ACCOUNTABILITY AFTER STRUCTURAL DISAGGREGATION: COMPARING AGENCY ACCOUNTABILITY ARRANGEMENTS
| Date | 01 December 2015 |
| Published date | 01 December 2015 |
| DOI | http://doi.org/10.1111/padm.12185 |
| Author | SANDRA VAN THIEL,MARIEKE VAN GENUGTEN,SJORS OVERMAN |
doi : 10. 1111/p adm .12185
ACCOUNTABILITY AFTER STRUCTURAL
DISAGGREGATION: COMPARING AGENCY
ACCOUNTABILITY ARRANGEMENTS
SJORS OVERMAN, MARIEKE VAN GENUGTEN AND SANDRA VAN THIEL
New accountability instruments – performance indicators, audits, and nancial incentives – are
expected to replace traditional accountability instruments in NPM reforms. Wetest this expectation
by looking at the accountability arrangements of semi-autonomous agencies as a typical example of
NPM reforms. Our ndings are based on survey data on 342 agencies in six countries. We identify
four types of accountability arrangements in semi-autonomous agencies, in line with the four
trajectories of public management reform identied by Pollitt and Bouckaert (2004). Accountabil-
ity relations between agencies and ministries are determined by country-specic administrative
regimes, types of agency, and to a lesser extent, agencies’ policy elds. We identify new avenues
for theory and research into the effects of reforms on accountability in the public sector, and on
semi-autonomous agencies in particular.
INTRODUCTION
With the advent of New Public Management (NPM), many new mechanisms of account-
ability were introduced in the public sector, including explicit standards, performance
measures, and output controls (Hood 1991; Power 1999; Pollitt and Bouckaert 2011). This
has led to a shift from accountability for processes to accountability for results (Bardach
and Lesser 1996; Behn 1998). The introduction of such new accountability mechanisms is
considered important because in many cases NPM reforms like structural disaggregation
and contracting out fundamentally alter the democratic legitimacy of the public sector
(Behn 1998), as they substitute traditional accountability mechanisms via political princi-
pals and democratically elected members of parliament (Maggetti 2010).
In this study, we analyse the shape of accountability arrangements in the context of a
specic reform, namely agencication. Agencication refers to the delegation of public
tasks to so-called semi-autonomous agencies (Van Thiel 2001; Verhoest et al. 2010). These
agencies carry out public tasks, such as policy implementation, public service delivery,and
regulation, but they are structurally disaggregated from their political principals. Politi-
cians and NPM champions justify agencication by claiming that agencies will bring polit-
ical, economic, and organizational benets (Pollitt et al. 2001; Smullen 2010; Verhoest et al.
2012). However, these gains may come at the cost of agency loss for political principals, as
well as problems for political and democratic legitimacy, because agencies are non-elected
bodies, operate at arm’s length, and are not under direct, hierarchical ministerial control.
To deal with the potential problems that arise from the structural disaggregation of
agencies, NPM prescribes a shift from ex ante control to ex post control and accountability,
based on contracts, audits, and nancial incentives (Behn 1998; Pollitt et al. 2001). In this
line of reasoning, agencies are made responsible for their performance (Gilardi 2008) and,
therefore, can be held accountable for their performance (Majone 1999; Busuioc 2009).
Sjors Overman, Marieke van Genugten and Sandra van Thiel are at the Institute for Management Research, Radboud
University,Nijmegen, the Netherlands.
Public Administration Vol.93, No. 4, 2015 (1102–1120)
© 2015 John Wiley & Sons Ltd.
ACCOUNTABILITY AFTER STRUCTURAL DISAGGREGATION 1103
However, administrative reforms, including agencication and the implementation of
new accountability arrangements, are seldom part of comprehensive reform packages
(Pollitt 2001; Andrews 2010). Instead, politicians and top civil servants design their
reforms using various ideas and best practices elsewhere. Indeed, these reforms are often
subject to compromises, adapted to t within existing political-administrative institutions,
and therefore, are partly implemented or only paid lip service.
In this study, we test which parts of the new ex post accountability mechanisms are put
in place, and whether these accountability arrangements differ between agencies, operat-
ing in several policy areas, at various degrees of disaggregation from political principals,
in six countries. Using data from 342 organizations, we are able to identify four types
of accountability arrangements in semi-autonomous agencies, in line with the four pub-
lic management reform trajectories labelled by Pollitt and Bouckaert (2004): maintaining,
modernizing, marketizing, and minimizing. The occurrence of these types of accountabil-
ity arrangements is correlated with country origin, type of agency, and policy sector.
There is some literature on the accountability of semi-autonomous agencies (Curtin
2005; Verschuere et al. 2006; Schillemans 2008; Busuioc 2009; Reiss 2009; Koop 2014), but
international comparative research on this topic is still scarce. This article contributes to
this lacuna by using data on a substantial number of agency accountability arrangements.
Moreover, we suggest that the reform trajectories labelled by Pollitt and Bouckaert (2004)
are well suited to classify individual organizations, rather than entire country reform tra-
jectories. In the next sections we will present a number of expectations based on the litera-
ture on public management reform, accountability, and agencies. These expectations will
be analysed, using survey data in a cluster analysis and a multinomial regression. Based
on our ndings, we will discuss important implications for theory and practice.
ACCOUNTABILITY MECHANISMS
The literature on accountability has a vast scope and attributes a multitude of different
meanings to the main concept (Bardach and Lesser 1996; Mulgan 2000; Dunn and Legge
2001; Dubnick 2005; Bovens 2010; Yang 2012; Tenbensel et al. 2013). Bovens (2010) identi-
es two distinct forms of accountability: accountability as a virtue and as a mechanism.
Accountability as a virtue leads to legitimacy of an actor. Accountability as a mechanism
can lead to, for example, popular control, prevention and detection of corruption, and
reection or learning (Bovens 2010). Here, we will apply the perspective of accountability
as a mechanism. Following this perspective, an accountability relationship is character-
ized by an agent and a forum, where the agent has an obligation to provide information
and clarication about its conduct, and the forum can ask questions, pass a judgment, and
make the agent face the consequences of its behaviour (Bovens 2007, pp. 450–52).
The process of accountability can be divided into three separate phases: information,
discussion, and consequences (Bovens 2005; Brandsma and Schillemans 2012). In an
agency-specic context, in the information phase the agency provides the forum (i.e. its
parent ministry) with an account of its behaviour and performance, leading to a high
frequency of formal meetings between representatives from the agency and the parent
ministry.In the countries under study, agencies typically report to the ministry,not to par-
liament (cf. Verhoest et al. 2012). Horizontally arranged forums, such as public reporting
and citizen panels (Schillemans 2008), are not included in this study.The discussion phase
follows, and encompasses the process that eventually leads to a judgment by the forum
about the behaviour and performance of the agency. Audits are tangible instruments
Public Administration Vol.93, No. 4, 2015 (1102–1120)
© 2015 John Wiley& Sons Ltd.
To continue reading
Request your trial