Académique Documents
Professionnel Documents
Culture Documents
http://ppa.sagepub.com
Published by:
http://www.sagepublications.com
On behalf of:
Additional services and information for Public Policy and Administration can be found at:
Email Alerts: http://ppa.sagepub.com/cgi/alerts
Subscriptions: http://ppa.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.co.uk/journalsPermissions.nav
Citations http://ppa.sagepub.com/cgi/content/refs/24/1/23
Abstract
This article focuses on the role public agencies play in the policy making
process. Based on a single embedded case-study of one Flemish public
agency and two policy-programmes in which this agency is involved, the
normative practitioner model of the policy/operations divide is tested. This
model assumes (or prescribes) that policy making is a political prerogative,
while public agencies as administrative actors should stick to implementing
policy. The evidence shows that reality is more complex than the rhetoric of
the practitioner model. Agencies may be more involved in policy preparation
and policy decision-making than assumed, and political actors may in some
cases have a large say in policy implementation. The evidence from my
case-study shows that in reality the policy/operations divide might not be that
clear-cut, and that this normative model should be revised on some points.
Next to that, the findings may be discussed in the light of the current
administrative reform projects in various countries that propagate a strict
labour division between the political sphere (policy) and the administrative
sphere (implementing policy). This discussion is particularly relevant for the
Flemish public sector that is currently facing a major reform of that kind.
Keywords
Introduction
The focus of this article is the task division between political actors and administrative actors in the policy cycle. As I will show, the normative practitioner doctrine of the policy/operations divide assumes that politicians decide on policies,
while administrative agencies should implement these policies. This philosophy is
DOI: 10.1177/0952076708097907
Bram Verschuere, Dept. Handelswetenschappen en Bestuurskunde, Hogeschool Gent,
Voskenslaan 270, 9000 Gent, Belgium. [email: bram.verschuere@hogent.be]
23
also one of the cornerstones of the current administrative reform that is transforming the structure of the public sector in Flanders (BBB: Beter Bestuurlijk Beleid,
or Better Governmental Policy [author translation]). The aim of this article is
therefore to investigate empirically the extent to which this model is also found in
reality. Many authors have questioned the empirical validity of a strict task division between politics and administration (e.g. Aberbach et al., 1981; Jacobsen,
2006; Pollitt et al., 2004; Svara, 1998, 2001). In this article, I will first discuss the
rhetoric of the model and how this model has been an inspiration for the ongoing
Flemish public sector reform. Second, I will empirically test the prevalence of the
policy/operations divide in the real world, via a single embedded case study of
one Flemish public agency. Third, I will discuss the research findings and their
relevance for the theoretical and practitioner debates.
24
the agency form of organizing government activities has become the subject of reform
programs in dozens of countries in the past decade and a half . . . these initiatives can be
grouped under diverse labels, from reforming pre-existing agencies (in countries such as
the USA, Sweden, and Finland), through creating new and only moderately autonomous
agencies (in the UK and the Netherlands), to a fairly radical separating out of a class of
organizations with extensive autonomy (New Zealand e.g.). (Pollitt et al., 2004: 31)
DET
IMP
Principal
Agency
Figure 1 Hypothetical task division in the policy cycle between arms length
agencies and their oversight authorities in a simplified manner
the policy/decision-making process. Figure 1 shows this hypothetical task division in the policy cycle between arms length agencies (agents, full line) and their
oversight authorities (principals, dotted line) in a simplified manner.
On the horizontal axis, the stages in the policy cycle are represented in a simplified manner. The vertical axis shows the influence of agents and principals in
policy decisions made in the different stages of the policy cycle. The preparation of
policies (PREP) and the determination of policies (DET) are tasks that are the
responsibilty of the oversight authorities of the agencies (administrative and political). The key task of arms length agencies is implementing these policies (IMP).
The figure thus represents in a simple way the so-called policy/operations divide,
which suggests that there is a gap between administrative tasks (implementing
policies) and political tasks (preparing and deciding on policies): agencies are
primarily involved in decisions made during the implementation stages of the
policy, while principals make most significant decisions during the preparation and
decision making stages of the policy. If these assumptions by the normative practitioner theories are correct, then one can expect that the involvement of agencies in
the policy making process is low in the preparation and determination stages of the
policy cycle. In these phases, the important policy decisions will be made by the
political and administrative principals of the agency. In the implementation phase,
on the other hand, one can expect that agencies will have higher levels of involve26
ment (resulting from the agencys larger role in the implementation of the policy).
However, some authors do not consider the policy/operations divide as a key
feature of the NPM ideal type agency. They observe a discrepancy between official rhetoric and empirical reality. For example Pollitt et al. (2004: 41ff.) do not
include the policy/operations divide in their normative tripod-model of agencification. Although the same authors declare that in many accounts of the agency
movement it is suggested that this includes a split between policy-making and
policy implementation, they doubt this feature is fundamentally part of the
agency idea as there are both doctrinal and empirical examples of where the
policyoperations divide does not form a part of the agency movement. Pollitt et
al. (2004: 412) give an overview of agency-practices in which a strict
policy/operations divide seems absent. In Sweden for example agencies and ministers both play important roles in policy making historically. Carpenter (2001)
found that many agencies in the USA seem to have a great deal of policy making
competences (policy autonomy). Talbot (2004) observed that in the UK a considerable number of large agencies not only provide services, but also are involved in
policy. As a final example, some authors claim that service provision itself (an
assumed task of agencies) includes not only services or products, but also may
include policy advice (Boston et al., 1996; Kemp, 1993).
In the remainder of this article I take a look at the Flemish situation. First I
describe the rhetorics at work by looking at how the policy/operations divide is
normatively conceived in the public sector reform in Flanders. Next, I take a look
at the empirical reality, in order to check whether agency practice matches with
normative rhetorics.
According to BBB, the minister and the government are vested with extensive
policy determination competences that range from setting the goals to choosing
the policy-instruments. The administrative actors role in the policy decisionmaking process is restricted to advice and implementation. Ministries advise on
policy, arms length agencies implement these policies.
Empirical Scope
For answering my research question (to what extent do we observe the model of
the policy/operations divide in reality in Flanders?), I have performed a single
embedded case study. I have selected one executive Flemish public agency, and
two policy programmes in which this agency is involved. The case organization
(the agency) is the Flemish Public Transport Company (hereafter VVM). VVM is
an agency with an implementing role. It has not many regulative or other policyrelated tasks. Its main goal is to provide public transport, hence driving public
buses and tramways. This is a crucial organizational feature on which I rely for
selecting this agency, as I need to be able to test the assumption that agencies with
a large implementation role are not involved in policy related matters. Hence, I
choose for an agency that has a large role in policy implementation (or service
delivery). The VVM is organized via the principle of territorial decentralisation.
Beside the central management and the staff functions in the central headquarters,
it has five provincial entities that are responsible for delivering public transport in
the province for which the entity is responsible (Flanders has 5 provinces). Next to
that, it is a so-called arms length agency, structurally disaggregated from the core
28
department, with its own public law legal personality. This is also an important
criterion for choosing the VVM as my case organisation, as I want to test the task
division in the policy between politicians and arms length agencies (which are
assumed to be policy implementing vehicles according to normative assumptions). Moreover, the VVM seems to be more or less representative for the group
of arms length public agencies in Flanders in terms of observed autonomy and
control practice by the oversight authorities (Verschuere, 2006). Another important point is that the organization is not too much affected by changes resulting
from the reform process that is currently taking place in Flanders (BBB). It is
important that the managerial and organizational structure of my case organization
remains more or less the same before and after BBB, because the data-collection
took place in the middle of the reform process. Additionally, the organization must
be accessible for collecting data. Practical reasons urge for choosing an organization that is accessible and has sufficient critical mass to enable participation in a
study. A final important criterion is that the organization I have selected allows for
a valid data collection and analysis strategy. This means that it needs to be
involved in policy initiatives that are accessible and comparable, and that contain
policy implementation initiatives that are also accessible and traceable. For more
features of the case organization I refer to Appendix 1 (derived from Verschuere
2006).
After a first round of discussions with people closely related to the selected
agency I have chosen two policy initiatives in which the VVM is involved: basic
mobility and Pegasus. The policy programme of basic mobility aims at increasing the supply of public transport in the Flemish region. Via a decree that was
voted in parliament in 2001, the norms for public transport delivery in Flanders
were set. Essentially, the decree sees public transport as a social right for everybody. This social right applies to all citizens, in all regions of Flanders, notwithstanding special attention is given to certain groups that are considered risk-groups
and are vulnerable for suffering from traffic-poverty (disabled people, elderly people, people living in rural areas, etc.). Therefore, the decree arranges basic mobility by specifying explicit output norms for the public transport services. Public
transport services are to be delivered within acceptable time-limits and distance
limits, and therefore norms and criteria have been developed (frequency of service, number of bus stops, amplitude of service, etc.).
The policy programme of Pegasus was launched when societal consensus
started to grow that one of the key policy issues for the future was the preservation
of the accessibility of the Flemish Diamond, the economic heart of Flanders
(region between Antwerp, Ghent, Brussels and Leuven). One of the measures
taken to fight ever growing traffic jams and traffic congestion, was the continuous
development of a high quality public transport net. The choices are made explicit
and determined in the Pegasus-plan, an overall plan for the Flemish Diamond, that
was taken up in the governmental agreement of the current Flemish government.
In this plan the policy goals and the desired policy output have been made
29
concrete. The plan contains mainly tactical choices about how public transport
service delivery may help to reach the policy goals expressed. The aim is to reach
extra travelers for public transport and to decrease traffic jams in the region. This
goal is operationalized in explicit figures and goals (80m extra travelers for public
transport by 2025), to be achieved by ameliorating the bus and tram net in the
region. One of the first concrete projects under Pegasus was the implementation of
measures that had to decrease the negative mobility effects during the restorationworks of the Antwerp ringroad in 2004. This very busy highway lost a considerable part of its automobility-capacity during these works. A large part of the
measure was to replace lost automobility-capacity by extended public transport
(buses, trams), the inspiration for which was taken from the general Pegasus
policy plan.
There are some important resemblances between both policy programmes. First,
both policy programmes have a relatively high impact and visibility. Basic mobility is the restructuring of the supply of public transport in the whole Flemish region,
and in every single local authority. This policy affects virtually every citizen, and
especially the users of public transport. Pegasus mainly affects the region of the
Flemish Diamond and the people who live there (57% of total population in
Flanders). But it also affects a lot of home-work travellers, as the region hosts many
employers (more than 60% of the workforce is located in the Flemish Diamond). In
both programmes, the VVM is the implementer of the policy, by delivering public
transport services. VVM has to implement basic mobility projects in the field, and
VVM has to implement parts of the policy of Pegasus in the field. Furthermore,
both policy programmes stem from policy needs that are politically and societally
recognized. This can be illustrated by the fact that both programmes address
problems that are related to two of the key-pillars of the Mobility Plan Flanders:
(1) guaranteeing access to mobility for every citizen in Flanders and (2) guaranteeing the accessibility to the Flemish economic centres. As such, basic mobility is
part of the social policy component of public transport, whereas Pegasus is to be
considered as part of the societal component of public transport.
Notwithstanding these similarities, there is one crucial difference between both
policy programmes. The level to which the policy programme of basic mobility is
formalized and detailed is much larger than the level to which the policy programme of Pegasus is formalized. In terms of Huber and Shipan (2002), this
means that the policy of basic mobility is spelled out with a much larger level of
detail compared to the policy of Pegasus. The bottom line is that basic mobility is
arranged via a decree that was voted in Parliament, and via accompanying
governmental decisions. These regulations arrange the way basic mobility is to be
implemented in the field with a large level of detail. On the other hand, the policy
of Pegasus is only formalized to a small extent. Its formal status is a policy plan
that has been taken up in the governmental agreement implying a commitment by
the government to take initiatives that start the implementation of the Pegasusplan. Such a commitment has less direct consequences than a decree, as it is not
30
implying a right that has to be fulfilled by law. Next to that, the way the implementation projects that are under the Pegasus-umbrella are to be implemented is
specified in a less detailed way as is the case for basic mobility.
AG MAN
10.00
AG ENT
PR POL
9.00
PR ADM
8.00
7.00
6.00
5.00
4.00
3.00
PREP
DET
Phase
IMP
questionnaire data are presented here, the additional documentary evidence has
shown to be in support of the data presented (Verschuere, 2006).
AG_MAN
(mean + N)
AG_ENT
(mean + N)
PR_POL
(mean + N)
PR_ADM
(mean + N)
BM_PREP
4,9 (18)
3,2 (17)
9,2 (19)
5,1 (19)
BM_DET
6,1 (20)
3,6 (18)
9,5 (20)
6,7 (19)
BM_IMP
7,5 (20)
8,4 (20)
7,6 (20)
6,7 (18)
Wilcoxon
PREP*DET (.03) PREP*DET (.32) PREP*DET (.25) PREP*DET (.00)
test (paired) PREP*IMP (.00) PREP*IMP (.00) PREP*IMP (.02) PREP*IMP (.02)
+ sig. of
DET*IMP (.01) DET*IMP (.00) DET*IMP (.00) DET*IMP (.65)
differences
actors in the policy process are statistically significant. One can read the table horizontally (mean scores of the actors, per policy phase, and N of respondents). In the
last column the results of the Wilcoxon test for comparing means are shown. This
test looks for statistically significant differences in the means of two related samples, showing for example that in the preparation phase, the mean scores of the
different actors are significantly different, except between the agency management (mean: 4.9) and the administrative principal (mean: 5.1).
We can also read the table vertically (scores of 1 actor over the different policy
phases). Also here, the significance of differences between mean scores has been
tested statistically. The results thus show that the agency seems to be more
involved in the implementation phase than in the preparation and determination
phases, which is consistent with the normative model (compare figure 1 on page 5
33
AG MAN
10.00
AG ENT
PR POL
9.00
PR ADM
8.00
7.00
6.00
5.00
4.00
3.00
PREP
DET
Phase
IMP
with figure 2). The provincial entities of the agency seem to have a large role in the
implementation phase. These findings are also statistically significant. Moreover,
we can see that during the first two policy phases, the political principal seems to
have a significantly larger influence than any other actor in the policy process,
which is also consistent with the normative model. On the other hand, the large
influence of the political and administrative oversight authorities in the implementation phase (no significant differences with other actors) is not in line with
the assumptions of the policy/operations divide model.
A similar analysis can be done for the policy process of Pegasus. Figure 3 shows
that the key decision maker in the preparation phase of the policy is the minister of
mobility and his cabinet.
Actors in the agency and actors from the administrative oversight authorities
were involved to a lesser extent in the preparation phase. During the determination
phase however when the policy vision is translated into a politically approved
policy plan the actors in the agency also start to play a role. The agency management and the agency provincial entities, together with the minister of mobility
(PR POL) seem to play an equal and large role in the decision making process
during the determination of the policy of Pegasus. In the implementation phase
this is also the case: the political oversight actors and the agency simultaneously
play a considerable role in the decision making process. Table 2 is similar to the
table presented above for the policy of basic mobility, and shows to what extent
34
Table 2 Mean scores of the actors per policy phase and over different policy phases
AG_MAN
(mean + N)
AG_ENT
(mean + N)
PR_POL
(mean + N)
PR_ADM
(mean + N)
PEG_PREP
6,5 (17)
5,2 (17)
8,4 (18)
5,6 (16)
PEG_DET
7,9 (18)
8,1 (18)
7,4 (17)
3,5 (15)
PEG_IMP
7,3 (19)
8,4 (20)
8,4 (20)
4,1 (16)
Wilcoxon
PREP*DET (.01) PREP*DET (.00) PREP*DET (.04) PREP*DET (.01)
test (paired) PREP*IMP (.09) PREP*IMP (.00) PREP*IMP (.56) PREP*IMP (.17)
+ sig. of
DET*IMP (.13) DET*IMP (.44) DET*IMP (.02) DET*IMP (.17)
differences
the differences between the mean scores of the actors per policy phase (horizontally), and of the same actor over different policy phases (vertical) are statistically
significant.
Contrary to the policy programme of basic mobility, the findings for the policy
programme of Pegasus seem to contradict the normative model to a large extent
(compare Figure 1 with Figure 3). In the preparation phase of the policy, the
influence of the political principal is significantly higher than any other actors
influence, but in the determination and implementation phases this is not the case.
The influence of the agency in both phases is equal to the political principals
influence (and high). The influence of the administrative is low throughout the
policy cycle.
35
Discussion
The results of my case study shed an interesting light on the policy/operations
divide. It seems that the strict policy/operations divide is absent in the case of the
VVM, and that this agency cannot be simply positioned in one or other position
along the divide. The political oversight authorities, and to a lesser extent the
administrative oversight authorities, have a substantial input during the implementation phase. Meanwhile, and especially for the case of Pegasus, the agency
has a considerable influence in the preparation and determination of the policy.
Based on these observations, we can conclude that there is no clear gap between
policy (task of democratically elected politicians and their staff, political and/or
ministerial) and operations (implementation as the exclusive task of the arms
length agency) as some practitioner theories would suggest (and sometimes advocate). This finding is in line with a growing number of empirical studies (e.g.
Pollitt et al., 2004). The case study shows that the level to which the agency is
involved in policy decisions may vary between different policy programmes, and
between different policy phases. The conclusion that the policy/operations divide
is rare among Flemish agencies, is also reflected in earlier research on Flemish
agencies by Verhoest et al. (2003). This suggests that many agencies perceive
themselves to be able to make significant policy choices. Next to that, and based
on data from the same survey, the majority (approx. 75%) of Flemish public
agencies seem to declare to have policy implementation as a primary task. In the
same survey however, there is also a considerable number of agencies that see
policy preparation as their primary task (20%), or even decision-making (5%).
Furthermore, research by Rommel (2006) of five Flemish public agencies also
concludes that the separation between policy and execution is unrealistic and
normatively disputed, and that agencies may be involved in the preparation and
the formulation of policy simultaneously.
Also, other authors have concluded that as far as the policy/operations divide is
concerned, the reality seems to be more complex than the normative model would
suggest. The dichotomy between politics and administration, which is rooted in
the belief that politicians and administrators should not interfere in each others
domain, has repeatedly been proven false by various scholars who suggest a
model of complementarity between political and administrative actors. According
to Jacobsen (2006) there has been much research (e.g. Aberbach et al., 1981;
Aberbach and Rockman, 1988; Kingdon, 1995; Svara, 1998, 2001) that shows an
extensive, though varying, intermeshing of the two spheres [political and administrative]. This view is best formulated by Svara (1998, 2001), who claims that the
politics-administration dichotomy is an aberration. The dichotomy model
remains important as a normative standard for practitioners, but it is not what it
seems. Based on research, Svara (1998) proposes an alternative model of
complementarity with shifting boundaries between the policy and the administrative arenas, implying changes in the behaviour of both politicians and bureaucrats.
36
Complementarity entails ongoing interaction, reciprocal input, and mutual deference between elected officials and administrators. Administrators help to shape
policy and give it specific content and meaning in the process of implementation.
Elected officials oversee implementation, probe specific complaints about poor
performance and try to fine tune in performance problems (Svara, 2001: 180).
Other authors also come to similar conclusions, based on empirical research in
different contexts. Jacobsen (2006), for example, shows that the border between
the political and administrative sphere is not absolute. His research in 30
Norwegian municipalities (Jacobsen, 2006: 303) shows that this border is a variable opening up for the possibility that it may vary among contexts, structures,
demographics and over time. Fedele et al. (2005) point at a similar phenomenon
based on research on Italian agencies:
the experience in many countries highlights examples of ministers involvement in
agency managerial and operational matters . . . [on the other hand, also] the study of two
agencies in Italy seems to indicate an influential role of the executive agencies in the
policy formulation process, with a potentially political role. (Fedele et al., 2005: 9)
A second point for discussion is an issue for further research. Given the observation that in some cases the agency (as an administrative actor) is involved in policy
making processes, and in other cases not, it might be interesting to investigate the
circumstances under which agencies have a considerable say in the policy process,
or not. Earlier research may provide us with some answers. A first factor that may
be influencing the involvement of administrative actors in the policy process is
capacity. Carpenter (2001) found that organizational capacity and expertise,
resulting from strong leadership, talented officers that are coherent, and offices in
which turnover is minimized (stability, experience), are important factors for
establishing the reputation and legitimacy of the agency upon which it can rely to
forge decision making autonomy. Similar claims about the importance of organizational capacity and expertise were advanced by Hammond and Knott (1999),
who argued that specific leadership and skills of the management may be a
prerequisite to advance the preferred policy of the agency. Krause (2003) argued
that organizational stability is important for the agency to be able to handle discretion, and Hawkins and Jacoby (2004) have argued that agency/expertise and
knowledge may be important resources to enable the agency to have discretion or
autonomy in (implementing) policy. Other factors may be related to the policy
programme, for example the political salience around the policy. One might argue
that in highly significant policy issues the agency will tend to be steered more
strictly (Pollitt et al., 2004; t Hart, 2002) or will have the incentive to demand less
discretion (Krause, 2003), hence having fewer opportunities to be involved in
policy making. On the other hand, one could also reason the other way around. It
could be the case that in times of high political turbulence around policy, the political oversight will tend to pass the hot potato to the agency, and give them the
discretion to taken policy related decision (Verschuere, 2006). Another factor may
37
be the attitude of the agency towards the programme. This factor is likely to be
related to cultural-institutional theories, which would explain decision making
behavior of agencies as being largely culturally determined and based on internalized normative and value frameworks (Laegreid et al., 2005, 2006; Thoenig,
2003). The attitude of the agency towards the policy programme can be induced by
a normative belief of the agency about whether the proposed policy seems to
deal with some urgent societal problems or needs, and whether the issues with
which the proposed policy deals are the right issues to be addressed as a policy
priority (according to the agencys perception). Still other factors may be internal
to the actors in the policy process, such as the rather normative role perception of
the agency. The role perception of agencies may be defined as the belief of how
actors are supposed to behave (or what behavior is appropriate). The assessment
of appropriateness will be done by the actor via comparing actual behavior to the
normative value frameworks, roles, and routines that apply and are internal to
the organization. Role perceptions of the agency may determine whether it is
appropriate to be involved in the decision making process (are we an agency with
a policy role?). Gains (1999, 2004) showed that customs and traditions derived
from the past such as existing formal and informal constitutional and institutional
arrangements in governmental departments to a large extent determine the
features of the minister-agency relations that are at work today in Next Steps agencies (via mechanisms of path dependence). Moynihan and Pandey (2006) found
that features of organizational cultures may help to explain the level of (managerial) autonomy in public agencies. According to the latter authors, agencies with a
strong managerial and entrepreneurial culture tend to demand and receive more
managerial autonomy (contrary to agencies with a bureaucratic culture). Finally,
the relationship between political principal and administrative actor may influence
the extent to which administrative actors are involved in the policy making
process. According to Laegreid et al. (2005, 2006), high levels of mutual trust and
support between the agency and the oversight authorities may be assumed to
decrease levels of cultural collusion between agency and minister, increasing the
levels of leeway that is left to agency, because the agency is trusted to make the
right decisions. Also, other authors found that large political support for the
agency may lead to greater autonomy of the agency. Moynihan and Pandey (2006)
claim that politicians are likely to provide autonomy to an organization they trust,
has a strong track record and is unlikely to use autonomy in an undesirable way.
Carpenter (2001) found that agencies with a good reputation, a strong legitimacy
and embedded in strong political networks may be able to forge their policy
autonomy. Verhoest (2003) found that agencies showing an open and consistent
attitude towards the political principals may enjoy larger levels of autonomy (or
lower oversight control).
The empirical evidence is too narrow for being able to test the existing explanations for policy autonomy or policy influence of administrative agencies. This
discussion is beyond the scope of the article. At most, the article can feedback
38
single case study results to this literature, and develop a refined hypothetical
model based on this evidence (see Verschuere, 2006). It might be worthwile however in future research to test existing explanations based on large-N or multiple
case study research. In relation to the case study, some initial statements about the
reasons for agencies involvement in the policy decision-making process can be
made. The fact that in Pegasus the influence in the policy making process by the
VVM was substantially larger than in the policy of basic mobility, can be related
to some of the factors discussed above (Verschuere, 2006). First, the fact that the
VVM has evolved over time into an agency with extended policy making capacities, and the subsequent normative shift in the agency that started to see itself as an
agency with a policy role, may be related to the fact that in the older policy programme (basic mobility, preparation started in 1995) the input by the VVM was
smaller than in the younger policy programme (Pegasus, preparation started in
2000). Second, the attitude of the VVM towards the policy programme of basic
mobility was less positive than their attitude towards the policy of Pegasus.
Extending service delivery and developing high quality transport nets in the
densely populated region of the Flemish Diamond has been priority number one
for the agency. The VVM had a firm belief that the policy of Pegasus was addressing the right policy problems in the field (fighting congestion in the Flemish
Diamond via extended public transport). Hence when the debate about the
accessibility of the Flemish Diamond is launched politically (e.g. parliamentary
debates around the Mobility Plan Flanders), the VVM has an incentive to be
closely involved in preparing public transport policies that address this issue. To
the contrary, the VVM was not that much in support of public transport policy/
programmes as proposed with basic mobility. They believed that basic mobility
was the wrong policy in the field of public transport. The VVM has always been
used to demand related work (delivering where the needs are highest) and in an
efficient and economical way (a heritage from the past when the resources for
public transport were poor). Basic mobility to the contrary stood for a supply
model of public transport, and induced the perception in the VVM of spending
money (delivering services where the demand is low is very costly, empty buses
also have to be paid).
As a final point for discussion, it is interesting to compare the Flemish governmental intention to create agencies in the pure form as suggested by the normative models, with the results of the empirical analysis. As found in the case, there
is no clear cut policy/operations divide between agencies and oversight authorities
today. This might lead to a paradoxical situation between the real world (interplay between agency and oversight authorities throughout the policy cycle), and
the formal world (the strict policy/operations divide as proposed by Beter
Bestuurlijk Beleid). Not only in this case organization, but also in other Flemish
public agencies a similar observation can be made, as is argued by De Rynck
(2004 [author translation]):
39
In many instances, all the policy capacity is within the agency itself [by referring to the
example of the VDAB, Flemish Employment Agency] . . . such agencies deliver the
knowledge to the cabinet to steer and control themselves . . . BBB makes an explicit
choice for the policy/operations divide whereby politics steer and administrations implement . . . I believe that we need more interwovenness between both spheres instead of
strict divisions for dealing with complex policy problems. (De Rynck, 2004)
Besides the fact that the involvement of agencies in the policy decision making
process is commonplace in many cases, this practice also seems to be defended by
the agencies because they want to keep close ties with their policy domain, their
clients, in other words the field demands participation (Rommel, 2006). The
formal world of the strict policy/operations divide is thus to a certain extent also
perceived as a threat by many agencies (e.g. year report 2005 of MOVI, the
representative organization of Flemish agencies). A quote from an interview is
exemplary for that:
We are against a strict policy/operations divide, because (today) in reality we perform
tasks of policy preparation, management and implementation . . . next to that, it is a threat
for our company and we also think that this will not improve policy: we have built up
the knowledge, and it would not be good to take this knowledge away, secondly, the
department will have to built up this knowledge, but their knowledge will inevitably be
theoretical. (interview with author [author translation], 16 May 2006)
Knowing that agencies today are to a large extent involved in the decisions made
during the policy process, and that this is a role they are aware of, and to a certain
extent even claim, the future will have to show to what extent the rhetoric of BBB
(a reform programme of which the implementation has only started recently) will
match with the practice in the field.
Conclusion
In this article I have shown for the Flemish case that the rhetoric of the
policy/operations divide does not always match with the practice of how different
actors are involved in the policy process. The results of the case study are to a
certain extent in line with a growing range of evidence provided on this issue by
other research in diverse settings. In the case of basic mobility, the agency was
hardly involved in the preparation and the determination of the policy, while on
the contrary this was the case for the policy of Pegasus. Whether the agency is
involved in policy matters or not seems to vary largely with policy programmes,
or with features of the agency (that may change over time, e.g. policy making
capacity). One important direction for further research will therefore be to investigate the circumstances under which administrative actors such as agencies are
involved in the policy process, or not. Earlier research may provide us with some
potential hypotheses to address this question.
40
Acknowledgements
I would like to thank the colleagues at the Public Management Institute at K.U. Leuven
(especially Geert Bouckaert and Koen Verhoest), for the support and inspiration during
the research that resulted in this article. Also I would like to thank the anonymous
reviewers of PPA for their valuable comments and suggestions.
References
Bram Verschuere is lecturer in public management and third sector policies at the
University College Ghent (part of the Association of the Ghent University) in Belgium. He
holds a PhD in social sciences, and his main research interests are organization of the public
sector, organization of public service delivery, welfare policy and third sector organization.
Part of his work has been published in Governance, Public Management Review, Public
Administration and Development, and Management International.
STRUCTURE
SIZE
43
euros of own revenue in 2004, makes the VVM one of the largest
agencies in Flanders, in terms of budget. The governmental transfers
have increased dramatically over the last five years, while the own
revenues decreased in the period 1999-2004.
Clients/travellers: There is a constant and impressive growth in number
of travellers over the last 10 years (from 215 million in 1996, to 413
million in 2004).
TASK
44
45
PREP: AG_MAN
PREP: AG_ENT
PREP: PR_POL
PREP: PR_ADM
DET: AG_MAN
DET: AG_ENT
DET: PR_POL
DET: PR_ADM
IMP: AG_MAN
IMP: AG_ENT
IMP: PR_POL
IMP: PR_ADM
Minimum
Maximum
Mean
SD
18
17
19
19
20
18
20
19
20
20
20
18
1,00
1,00
7,00
1,00
1,00
1,00
8,00
1,00
3,00
5,00
4,00
1,00
9,00
9,00
10,00
8,00
9,00
8,00
10,00
10,00
10,00
10,00
10,00
10,00
4,8889
3,2353
9,2105
5,0526
6,1500
3,6667
9,4500
6,7368
7,5000
8,4500
7,6000
6,7222
2,44682
2,22288
,91766
2,17239
2,20705
2,24918
,68633
2,25689
1,93309
1,35627
1,87504
2,13667
Minimum
Maximum
Mean
SD
17
17
18
16
18
18
17
15
19
20
20
16
3,00
1,00
6,00
1,00
5,00
6,00
3,00
1,00
4,00
4,00
5,00
1,00
10,00
10,00
10,00
9,00
10,00
10,00
10,00
8,00
10,00
10,00
10,00
9,00
6,5294
5,1765
8,4444
5,6250
7,8889
8,0556
7,4118
3,4667
7,3158
8,4000
8,4000
4,1250
2,03463
2,72110
1,09664
2,41868
1,67644
1,34917
2,12305
2,47463
1,88717
1,72901
1,27321
2,72947
Pegasus
PREP: AG_MAN
PREP: AG_ENT
PREP: PR_POL
PREP: PR_ADM
DET: AG_MAN
DET: AG_ENT
DET: PR_POL
DET: PR_ADM
IMP: AG_MAN
IMP: AG_ENT
IMP: PR_POL
IMP: PR_ADM
46