‘The Minister Wants it’: Self-Politicisation and Proxy Politics among Senior Civil Servants

In this article, I ask how senior civil servants (SCSs) practice functional politicisation. The literature suggests that they balance responsiveness with astuteness towards ministers, while maintaining neutral competence. However, functional politicisation is prone to affect this balance. Drawing on 160 h of ethnographic shadowing in Dutch government, I show three faces of functional politicisation; while directly advising ministers, in the preparation of policy advice and while working in the public eye. The findings suggest that senior civil servants actively try to align their fellow civil servants with their version of the minister’s wishes. This practice of ‘proxy politics’ calls for a shift in functional politicisation research from political ‘skills’ to ‘authority claims’ among senior civil servants. I conclude with urgent implications for politicisation theory and civil service practice.


Introduction
How senior civil servants (SCSs) interact with their political masters is a returning question in the literature on political-administrative relationships.The job of SCSs is usually considered to be a dual craft.Senior civil servants are both responsive to political imperatives, while maintaining a responsibility for neutral competence.So, to do their work well, SCSs must navigate 'responsiveness and astuteness' and maintain a degree of competent neutrality (Hartley et al., 2015;Hustedt and Salomonsen, 2018).
In recent times, forces of politicisationwhether in the hiring and firing of SCSs, their access to ministers relative to lobbies and other partisan actors, or the rise of political advisershave impinged upon this dual or balancing act, to the point that some scholars are weary of its effect on the civil service's capacity to give speak truth to power and remain neutral and non-partisan (Hustedt and Salomonsen, 2014;Peters and Pierre, 2004;Rhodes et al., 2009;Shaw and Eichbaum, 2020).Even in the absence of blunt moves towards de facto political control of SCS appointments, these more subtle influences on the nature of political-administrative relations have been dubbed 'functional politicisation' and are considered a key factor shaping the beliefs and practices of SCSs (Mayntz and Derlien, 1989).
Extant literature mentions broad containers of work and behaviour repertoires which indicate functional politicisation such as 'political-tactical advice vs substantial policy advice' (Christiansen et al., 2016), 'coalition building and conflict reduction (ibid: 1233), 'communicating and planning strategic external communication for pure political purposes ' (Salomonson et al., 2016: 210), 'Thinking politically' (Hustedt and Salomonson, 2018: 79) or the 'incorporation of political considerations into the dispositions and activities of bureaucrats' (Shaw and Eichbaum, 2020: 843).Although these labels give plausible insights into what kind of activities SCSs do and what functional politicisation means for their everyday tasks, they stop short of being observed in the lived realities of SCS's work practices.Such 'up-close and personal' studies at top levels in government hierarchies remain rare (Rhodes et al., 2007) but are much needed to advance our understanding of political-administrative relations beyond typological approaches, as helpful as they may be (e.g.Hood and Lodge, 2006), survey-based (Hammerschmid et al., 2017), and interview-based (e.g.'t Hart and Wille 2006) 'outside-in' research that continue to dominate it (cf.Bach and Wegrich, 2020;Peters, 2020).To me, this is not a dismissal of existing and insightful studies, but an attempt to diversify our methodological reach.
In this study, I aim to dig deeper into the forms and impact of functional politicisation by reporting 160 h of shadowing among SCSs, that is, secretaries-general and directorsgeneral in Dutch central government.Based on this observational data, I shed light on the 'craft' of public servants and contribute to a more nuanced everyday life sense of politicaladministrative relationships than much of the existing literature does (Goetz, 1997;Rhodes, 2016).By focussing on what SCSs actually doobserving their work in practicewe can better understand how they anticipate and respond to political actors and influences in their everyday practices (cf.Wagenaar, 2004).Additionally, this article contributes a study of politicisation in a national context which is generally considered non-politicised.
I argue that functional politicisation is in practice not just about command of political skills and the prevalence of political-tactical advice over policy advice (even though that seems the case), but that it also manifests an authority claim allowing some SCSs to represent the minister and her authority by proxy vis-à-vis fellow civil servants.I refer to van Dorp this as 'proxy politics'.Although political-administrative relationships are generally understood as interfaces between political principals and civil servants, proxy politics are a practice of anticipating and strategizing about these political-administrative relationships in interactions that occur among civil servants.
In the remainder of this article, I first discuss current research on functional politicisation, particularly as it applies to SCSs in the Netherlands' national government, the empirical setting for this article.I then account for the methodology of ethnographic shadowing.The findings section provides an 'up-close' narrative of how SCSs interact with their principals, how they represent these principals by proxy to their colleagues and how this is shaped by the realities of working 'in public'.Finally, I discuss the theoretical and practical implications of the findings and suggest ways for future research.

Functional politicisation in the core executive
Politicisation of the civil service is often understood as a mechanism for politicians to control government bureaucraciesa way for Ministers to get what they want.It makes a difference whether civil servants are recruited based on their professional merit or based on partisan allegiance (Cooper, 2020).However, politicisation is not just found in recruitment and selection procedures of civil servants.Politicisation comes in many shapes which carry different implications for the relationships between politicians and civil servants.In their literature review, Hustedt and Salomonsen (2014) present a threefold typology: formal, administrative and functional politicisation.
Formal politicisation is usually defined as 'the substitution of political criteria for merit-based criteria in the selection, retention, promotion, rewards, and disciplining of members of the public service.'(Peters and Pierre, 2004: 2).This can take many forms, including the personalisation of the bureaucracy, meaning Ministers recruit civil servants whom they connect well personally and patronage (Hustedt and Salomonsen, 2014: 749).This does not necessarily mean that these politicised or personalised recruits are unprofessionalthe German case with politicised executive level appointments proves to result in professionally trained civil servants with a career in government (Bach and Veit, 2018).So, although ministers may seek to appoint loyal officials, they require 'some modicum of ability' too (Hood and Lodge, 2006: 86).
Second, administrative politicisation refers to attempts to control the functioning of the public service in ways other than through targeted recruitment (cf: Shaw and Eichbaum, 2020).The most notable form has been the introduction of ministerial advisers or other explicit partisans such as chiefs of staff within the bureaucracy (Shaw and Eichbaum, 2018).Specifically, administrative politicisation occurs when political advisers behave towards civil servants in ways that 'offend against the principles and conventions associated with a professional and impartial civil service' (Eichbaum and Shaw, 2008: 343).For example, partisan staff in the department can obstruct civil servants' access to Ministers or add a partisan 'spin' to impartial civil service advice before the Minister gets to see it.The emergence of ministerial advisers has sparked concerns about the civil service's ability to deliver 'frank and fearless' advice and 'speak truth to power' (Hustedt et al., 2017).
The third type is the politicisation of work and behaviour of civil servants, rather than actors or recruitment and selection.'Functional politicisation represents a mechanism by which the civil service performs politically responsive bureaucratic behaviour' (Hustedt and Salomonsen, 2014: 750).'Political' work is part and parcel of the senior civil servant's craft (Goetz, 1997).They balance 'responsiveness' with 'political astuteness' (Hartley et al., 2015;Mulgan, 2008) as they anticipate and influence the political nature of context in which they work.The demands placed upon them by their political principals and the shadow of ministerial responsibility and accountability permeates the policymaking process.This type of politicisation implies that the craft of SCSs is more about being strategic and less about being the neutral purveyor of subject matter knowledge and advice about practical feasibility of policy ideas (Belloir and Van Den Berg, 2020;Christiansen et al., 2016).
Shaw and Eichbaum suggest an important factor that sets functional politicisation apart from the first two types.Rather than responding to exogenous stimuli from Ministers or Ministerial advisers, functional politicisation is more likely, they argue, endogenous: it is 'behaviour that bubbles up from within the civil service ' (2020: 843).Seen in this way, the civil service self-politicises, anticipating the Minister's needs and desires, for example, through 'mindful self-censorship' or 'anticipatory compliance' ('t Hart, 1994).Reading 'what the Minister wants' is considered a sine qua non for SCSs.However, others do find evidence for external stimuli.For example, research in Danish bureaucracies found a positive relation between functional politicisation and media awareness (Salomonson et al., 2016), reminding us that SCSs believe to work in the public eye (cf: Grube, 2019).
Situated nearer to the Westminster and Scandinavian than to the Napoleonic administrative tradition, the Dutch civil service system does not have a tradition of formal politicisation ( Van der Meer, 2004), nor have its appointment processes been ranked as highly politicised (Bach et al., 2020).Although Van Dorp and 't Hart (2019) provide a case study of intricate 'court politics' (Rhodes, 2013) in the Dutch Prime Minister's Office in the early 2000s, Van Den Berg's (2018) overview study finds that ministerial adviserseach minister has oneare by and large working in tandem with the civil service rather than acting as agents of administrative politicisation.Still, in other work he does observe that, 'political considerations in the policy process seem to have become more pronounced through the work of the Minister's direct entourage ' (Van Den Berg, 2017: 78).Elements of functional politicisation may well be slipping into the operating model of the Dutch core executive (Belloir and Van Den Berg, 2020).Recent governmental audits concluded that 'unwritten rules' incentivise SCSs to shield their Ministers rather than advise them frankly and fearlessly and report the strong sense of duty to deliver on political wishes among SCSs (ABDTOPConsult, 2020;ADR, 2019).

Political work and political behaviour in the Dutch civil service
The analysis of functional politicisation and civil service behaviour usually covers two broad categories: 'neutral' policy advice and 'political' advice regarding coalitions, tactics and media spin, including anticipating political implications of decisions, tailoring proposals to the government of the day (e.g.Christiansen et al., 2016).Working towards van Dorp favourable media perceptions are considered elements of functional politicised behaviour.The neutral competence versus political tactics binary is a convenient trope but cloaks the more subtle realities in SCSs' work contexts and coping styles.Van Dorp & 't Hart (2019) also reject this divide: in their analysis, Dutch SCSs consider their policy advice politicalstrategic, and their political-strategic advice about policy.So, as long recognised, SCSs' work includes both policy and (non-partisan) politics (Mayntz and Derlien, 1989).To phrase the indivisibility of policy and politics, Putnam (1975) coined the 'political bureaucrat', while Rhodes (2011) prefers 'political-administrators'.An in-depth analysis of functionally politicised work of SCSs cannot do with binaries alone.
Moreover, as Mulgan (2008: 347) points out, being responsive implies that SCSs' neutral advice should take account of government priorities, rather than attempting to be impartial between competing political viewpoints.In this view, SCSs serve this government (and the next), so they are by no means impartial.In normative terms, he proposes the principle of 'constrained partisanship' arguing that although inevitably political, SCSs should refrain from being too obviously partisan, particularly in partypolitical matters and the open advocacy of government policy (2008: 348-350).Their command of 'political' skills, however, is crucial to successful policy advice.Aucoin's (2012) rival description of 'promiscuous partisanship' argued that SCSs are (increasingly) expected to be open advocates of government policies, either within or outside the departmental organisation, suggesting an explicit functional politicisation.
Though my purposes are not normative but empirical-interpretive, the containers of functionally politicised work and behaviour provide a bandwidth ranging from 'too much' to 'too little' responsiveness and astuteness.They also highlight three faces of functional politicisation.In the empirical part of this article, I will study these three faces: the direct inter-personal political-administrative interaction; the representation of the minister's authority by proxy to fellow civil servants within the administrative organisation; and the ways SCSs deal with being in the public eye.
Observing SCSs in the field I set out to shadow among 18 senior executives in the Dutch civil service as they went about their business (Noordegraaf, 2000;Rhodes et al., 2007).This qualitative approach gives unique insights into the practice of SCSs and allows us to open up the black box of functional politicisation.These SCSs led ministries, headed up service-delivery agencies or oversaw public regulators.They were either chief executive officers or members of the executive board of their organisations.Their formal titles included 'secretary-general' (SG), 'director-general' (DG) or 'inspector-general' (IG).
18 participants (12 male, 6 female) agreed to my request, four denied my request and six did not respond.All field visits were between February 2018 and March 2020.This period falls within the prime minister Rutte-III cabinet.The SCSs were targeted using pragmatic sampling resulting in an even spread across multiple departments and agencies (common in elite studies, cf : Peters, 2020: 218).
The claims of this study are based on notes (84,000 words) that I made during 160 h of shadowing 18 SCSs.Shadowing each SCS for a day or more has provided a unique opportunity to study their work up-close.For any observational study, duration in the field matters.Between 2018 and 2021, I have been moving in and out of the field.This approachcolloquially called 'hit-and-run ethnography'balances breadth with depth (Rhodes, 2014).Observing a larger number of participants has been tremendously helpful in spotting differences (breadth), whereas observation (even if 'just' for 1 day) has revealed aspects of civil service politicisation that interviews are unlikely to reveal (depth).Furthermore, the individual participants often interacted during the observation days, meaning I got to observe participants multiple times in different settings (one or 2 days as protagonist, other days as incoming colleague, antagonist or bystander).
The fieldnotes contain observations, verbatim excerpts of conversations and 'interviews-to-the-double' to capture the beliefs underpinning the observed practices (Theron, 2020).Fieldnotes relate to direct political-administrative interactions, policy preparation and various other forms of work.I made sure to include headnotes too, my running commentary on what 'I wondered as I wandered' during the fieldwork days.To me, headnotes make plain that the researcher is 'not inoculated from the effect of what he bore witness to ' (De Rond andLok, 2016: 1970).

Access and positionality
During a brief interview, I explained the purpose of the study and got acquainted with the participant.After gaining the participants' consent, access was still to be negotiated.All participants were willing to introduce me to their guests and allowed me into most meetings they attended.Sometimes I was temporarily denied access, for example, if personnel matters were discussed, when the meeting was unusually (politically) charged or when my participant was uncomfortable in bringing a guest to a meeting where s/he him/herself was invited as guest.Participants would usually debrief these meetings afterwards to excluded colleagues and me.
In assuming positionality, I aimed for 'social shrinking' as Goffman (2014: 235) coins it; I was in the room but tried to blend into a social non-factor by dress, silence unless directly prompted, posture and chair placement.Still, a degree of 'undesigned relationality' remained, as these understandings were not fixed and sometimes implicit (Bell, 2018).When I asked, participants said they were not bothered by my presence.
The up-close and personal approach priors me to try to feel with SCSs and see the politics-administration interface from their perspective(s).That said, I am not speaking for them.I invited all participants to reflect on the manuscript.Although they recognised and authorised this article's interpretation, three participants wished for more nuance or alternative phrasings.Therefore, this account is mine and not all participants fully agree.

Data analysis
The fieldnotes were thematically analysed by means of coding and memoing, aiming for theory refinement based on the qualitative material (Eisenhardt, 1989).Formal coding and discussing the data with colleagues took place parallel to the fieldwork.'Politicaladministrative relations' was a sensitising concept throughout the fieldwork and van Dorp analysis, but the concept of 'functional politicisation' emerged only after the fieldwork was done.The first rounds of coding revealed profound but well-documented findings about the work of SCSs, foregrounding their busy days, the centrality of meetings and department-agency relationships (cf: Noordegraaf, 2000;Rhodes, 2011).However, in next rounds of coding, a fresh perspective emerged: the practice of proxy politics, see Table 1 below.
In the analysis, I argue that civil servants do 'political' work -I rid myself of unhelpful binaries including the politics-administration dichotomy.In making this conceptual leapin recognising work as 'political' -I was guided by abductive coding, iteratively interpreting fieldnotes with elements from 'functional politicisation' discourse.I began to see work as 'political' when SCSs defended their claims with the approval of the political principal they served, whose legitimacy was generally unquestioned, thereby attempting to load some of that legitimacy upon themselves.How SCSs did this in practice, is subject of analysis below.

Findings: Three faces of functional politicisation
The fieldwork shows three faces of functional politicisation.First, functional politicisation was visible in the direct interaction between civil servants and ministers.A second face showed when SCSs represented ministers towards their fellow civil servants.The third face appeared when SCSs realised they were working in public.
Face 1: Serving ministers responsively and astutely Most SCSs I observed worked directly to serve their principals.For all but one, this meant they were personally serving the (junior) Minister responsible for their department or agency.This implied they performed roles as advisers to the Minister, were responsible for executing the Minister's policy directives, warned against harmful policy proposals, and counselled on any and all issues upon request.Even if they would not see the Minister This vignette illustrates how SCSs were 'responsive' by partaking in a daily meeting with the Minister where he is free to speak his mind, complimenting the Minister and immediately putting his plans in action by providing follow-up.Yet, they were also 'astute' by pushing back on the generation of all too many new ideas, knowing the department can only do so many 'extras', given upcoming elections and the going concerns of the policy agenda.
The practice of serving the principal was the most prominent feature of the working lives of the departmental SCSs I observed.They 'cocooned' in adjacent offices and had but to cross the hallway to see their Minister(s), sometimes multiple times a day (Fieldwork, June 2018).Their language for these impromptus was 'walking in' or 'tackling the minister', bearing connotations of astuteness and low threshold access.One departmental DG had seven face-to-face interactions with his ministers, accounting for 3.75 h of this observed day (Fieldwork June 2018).This facetime could include reassuring the Minister they were keeping his interests and credibility in mind, while acknowledging to him that there are other forces (and Ministers) at play too (Fieldwork, July 2019) or sitting in on negotiations and coming up with fresh ideas that help the Minister's position in a negotiation, without setting a damaging precedent a next Minister would have to deal with (Fieldwork, June 2018).
For SCSs in executive and regulatory agencies, the intensity of direct interaction with Ministers was considerably lower.Some had regular appointments with the Minister, and these would be complemented by ad-hoc meetings and calls, but the imperatives of dealing one-on-one with the principal, catering to their needs were less salient to them.In van Dorp effect, an agency director lamented his troubles reaching the DG of the department to meapparently speaking to the minister was beyond reach for him (Fieldwork, May 2018).Departmental SCSs would be his first point of contact.Still, their commitment to the 'primacy of the principal' was palpable on many field visits.In a meeting, one IG voiced this belief pleading with his fellow IGs: 'We are the Minister's inspectors.Our work falls within the Ministerial responsibility ' (Fieldwork, March 2020).With Mulgan's (2008) words, they professed neutrality, but not impartiality.An agency head explained that, although she ran an arm's length agency, she would make efforts to 'comfort' the Ministeragain a fine line between responsiveness and astuteness: 'One of the things I do, is to comfort the minister because he's insecure about not being a lawyer, while I am.I also add comfort to his position by reframing a parliamentary debate about an issue at my agency from ''Misconduct at'' my agency to ''Problems at'' my agency.Furthermore, I cut my vacation short to attend the debate.I didn't think I was needed, but my people were quite tense, so I thought it was better if I'd be there ' (Fieldwork May 2019).
Ministers were not the only principals -Parliament was always around the corner.The Dutch dualist contextwhere Ministers are not MPsfosters a dichotomy between the executive and the legislative branch.In Dutch parliamentary democracy, Parliament unambiguously has the final word, SCSs believed.Because the executive is not part of the legislative, this suggests a dual loyalty for SCSs: they serve both the legislative and the executive.However, as direct contacts with the legislative were actively discouraged, SCSs perceived it as stepwise: ministers embody parliament's legitimacy.Though they would not necessarily endorse Parliament's majority view, they acknowledged it as legitimate and as an important impetus for their work.Senior civil servants managed up to Parliament directly, through hearings and letters, but usually indirectly by preparing ministers for upcoming debates.
On occasion, a SG briefed MPs in person during a hearing setting.In this instance, he sat in front of a parliamentary committee, reporting on the progress of the department's organisation culture and change management program.Given the discouragement of direct parliamentary-administrative interaction, these 'technical briefings' were considered rather unusual.On a car ride, he confessed he was pleased with how it went: 'we hit a good groove' he said, 'much better than the previous hearing'.His nerve had been tested, but it went well.During the follow day, the SG was repeatedly asked about it by his departmental colleagues, stressing the alien nature of such hearings.(Fieldwork, September 2018).
Interestingly, some SCSs saw Parliament as their primary political master, coming even before their Minister.During a meeting, an IG urged her colleagues to provide her with policy advice, so she could accommodate 'Parliament and Minister,' stressing the former not the latter (Fieldwork, November 2019).Though, when she voiced this minority view to fellow regulatory chiefs, most disagreed, reaffirming the predominance of the Minister as primary principal for departmental, frontline agency and regulatory agency heads alike (Fieldwork, March 2020).
Regulatory SCSs often had the opportunity to consult MPs or be consulted by them, without a Minister or departmental officials mediating the relationship.They could send letters to Parliament, signing them without a Minister's signature.A regulatory chief was quite confident about her position vis-à-vis Parliament.She counted on Parliament as her main authorising environment.Of course, her respective Minister could give directions, but only through a public letter notifying Parliament.She counted on the publicity side-effects to make sure her minister would ration his/her decrees wisely (Fieldwork, November 2019).
Though SCSs seemed generally to Parliament's wishes, they practiced forms of astuteness too.A brief excerpt from a departmental board meeting highlights how SCSs bought time by writing a trivial letter in response to parliamentary questions: Next item on the agenda: incoming parliamentary questions.A DG briefs his colleagues: 'the letter is written, with meaningless answers, because the issue is discussed in Cabinet only later this week'.The discussion of all questions takes two minutes.(Fieldwork, June 2018).
It was not hard to recognise some of the 'functional politicisation' traits in the SCSs' serving and steering of ministers.Though they dealt with personnel, policy and implementation, they practiced 'thinking politically' (Hustedt and Salomonson, 2018: 79) too and made efforts to incorporate 'political considerations into the dispositions and activities of bureaucrats' (Shaw and Eichbaum, 2020: 843).

Face 2: Politics in the shadow of politicians
Although I saw SCSs with their ministers, the bulk of the observed days was spent without politicians in sight.This could, of course, be a sampling consequence of my field access (although I do not have any indication that suggests this), but it was also a perennial complaint among SCSs and their staff (Fieldwork March 2018).During observation of a departmental SCS in the summer of 2018, the daily morning meeting with the minister was cancelled -'again'and the weekly staff meeting shortened to a mere 30 min.As they sighed under their breath, a DG concluded the minister 'apparently does not consider it important enough' (Fieldnotes June 2018).Likewise, a SG lamented that his weekly bilateral with the minister was the first thing that would be cancelled when the minister was busy (Fieldwork July 2019).
Even if politicians were absent, politics was all around.SCSs practices a distinct form of politics, I call 'proxy politics.'Proxy politics refers to the practice of SCSs claiming the authority of their principal for their own while thinking and/or acting on his/her behalf.A reoccurring experience throughout my fieldwork was that SCSs would encounter a fellow civil servant who would back up an argument with the phrase 'The Minister wants this', assuming that would be the end of it (e.g.Fieldwork, June 2018).This was particularly observable in interactions between departmental SCSs and agency SCSs, but not exclusively.In a similar vein, departmental civil servants would encounter fellow departmental civil servants, who claimed a still more significant authorityusually the Prime Minister's or the Minister of Finance's.

van Dorp
A simple example illustrates what 'proxy politics' looked like in practice.A SG made a site visit to a frontline agency of her department.In a meeting she was asked to respond to case studies by saying how the Minister would decide in this situation (Fieldwork, May 2019).In this example, the SG was perceived as a proxy of the Minister, providing upclose insights into the Minister's brainsomething the frontline agency employees lacked.In such cases, SGs were perceived to speak for the Minister beyond rebuttal.
A lot of SCSs' work is paper driven, as can proxy politics.Consider when a departmental DG was correcting departmental notes to the Minister one early afternoon: Using a pen, he is correcting a departmental brief to the Minister.He erases, highlights, and draws a circle around specific words.Without looking up, he comments: 'look, it says that ''it is in the Minister's interest to read this carefully ''.That line will just annoy her,' as he erases the sentence from the brief.'She is swamped with briefs and has no idea what this is about.'He goes on to explain that the Minister cancelled her daily morning meeting due to other urgent business and a trip to Sweden this afternoon (Fieldwork, June 2018).
Put simply, some SCSs thought and spoke for their Minister.They committed proxy politics while judging what their Minister would (not) like to hear or should hear or would time incoming issues to a more opportune moment.In doing so, they blended gatekeeping with substantial directives by proxy.
How SCSs dealt with the proxy politics of their colleagues was shared by a yet another SCS telling of his experience as IG.Every now and then, there is a suspicion that specific products malfunction or otherwise deviate from the appropriate regulations, he explained the context.When that is the case, products can be recalled, often at great expense of businesses and plausibly the Minister's political capital.Clearly an incendiary decision: 'I was informed that the situation was quite serious, so I decided to recall all products.When I informed the Director-General of my intention, he was shocked and argued that I could not do that.Referencing legislation, he claimed he had the authority to make that decisionto him it was a policy matter.I thought, based on different legislation, that the mandate to recall products was mine.We had to argue before the Minister to find out who was right ' (Fieldwork, February 2020).
As the interview excerpt illustrates, SCSs were involved in proxy politics.Both acknowledged that the Principal is always right.However, many civil servants had little access to the Minister, contrary to the IG above, meaning they could not easily appeal and will need to cope with the Minister's proxy in front of them.
In May 2018, a manager of a large arm's length urged his staff to provide him with more information and a convincing letter he could send to the Minister, which would suggest 'please sign here' to the Minister, warding off further dispute (FNWB).This agency manager made use of both careful preparation of staff and statutory standing (after all, the agency sits at arm's length) to pre-emptively overcome proxies of authority, a strategy also endorsed by other SCSs (e.g.Fieldwork, March 2020).
Departmental SCSs thus acted as authority proxies but encountered them too.Like all members of the core executive, they are bosses and being bossed.This may look like this: On March 19 th 2018, a secretary general discussed upcoming budget negotiations with his departmental CFO and a financial adviser before he would go to the Ministry of Finance.'What's the minimum we need?' is his simple question to present staff.The department faces urgent challenges which require staff and budgets.As they present the figures to him, the SG does the maths on his notepad, in millions of Euro's This vignette illustrates that inter-departmental relations can be a source for zero-sum competition.However, like the preceding vignette of the IG recalling products, the SCS's approach was aimed at undoing the proxy politics of the Ministry of Finance negotiator, not at bargaining for a bigger part of the pie.In this instance, a civil servant's relationship with the minister can fuel proxy politics, but also a way to overcome it: calling the interdepartmental bluff and taking the matter to Cabinet.
The practice of proxy politics is more than shadow boxing by and among civil servants; it has real import that SCSs could not undo even if they wanted to.'t Hart and Kuijken (2018: 56) describe the predicament of a SG in the Prime Minister's Office who brings word to a Minister saying the Cabinet cannot likely sustain his position any longer, upon which the Minister draws his conclusions.A clear example of a civil servant acting on behalf of unquestioned authority in the core executive which leads to compliance, not undoing.In my own fieldwork, a manager of a frontline agency explained how departmental civil servants would pressure him into compliance and how he had difficulty of countering this.The frontline agency's manager spoke of 'Implementation tests' [Dutch: Uitvoeringstoets or HUF-toets].These tests are designed to quickly evaluate whether proposed legislation can actually be implemented.Some legislation cannot.Sometimes because it is unlawful, not supported by current ICT infrastructure, or the respective agency does not have the necessary skills, specialised staff or resources.If the outcome is 'no', the test provides a quick feedback loop for departmental policy makers.However, the manager explained, 'the department does not accept a 'no'.If they want X, we need to give them X.Especially when the proposal is already published in the Cabinet Agreement.' (Fieldwork, May 2018).
The frontline agency, their manager explained, has a strong 'sense of duty' to comply with political requeststhe SCSs did not always want to undo the proxy politics, as they had internalised the proxy's authority.Later that day, when the agency's Board discussed a proposal which was not practically feasible, this was epitomised by the chairperson's words 'I want us to be able to do this':

van Dorp
The chairman reads that the implementation test labelled the proposal as 'not feasible'.It means the agency would not be able to deliver on the Minister's request.His voice sounds dismissive.'I want us to be able to do this anyway', he adds.The attendants quickly come up with ideas of how they could work around the agency's shortcomings; using the institutional, legal and financial know-how of other agencies, such as the Tax Office.One civil servant protests and stresses the unfeasibility of the proposal that was underlined by the implementation test.When colleagues appear prone to the chairman's voice, the civil servant acquiesces (Fieldwork, May 2018).
This subparagraph illustrates how 'the minister wants it' was both a rhetorical trope that civil servants used to subtly whip their colleagues in line, but also a dominant force among Dutch civil servants to deliver for the government of the day.Although proxy politics were countered, the final example illustrates that in the end some SCSs and their subordinates would acquiesce.

Face 3: Seeking privacy in public
The pervasiveness of being in the public eye was ingrained in the SCSs I observed (see also Grube, 2019).This publicness has a distinct effect on being responsive and astute, both directly and by proxy.I saw them come to terms with the fishbowl realities of working in the public eye in two ways; they tried to save face by anticipating the potential fallout of unpopular decisions of policies gone awry and, second, they tried to seek privacy in public.For saving face, consider the following vignette from the fieldnotes: An agency executive makes sure that his other board members will not accidentally sign an invoice for new furniture in the executive rooms, totalling €130,000.Apparently, the invoice had turned out much higher than expected.'We're not going to sign that invoice', he stresses, seeking active confirmation from his fellow board members.They nod.In response, a present non-board member points to the threadbare furniture in the room and reminds the board members that they do need new furniture.The director concurs, but points to the publicity it may arouse: 'we're really careful with this'.Though possibly unaware, his caution is not farfetched.Some 15 years ago, a similar invoice for office furniture (admittedly a much higher sum) in a neighbouring agency led to the immediate removal of its director, heated debate in Parliament about alleged 'golden faucets', and a dint in the department-agency relationships (Fieldwork, May 2019).
SCSs imagined what something on their table would look and feel like to the general public.This form of political nous was considered essential to the SCS' craft (Rhodes, 2016).So, it made sense SCSs were not worried about public expenditure as such, but about the reputational effects of 'bureaucrats touching up their offices'.Similarly, an agency manager was approached to finance a theatre production that dramatised the agency's work.When she voiced her concern to the aspiring theatre producer, she did not worry much about donating €100K to the production but with the possibility of being depicted as buying influence over the production 's content (Fieldwork, May 2019).
Keen on upholding the reputation of their organisations, most SCSs were careful on what they would voice in publicoften limiting themselves to things they would not regret (Schillemans, 2012).Consider this example from the fieldwork: In a departmental board meeting, a director points to the notes that have just been distributed around the table.A line about the department's relationship with the sector it serviced reads that 'no consensus has to be reached with the sector'.He draws attention to the line and points out that while it may be legally correct, they would have trouble justifying it in public when it would be made public through a FOIA request (Fieldwork, March 2018).
In yet another case, an agency head was asked to comment on the slogan for the agency's annual staff party.She wondered out loud what it would look like on the papers' frontpages.She justified her hesitation to her staff member with: 'I think in headlines ' (Fieldwork, May 2019).
In contrast, good news was made public swiftly and gregariously.Some SCSs enjoyed a policy victory with the minister signing an agreement in a carefully staged, well photographed setting (Fieldwork, June 2018).In a different setting, a departmental SCS urged his staff to send a letter containing reports of considerable policy progress to Parliament before the upcoming May holidays (Fieldwork, April 2018).Simply put, the Minister could always use some good newsand SCSs made sure everyone was well attuned to it.This was very apparent in meetings I observed.For instance: A departmental DG opens an interdepartmental meeting about a new policy plan.Due to the ill-fated history of the policy plan's predecessors, he stresses that the Minister considers visible progress on this policy plan very important.He adds to the six colleagues around the table that he wants to 'underline her concern, before they would start off'.Later, but in response, an executive agency DG proposes to cut up the new policy plan into ten policy plans, explaining this would allow both their Ministers to 'display' quick wins and better compete for attention.[…] When the departmental DG closes the meeting, he thanks everyone and says he's glad they had sent a few 'shots across the bow ' (Fieldwork, June 2018).
Though SCSs worked hard to save face for their organisations and principals, they could not always control their agency's behaviour and reputation.'Unauthorised' information leaks from anonymous sources were a source of professional embarrassment.A discussion in a departmental board meeting is exemplary: At the meeting's close, they discuss the recent coverage in De Telegraaf about the newly refurbished government building Rijnstraat 8.The newspaper item mentioned the cramped office facilities, including staff complaints about the building and the public servant's wellbeing.'It includes a picture of public servants queuing for coffee', stressing existing stereotypes.'Nice pictures', is the sarcastic reply.Though they do not deny the problem, they regret that this was leaked through the largest national newspaper.They collectively suspect the employee's representatives as the source but concede they cannot prove a thing.The chairperson gives the board members an update of how the Government's Real Estate Agency (the building's legal owner) has acted since the coverage broke.Now they regret that they could not get them to the cooperate, while now it is in the papers, they are eager to help.'We're living in a mediacracy ', a DG murmurs (Fieldwork, June 2018).
Similarly, when an agency manager gave her colleagues an update on upcoming reports, she regretfully anticipated plausible leaks referring to icy relations with a wellread Dutch broadsheet paper.'The [name of broadsheet paper] is not our friend at the moment ' (Fieldwork, May 2019).In this light, my presence as observer sometimes prompted questions about the confidentiality of the meetings I attended.Some attendants wanted reassurances that they could 'speak freely', meaning their words would not leak (Fieldwork, July 2019).For instance, a departmental spokesman warned me to keep quiet about what I heard, while making a gesture zipping up his lip.Still, it took less than an hour, before he proposed to 'share' something 'on background'which was code for anonymously leaking the departmental position on a policy issue (Fieldwork, June 2018).This marks the ambivalence SCSs felt about working in public: it was considered inevitable, but in practice they also sought privacy to 'speak freely' (Fieldwork July 2019).So, paradoxically, when working in public, SCS preferred to remain invisible.If possible, they were one step removed from shows of publicness and let their Ministers champion them in public: ministers and media awareness, rather than formal controls of the bureaucracy or its colonisation by partisan staff.
As a way to control bureaucracies, functional politicisation starts with ministers.The fieldwork provides many clues that support this.For example, a minister who instructed her DG that she needed 'visible' policy progress, directing this DG to make that his main message in meetings with colleagues.However, of course, open to interpretation, such instructions arguably steer SCSs to 'think along' with their principal and strive for the minister's increased standing in parliament and media.It makes an explicit case of 'incorporation of political considerations into the dispositions and activities of bureaucrats' (cf.Shaw and Eichbaum, 2020: 843).However, the main thrust of the empirical paragraph is not that SCSs thought along with their ministers, but that they thought for their minister.Although ministers were always on their mind, SCSs seemed prone to selfpoliticisation.Or, in Shaw and Eichbaum's (2020) phrase: it bubbled up inside of them.
If functional politicisation is a mechanism, as Hustedt and Salomonson (2014) hold, this study stresses the practice of SCSs to anticipate and strategize about politicaladministrative relationships within the bureaucratic organisation.To unpack this, I used the phrase proxy politics.Some SCSs, most notably departmental managers, were so attuned to the principal's preferences that they would easily substitute for the Minister by evoking their wisheswhich was met by few objections in the department.Theymy fieldwork suggestsacted and thought on behalf of the principal, with or without them knowing.More importantly, bystanders and colleagues were often in the dark about the principal's actual needs and desires.This makes 'The Minister wants it' both an in vivo trope that was used by SCSs, but also a metaphor for politicisation.Though never partisan, and always with a keen eye for professional distancing from the rubble of politics, this comes close to what Aucoin (2012) coined 'promiscuous partisanship' but within their organisations.
When observing SCSs this practice was hard to pinpointwhen siding with their perspective it was the rational, if not, only thing to do.Translating the Minister's latest ideas was their prerogative, as they were often the only ones who knew about them.Likewise, working towards the minister's future wishes was simply the professional thing to do, and self-censorship could save them a pointless row (cf: Kershaw, 2010).Only after observing civil servants in executive and regulatory agencies, I saw how seemingly coherent perspectives, professional pride andat face valuereasonable requests were met with an alienated and awkward response from departmental ventriloquists condensed in 'the Minister wants it'.
The interesting thing about proxy politics is that it describes political work being performed without politicians present to perform it.It is the realm of the 'political bureaucrat' (Putnam, 1975), 'Political-administrators' (Rhodes, 2011) and 'stand-ins' (Askim et al., 2017).Though ministers, MPs and the public loomed large in the conversations, considerations, negotiations and practices of the SCSs, they wereto the civil service's dismayoften not around.However, the wellbeing and success of the Minister shadowed over conversations.Civil servants spoke for their principal, considered what they needed (or not), they interpreted parliamentary debates and anticipated media headlines and public opinion.Although functional politicisation usually regards political skills and nous, gearing policy and advice processes towards the wellbeing and survival of the principal, proxy politics transcend skills.Its rationale seems to have been internalised in the hearts and minds of Dutch SCSs.It is not just an exogenous mechanism with behavioural consequences; it has become an endogenous practice of the senior civil service's habitus by means of self-politicisation (cf: Shaw and Eichbaum, 2020).
Strikingly, the most noticeable element of functional politicisation among the Dutch SCSs is not their command of political skills (although craft skills matter).It is their inclination to act and think for their principals with or without them knowing.Enacted through seemingly mundane practices, which I show in this paper, proxy politics resemble an authority claim civil servants make vis-à-vis their fellow civil servants.All professed competent neutrality, yet a number of them synthesised civil service neutrality with an ambassador role representing the Minister to their colleagues.As Dutch Ministers are not members of Parliament, the parliamentary-administrative relationships deserve more serious attention.The varying degrees of loyalty to Parliament that SCSs experience and how that shapes their behaviour remains an understudied dynamic.
This study is not a systematic comparison of different contexts (departments and/or agencies).However, it did seem that where SCS stand depends on where they sit.This then, nuances the claim that the Dutch core executive is increasingly functionally politicised (Belloir and Van Den Berg, 2020).Instead, my fieldwork indicates a differentiated picture of a contested civil service.Though some civil servants may have resembled 'promiscuous partisans' who practiced proxy politics (usually departmental managers, but variations existed), most seemed more likely to be 'constrained partisans' and others even tried to undo these functional authority claims altogether or were relatively unbothered by these dynamics.If these dynamics are found in other contexts too, it is important to study why SCSs turn to proxy politics rather than to a different set of practices in order to better understand expressions of functional politicisation.
Also, the practice of proxy politics prompts questions regarding the balance (or imbalance) of clout among and between departments and agencies and the opportunities for effectively 'speaking truth to power' by those without proximity to the principal.If such opportunities are impaired by functional politicisation, policy results may be hampered as critical voices are suppressed.New research could look into this relationship further.Hustedt and Salomonson (2014: 761) called for parliamentary opposition and the broader public as sensitive observers in this, butdespite fishbowl experienceswhether these are equipped to do this remains in question.
On balance, the fieldwork stresses SCSs' propensity to serve their principal, rather than their relative autonomy as stewards, executives and/or regulators of a public policy domain.As their craft entails serving both this government and the next, a degree of autonomy remains a critical feature of a neutral bureaucracy.If SCSs would want to increase their autonomy, simple solutions are elusive.Inevitably, it includes recalibrating the balance between responsiveness and responsibility, while maintaining both, at a pace all can stand.It may be necessary, but who knows if that is what the minister wants.
politicisation literature, I conducted ethnographic fieldwork among SCSs in Dutch government.I attempted to open the black box of functional politicisation behaviour, in particular when working in the public eye.The analysis shifts the perspective from political skills to authority claims by SCSs.In closing, I recap the argument and suggest ways for further research.

Table 1 .
Themes and sub-themes in data analysis.'look,itsaysthat''itisinthe Minister's interest to read this carefully''.That line will just annoy her,' as he erases the sentence from the brief Working in public 'I think in headlines'every single day, or not even on a regular basis, 'the Minister' would be prominent in their reasoning, arguments and discussions.'Serving'the principal meant both enabling the Minister to govern and execute a political program, but it also meant keeping the Minister out of trouble, for instance by steering him/her away from bad ideas, or trying to deflect mistakes from the Minister's reputation.The picture is one of close relationships with frequent and wide-ranging interactions, including:A SG, a DG and two junior advisors sit around a conference table for the daily early morning phone call -'morning prayer'with their Minister.The latter is en route to the department, digesting today's papers and coming up with new ideas.He pitches an idea and wonders out loud if it may be a solution to the employment issues in their sector.The SG compliments the Minister calling it a 'good idea' and proposes to contact the SG at the Defence department to discuss it tonight in the margins of a planned meeting.The Minister speaks about 80% of the duration of the call, one of the DGs does most of the talking on this end of the call.The two junior advisers remain entirely silent throughout.
The last topic of the call is the Mid Term Review of the department.The coalition parties in Parliament were asked to come up with new ideas for the second half of the government's term, but the yield was low.The Minister wants more ideas.The DG demurs and claims 'we already have a lot of ideas, let's first see what they are worth'.The secretary-general grins quietly and gives a thumbs up.But the Minister insists on more ideas (Fieldwork, July 2019).
. They conclude it is 50 million Euro's.The CFO urges the secretary to open negotiations with €60M, as this raises their chances of getting what they want.It sounds like negotiations 101, but the SG is not inclined and counters the CFO: 'If we need 50, we'll ask the Ministry of Finance for 50.She knows me.If that's what we need, we will get it.If not, we'll make it political', meaning he would ask the Minister to take it to Cabinet (Fieldwork, March 2018).