; through a study of

critical incidents we may better be able to define and model the interactions within a

complex principal and between its elements and an agency.

Finally we need to consider the motivation for and effectiveness of attempts at reform.

Reforms have been proposed and/or implemented in relation to agencies in all three

countries, for example: the feasibility tests and peer review in the Netherlands, changes

to the system of departmental sponsorship and the role of the Cabinet Office in the UK,

and attempts to reduce the role of cabinet de minister and redefine the relationship

between departments and agencies in Flanders. Reforms typically alter the institutional

arrangements – the rules of the game that define which actors can play, in what roles,

and how the relationships between them should be expressed and regulated. Reforms

are like a moment of crisis in the sense that they throw the system into relief and expose

something of the underlying power relationships, and thus will enhance the study of

multiple principals.

28

References

Andeweg, R.B. & G.A. Irwin (2005). Governance and politics of the Netherlands.

Basingstoke, Palgrave MacMillan.

ARK: Algemene Rekenkamer. (2012). Publieke organisaties en private activiteiten.

[report 20 June 2012]. Download: www.rekenkamer.nl

ARK: Algemene Rekenkamer. (2013). Toezicht bij de Nederlandse Voedsel en

Warenautoriteit na de fusie. [report 20 November 2013]. Download:

www.rekenkamer.nl

Bertelli, A.M. (2006a) The role of political ideology in the structural design of new

governance agencies. Public Administration Review. 66 (4), 583-595.

Bertelli, A.M. (2006b) Delegating to the quango: ex ante and ex post ministerial

constraints. Governance. 19 (2), 229-249.

Bhatta, G. (2003) Post-NPM themes on public sector governance. State Services

Commission Working Paper 17 [online] https://www.ssc.govt.nz/sites/all/files/Post-

NPM_Themes_In_Public_Sector_Governance.pdf.

Blondel, J. (2005) The central role of comparative politics in political analysis.

Scandinavian Political Studies 28(2): 183-191.

Bruce, A., Buck, T., and Main, B.G.M. (2005) Top executive renumeration: a view from

Europe. Journal of Management Studies. 42 (7), 1494-1506.

Caers, R., Du Bois, C., Jegers, M., De Gieter, S., Schepers, C., and Pepermans, R. (2006)

Principal-agent relationships on the stewardship-agency axis. Non-Profit Management

and Leadership. 17 (1), 25-47.

Cabinet Office (2011) Guidance on Reviews of Non-Departmental Public Bodies [online]

http://www.civilservice.gov.uk/wp-content/uploads/2011/09/triennial-reviews-

guidance-2011_tcm6-38900.pdf

Cabinet Office (2014) Cabinet Office Controls Guidance Version 3.2 [online]

https://www.gov.uk/government/publications/cabinet-office-controls/cabinet-office-

controls-guidance-version-32

Child, J., and Rodrigues, S.B. (2003) Corporate governance and new organisational

forms: issues of double and multiple agency. Journal of Management and Governance. 7,

337-360.

Davila, A., and Elvira, M.M. (2012) ‘Understanding organisations in complex, emergent

and uncertain environments: an introduction’, in Davila, A., Elvira, M.M., Ramirez, J., and

Zapata-Cantu, L. (eds.) Understanding Organisations in Complex, Emergent and Uncertain

Environments. Basingstoke: Palgrave Macmillan.

29

De Leeuw. (2013). Onderzoek naar de herpositionering van zbo’s. [rapport]. Den Haag.

Dixit, A. (2002) Incentives and organisations in the public sector: an interpretive

review. The Journal of Human Resources. 37 (4), 696-727.

Dommett, K. and Flinders, M. (forthcoming 2014) The centre strikes back: meta-

governance, delegation and the core executive in the United Kingdom, 2010-2014.

Public Administration.

Dommett, K., Flinders, M., Skelcher, C., and Tonkiss, K. (2014) Did they read before

burning? The coalition and quangos. Political Quarterly.

Eisenhardt, K.M. (1989) Agency theory: an assessment and review. The Academy of

Management Review. 14 (1), 57-74.

Environment, Food and Rural Affairs Committee (2014) Winter Floods, 2013-2014.

London: Stationery Office.

Flinders, M. and Tonkiss, K. (forthcoming) From ‘poor parenting’ to micro-management:

coalition governance and the sponsorship of arm’s length bodies in the United Kingdom,

2010-2013. Under review at International Review of Administrative Sciences.

Holmstrom, B., and Milgrom, P. (1991) Multitask principal-agent analyses: incentive

contracts, asset ownership, and job design. Journal of Law, Economics and Organisation.

7 (sp.), 24-52.

Ingley, C., Rennie, M., Mueller, J., Warrick, D.D., and Erakovic, L. (2012) ‘The social

dimension of boards’ role in better corporate governance’, in Davila, A., Elvira, M.M.,

Ramirez, J., and Zapata-Cantu, L. (eds.) Understanding Organisations in Complex,

Emergent and Uncertain Environments. Basingstoke: Palgrave Macmillan.

James, O. & S. van Thiel. (2011). Structural devolution and agencification. In: T.

Christensen & P. Laegreid. (2010). Ashgate Research Companion to New Public

Management. Aldershot, Ashgate Publishing Ltd. [pp. 209-222]

Kiser, E. (1999) Comparing varieties of agency theory in economics, political science

and sociology: an illustration from state policy implementation. Sociological Theory. 17

(2), 146-170.

Koppell, J.G.S. (2003) The Politics of Quasi-Government: Hybrid Organisations and the

Dynamics of Bureaucratic Control. Cambridge: Cambridge University Press.

Miller, G. J. (2005) The political evolution of principal-agent models. Annual Review of

Political Science. 8, 203-225.

Moe, T.M. (1984) The new economics of organisation. American Journal of Political

Science. 28 (4), 739-777.

30

Moe, T.M. (1987) An assessment of the positive theory of ‘congressional dominance’.

Legislative Studies Quarterly. 12 (4), 475-520.

National Audit Office (2014) Progress on Public Bodies Reform. London: National Audit

Office.

POC: Parlementaire Onderzoekscommissie. (2012). Verbinding verbroken? Onderzoek

naar de parlementaire besluitvorming over de privatisering en verzelfstandiging van

overheidsdiensten. Eerste Kamer, vergaderjaar 2012-2013, B & C.

Pollitt, C. (2005). Ministries and agencies: steering, meddling, neglect and dependency.

In: M. Painter & J. Pierre. (eds.). Challenges to state capacity: global trends and

comparative perspectives. Basingstoke, Palgrave MacMillan. [pp.112-136].

Pollitt, C., Bathgate, K., Caulfield, J., Smullen A., and Talbot, C. (2001) Agency fever?

Analysis of an international policy fashion. Journal of Comparative Policy Analysis:

Research and Practice. 3 (3), 271-290.

Public Administration Select Committee (2009) Rebuilding the House. London: The

Stationery Office.

Public Administration Select Committee (2013) Revisiting Rebuilding the House: The

Impact of the Wright Reforms. London: The Stationery Office.

Shapiro, S.P. (2005) Agency theory. Annual Review of Sociology. 31, 263-284.

Skelcher, C. (2010) Fishing in muddy waters: principals, agents and democratic

governance in Europe. Journal of Public Administration Research and Theory. 20 (1),

161-175.

Smullen, A. (2010) ‘Translating agency reform through durable rhetorical styles: comparing official agency talk across consensus and adversarial contexts’ Public Administration

Tonkiss, K., and Noonan, A. (2013) Arm’s length bodies and alternative models of

service delivery. Public Money and Management. 33 (6), 395-397.

Van Ammers, M. & S. van Thiel. (2012). Kaderwet Zelfstandige Bestuursorganen: het

einde van Raden van Toezicht als toezichthouders bij ZBO’s? Bestuurswetenschappen,

66(4), 35-54.

Van der Meer, F.M. & J.C.N. Raadschelders. (1999). The senior civil service in the

Netherlands: a quest for unity. In: E.C. Page & V. Wright. (1999). Bureaucratic Elites in

Western European States . Oxford, Oxford University Press. [pp. 205-228]

Van Thiel, S. (2001). Quangos: trends, causes and consequences. Aldershot, Ashgate

Publishing Ltd.

31

Van Thiel, S. (2012). Comparing agencies across countries. In: K. Verhoest, S. Van Thiel,

G. Bouckaert, & P. Laegreid. (Eds.). Government Agencies: Practices and Lessons from 30

countries. Basingstoke, Palgrave MacMillan. [pp. 18-26].

Van Thiel, S. & C.Pollitt. (2007). The management and control of executive agencies: an

Anglo-Dutch comparison. In: C. Pollitt, S. van Thiel & V. Homburg. (eds.). (2007). New

Public Management In Europe: Adaptation And Alternatives. Basingstoke, Palgrave

MacMillan. [pp. 52-70]

Van Thiel, S. & R. Hendriks. (2014, in review). Aansturing van zelfstandige

bestuursorganen door ministeries: stijlverschillen.

Van Thiel, S. & J. Verheij. (2014, in review). ZBO’s in getallen.

Van Thiel, S. & K. Yesilkagit. (2014). Does task matter? The effect of task on the

establishment, autonomy and functioning of semi-autonomous agencies. International

Review of Administrative Sciences, 80(2), 319-341

Van Thiel, S., S. Soetekouw, M. Dresmé & P. van Goch. (2014, in review). ZBO evaluaties:

ter leringh ende vermaeck.

Verhoest, K., Demuzere, S., and Rommel, J. (2011) ‘Belgium’, in Verhoest, K., Van Thiel, S.,

Bouckaert, G., and Laegrid, P. (eds.) Government Agencies: Practices and Lessons from 30

Countries. Basingstoke: Palgrave Macmillan.

Waterman, R.W., and Meier, K.J. (1998) Principal-agent models: an expansion? Journal of

Public Administration Research and Theory. 8 (2), 173-202.

Yesilkagit, K. & S. van Thiel. (2012). The Netherlands. In: K. Verhoest, S. Van Thiel, G.

Bouckaert, & P. Laegreid. (Eds.). Government Agencies: Practices and Lessons from 30

countries. Basingstoke, Palgrave MacMillan. [pp. 179-190].