You ought to be careful. People will think you are … up to something.
Severus Snape in Harry Potter and the Philosopher’s Stone
Introduction
‘But we are the checks and balances?!’ (author’s interview with Hannah*)Footnote 1 one civil servant exclaimed, when I asked them about their role as defenders of democracy, surprised about my question and forceful about their position. Hannah* went on to explain how they defended democracy in the past and present. They gave examples of a civil servant’s everyday work: countering the political leadership’s proposals with practical considerations and advising them on better or alternative routes of action. But they also expressed their doubts about their role when a proposal seems not only impractical or infeasible, but also undemocratic. When this happens, there is no everyday response. Civil servants then enter a grey zone, between administrative neutrality and administrative activism, fraught with dilemmas. In this paper, I therefore ask and explore under what circumstances administrative democratic defence against incumbent-led autocratisation unfolds.
This paper explores how administrative democratic defence is a vital component of resilient democracies, considering that threats against democracy have increasingly come from inside the democratic system (Bermeo Reference Bermeo2016; Lührmann and Lindberg Reference Lührmann and Lindberg2019; van Lit and Van Ham 2025), being initiated by democratically elected incumbents, giving (apparent) legitimacy to the would-be autocrat. Furthermore, incumbent-led democratic recession is gradual and incremental (Diamond Reference Diamond2021; Maerz, Edgell, Wilson et al. Reference Maerz, Edgell, Wilson, Hellmeier and Lindberg2024), starting in the liberal components of democracy by eroding civil liberties and subverting checks and balances (Wunsch and Blanchard Reference Wunsch and Blanchard2023). Would-be autocrats’ democratic mandate and the incremental nature of contemporary autocratisation make it often hard to recognise autocratisation when it is in its early stages, especially when the would-be autocrat follows the ‘rules of the game’ to subsequently undermine the game. This makes the defence of democracy – actions taken to halt, revert, or slow down attempts to autocratise (Capoccia Reference Capoccia2005; Tomini, Gibril and Bochev Reference Tomini, Gibril and Bochev2023) – a more complex endeavour. In response to threats by incumbents, the institutional set-up of democracies is often not sufficient to counter incumbent-led autocratisation, and more importance is being placed on the actions of actors that stand up to defend democracy (Gamboa Reference Gamboa2022; Tomini, Gibril and Bochev Reference Tomini, Gibril and Bochev2023; van Lit, van Ham and Meijers Reference van Lit, van Ham and Meijers2023). Because the formal rules and regulations that should safeguard democracies are precisely the targets of incumbent autocratisers’ attacks, democratic defenders have to rely on political strategies, making the defence of democracy politically contentious, and these threats to democracy are not limited to newer or younger democracies (Svolik Reference Svolik2015). While it has been extensively studied in the paradigmatic cases of Hungary (Pirro and Stanley Reference Pirro and Stanley2022), Poland (Sadurski Reference Sadurski2018), or Venezuela (Gamboa Reference Gamboa2022), more recent scholarship has pointed to arguably consolidated democracies that appear under threat, such as the USA (Levitsky and Ziblatt Reference Levitsky and Ziblatt2018), Greece (Pappas Reference Pappas2020), and Israel (Gidron Reference Gidron2023). This indicates that threats to democracy (whether they are successful or not) can emerge in liberal, consolidated democracies as well.
Civil servants are part of this process: either as targets of autocratisation (Lotta et al. Reference Lotta, Tavares and Story2024; van Lit and Van Ham 2025) or as potential accomplices as they carry out policies that could result in autocratisation (Alon-Barkat, Gilad, Kosti et al. Reference Alon-Barkat, Gilad, Kosti and Shpaizman2025, p. 14). This suggests they can also play an important role in defending democracy.Footnote 2 However, we do not yet know under what circumstances civil servants make the decision to defend democracy, as they face a choice between administrative neutrality (the Weberian expectation that civil servants execute policies neutrally) and administrative activism (the Arendtian responsibility for a common good). Building on 29 elite interviews with Dutch national-level civil servants (Rijksambtenaren), I explore how they think about administrative democratic defence. Using an abductive approach, going back-and-forth between theoretical expectations beforehand and the collected interview data during and after the fieldwork, this paper is a first theory-building effort to understand the circumstances that contribute to administrative democratic defence.
Based on these interviews, I find that administrative democratic defence depends on three necessary requirements or circumstances: civil servants need to perceive they have the capability to recognise autocratisation, the access to legitimate instruments to halt or slow-down autocratisation, and the willingness to use those instruments. Only when each of these three requirements is met does administrative democratic defence occur.
Civil servants as democratic defenders
There is only limited research that recognises civil servants as potential democratic defenders (but see Bauer, Guy Peters, Pierre et al. Reference Bauer, Guy Peters, Pierre, Yesilkagit and Becker2021; Lotta, de Lima, Silveira et al. Reference Lotta, de Lima, Silveira, Fernandez, Pedote and Guaranha2023; Bauer Reference Bauer2024; Yesilkagit, Bauer, Guy Peters et al. Reference Yesilkagit, Bauer, Guy Peters and Pierre2024; Alon-Barkat, Gilad, Kosti et al. Reference Alon-Barkat, Gilad, Kosti and Shpaizman2025 and a brief mention of top civil servants in Tomini, Gibril and Bochev Reference Tomini, Gibril and Bochev2023 and van Lit, van Ham and Meijers Reference van Lit, van Ham and Meijers2023). Bauer (Reference Bauer2024, pp. 1106–1108) theorises that civil servants can ‘clash’ with would-be autocrats on four levels: (1) when would-be autocrats abuse their power outrightly, bypassing established procedures; (2) when would-be autocrats attempt to subordinate the bureaucracy to the primate of an illiberal majority, eliminating any administrative discretion; (3) when would-be autocrats attempt to push through policy changes that are at odds with standing policy; and (4) when the administrative workplace is changed as would-be autocrats appoint new leadership to upper and middle management (cf. Lotta et al. Reference Lotta, Tavares and Story2024). In each of these cases, the role of the civil service both as a policy-executing organisation and as a potential democratic defender is undermined. However, it is also clear in each case that the civil service has neither a monopoly on the truth nor on ‘the right course of action’. Lotta, de Lima, Silveira et al. (Reference Lotta, de Lima, Silveira, Fernandez, Pedote and Guaranha2023) demonstrate how civil servants and politicians legitimate their positions drawing on different legal frameworks (the law-vs-management dilemma). They find civil servants emphasise administrative autonomy and discretion, policy continuance and institutional memory, while politicians emphasise accountability of civil servants to politics and more centralised power (especially in contexts of democratic recession).
Despite these challenges, civil servants are prime candidates to be democratic defenders, as the civil service can be one of the first institutions confronted with autocratisation (Green Reference Green2019; Rockman Reference Rockman2019). As autocratising incumbents aim to keep a democratic façade (Mechkova, Lührmann and Lindberg Reference Mechkova, Lührmann and Lindberg2017), they often follow standard (albeit sometimes skewed, see Pirro and Stanley Reference Pirro and Stanley2022) procedure to implement policies. As such, the civil service is likely aware of autocratisation if it is proposed: it is asked for advice and subsequently asked to execute these policies. Especially since contemporary autocratisation is incremental (eg Maerz, Lührmann, Hellmeier et al. Reference Maerz, Lührmann, Hellmeier, Grahn and Lindberg2020), the role of civil servants to advise on (and potentially counter) each proposal that might harm democracy is pronounced. And just like other checks and balances, such as the courts and media, it is a target of autocratisation (Lotta et al. Reference Lotta, Tavares and Story2024; van Lit and Van Ham 2025). In other words, autocratisation occurs ‘in the surroundings’ of civil servants, either as potential accomplices or as potential targets.
Taken together, this highlights a central dilemma: are civil servants ‘serving the minister’ or ‘serving democracy’? And, more importantly, which allegiance prevails if civil servants perceive them to conflict? Yesilkagit, Bauer, Guy Peters et al. (Reference Yesilkagit, Bauer, Guy Peters and Pierre2024, p. 418) suggest that civil servants ‘have the responsibility to put guardianship in action […] as individuals and as a collective, [they] need to assume the role of the defender of the democratic order’. However, even if we accept this responsibility (which counters a political majority or plurality during incumbent-led autocratisation) as normatively acceptable, we still lack knowledge on how civil servants self-construct their role and its dilemmas: it is one thing to say that civil servants can defend democracy in theory; it is another thing to understand under what circumstances administrative democratic defence unfolds in practice. This leads to the central question in this study: under what circumstances does administrative democratic defence unfold?
Under what circumstances does administrative democratic defence unfold?
Democratic defence, in general, comprises all actions taken to halt, revert or slow-down attempts to autocratise (Capoccia Reference Capoccia2005; Tomini, Gibril and Bochev Reference Tomini, Gibril and Bochev2023). Administrative democratic defence (alternatively termed bureaucratic resistance in Bauer Reference Bauer2024; or guardianship in Yesilkagit, Bauer, Guy Peters et al. Reference Yesilkagit, Bauer, Guy Peters and Pierre2024) is those actions taken by civil servants. As with any democratic defence against proposals made by the governing incumbent, administrative democratic defence might be driven by (or perceived to be driven by) ulterior motives (Tomini, Gibril and Bochev Reference Tomini, Gibril and Bochev2023, p. 122; van Lit and Meijers Reference van Lit and Meijers2025). In fact, Alon-Barkat, Gilad, Kosti et al. (Reference Alon-Barkat, Gilad, Kosti and Shpaizman2025) show how the political ideology of civil servants influences which policies they see as threats to democracy and which as (potentially) legitimate changes. But, following Tomini, Gibril and Bochev (Reference Tomini, Gibril and Bochev2023), even when democratic defence is ideologically motivated, if the outcome is that attempts to autocratise are thwarted, it should be understood as democratic defence.Footnote 3 At the very least, however, the civil servants themselves should construct their opposition as a defence of democracy. Administrative democratic defence, in sum, is the actions taken by civil servants in their professional capacity that halt, revert, or slow-down proposals, and that are understood by civil servants themselves as protecting democracy.
Precisely because democratic defence against incumbent-led autocratisation goes against the (apparent) political will of the majority, administrative democratic defence is a process that is more complicated than ‘just sounding the alarm’. That is: in practice, the decision to defend democracy is hardly ever so clear-cut. I argue that for each individual instance of administrative democratic defence, there are three necessary circumstances that civil servants must meet: they need to have the capability to recognise autocratisation, access to legitimate instruments that halt it or slow it down, and willingness to act and use those instruments (see Figure 1). Other circumstances notwithstanding, these three circumstances emerged from the interviews as the most salient dimensions for civil servants themselves.Footnote 4
Figure 1. The three requirements for administrative democratic defence with the dilemmas civil servants face in resolving them. Only when all three requirements are met, does democratic defence occur.
The capability to recognise autocratisation
Before democratic defence can occur, a potential defender must have the capability to recognise that autocratisation is ongoing. Here, the basic question is as follows: can civil servants determine if a proposal has autocratic consequences?
The capability to recognise autocratisation entails both an actor’s awareness of the proposal, as well as the actor’s judgement that such a proposal is ‘wrong’ (Culiberg and Mihelič Reference Culiberg and Mihelič2017). Contemporary autocratisation is incremental, where single proposals are unlikely to lead to an undermining of a democratic system, but multiple proposals could. It therefore requires actors with extensive knowledge and awareness of the political and democratic system to recognise the threats of singular proposals. Political elites in general (cf. van Lit, van Ham and Meijers Reference van Lit, van Ham and Meijers2023), but civil servants especially, are likely to have this knowledge – even covert and incremental autocratisation. Tomini, Gibril and Bochev (Reference Tomini, Gibril and Bochev2023) and van Lit, van Ham and Meijers (Reference van Lit, van Ham and Meijers2023) suggest that civil servants understand the democratic system, their specific work topics, and how their craft was understood in previous cases to signal when proposals move away from accepted norms (cf. Wagenaar Reference Wagenaar2004; van Dorp Reference van Dorp2023). Where incumbent-led autocratisation is often characterised by its ambiguous nature (Bermeo Reference Bermeo2016) in which autocratic proposals follow the letter (and sometimes even the spirit) of the law (Pirro and Stanley Reference Pirro and Stanley2022; cf. van Lit and Van Ham 2025), civil servants are experts on their own topic of work – be it close to the ‘core of democracy’ (such as members of an electoral commission), or further away (such as civil servants working in a social welfare department). This topical expertise especially allows them to understand how proposals affect the system as a whole or parts of it, even without making a normative judgement about it. Put simply, civil servants should have the capability to look through the ambiguity of autocratisation. In fact, it is their job to recognise changes and advise on whether these changes are possible, probable, and permissible.
However, their recognition is challenged on two accounts. First, even if civil servants might recognise the potentially autocratic consequences of a single proposal, they might be hesitant to label it as such. Precisely because autocratisation is incremental, it is hard for a single civil servant to oversee the final outcome of the policy in interaction with other proposals. Second, ‘democracy’ itself (and conversely, autocratisation as a move away from democracy) is an ‘essentially contested concept’ (Gallie Reference Gallie1955; Held Reference Held2006). Even if civil servants might think a proposal is undemocratic, the fact that it is introduced and supported by the incumbent suggests that it carries at least some degree of democratic legitimacy, as it appears to reflect the will of the electoral majority or plurality that brought the incumbent to power. This tension makes it harder for civil servants to classify such proposals as autocratic: opposing them may feel like contesting the democratic mandate itself, even when the substance of the proposal undermines democratic standards.
Therefore, even if civil servants had access to legitimate instruments to halt autocratisation and were willing to use them, it is not a given that they recognise autocratisation when it occurs – despite their expert and topical knowledge.
The access to legitimate instruments to halt autocratisation
A democratic defence requires access to legitimate instruments that democratic defenders use to effectively defend democracy. The basic question is as follows: can civil servants do something about the proposal?
Courts, for example, can issue binding judgements (eg Zambrano, da Silva, Miron et al. Reference Zambrano, da Silva, Miron and Rodríguez2024) or citizens can demonstrate against autocrats or vote them out of office (eg Markowski Reference Markowski2024). But civil servants have different instruments they can access. Hirschman’s (Reference Hirschman1972) framework of exit and voice,Footnote 5 together with neglect (Farrell Reference Farrell1983; Golden Reference Golden1992) and subverting and obstructing internal procedures (Guedes-Neto and Guy Peters Reference Guedes-Neto, Guy Peters, Guy Peters, Pierre, Yesilkagit, Bauer and Becker2021; Schuster, Mikkelsen, Correa et al. Reference Schuster, Mikkelsen, Correa and Meyer-Sahling2022), provides a useful heuristic to understand which instruments are available to civil servants. Concretely, civil servants can issue critical advice about proposed policies; they can share confidential information to third parties in an effort to make (what they perceive as) wrongdoings public; they can delay policies by throwing up procedural hurdles; they can refuse to execute policies (especially if they are street-level bureaucrats, see Lipsky Reference Lipsky1980; Lotta et al. Reference Lotta, Piotrowska and Raaphorst2024); and they can quit their jobs.
However, which instrument to employ is again not an easy choice. Given that, by definition, democratic defenders aim to thwart autocratisation, they are likely to want to choose instruments they perceive to be effective (O’Sullivan and Ngau Reference O’Sullivan and Ngau2014). Yet not all the effective instruments are legitimate. For example, civil servants are observed to prefer voicing dissent internally, followed by a ‘silent exit’, rather than the active undermining of the executive (although it does occur, see Schuster et al. Reference Schuster, Mikkelsen, Correa and Meyer-Sahling2022). Kucinskas and Zylan (Reference Kucinskas and Zylan2023) suggest that this might be due to prevalent norms of administrative neutrality and obedience and loyalty to the political leadership. In addition, Gamboa (Reference Gamboa2022) explains that what she terms extra-institutional defence strategies might actually backfire, providing more legitimacy to the incumbent and undermining the democratic defender.
As such, even if civil servants are able to recognise autocratisation and willing to act on it, they might not end up defending democracy if they perceive effective instruments to be either inaccessible or illegitimate.
The willingness to act as a democratic defender
Finally, democratic defenders need to be willing to act and use the instruments they have access to. The basic question is fundamentally about civil servants’ personal courage, moral conviction, and the readiness to bear potential consequences.
On the one hand, the personal belief that it is (at least partially) up to civil servants to try and stop autocratic proposals should make civil servants more willing to defend democracy. Here, the democratic commitment of the potential defender as well as their conception of democracy (ie as either more majoritarian or more liberal) comes into play. For example, if a civil servant holds a more majoritarian conception, they are more likely to adhere to a Weberian role, making them less willing to use the instruments. On the one hand, civil servants, holding a more liberal democracy conception, are likely to value checks and balances in their own work as well.
On the other hand, professional risks, reputational costs, or threats to career progression should make civil servants less willing to defend democracy. Support from colleagues and managers (Akkerman, Sluiter and Manevska Reference Akkerman, Sluiter and Manevska2022; Demir, Reddick and Perlman Reference Demir, Reddick and Perlman2023) and an organisational culture that allows and encourages legitimate criticism (Goodsell, Reference Goodsell2011; Valentine and Godkin, Reference Valentine and Godkin2019) empower civil servants to defend democracy. In such cases, the professional risks are limited as the defence of democracy might be understood as a part of civil servants’ collective identity or professional duty (cf. Yesilkagit, Bauer, Guy Peters et al. Reference Yesilkagit, Bauer, Guy Peters and Pierre2024). But where such culture is lacking, or worse, where there is a real risk of repercussions, civil servants will be less likely to defend democracy. Gerschewski (Reference Gerschewski2013) describes repression as one of the core pillars of autocracies. However, repression does not only occur in full-fledged autocracies, but also in more covert forms in democracies, especially in contexts of democratic recession. In the US, for example, civil servants who were seen as too critical or disloyal were relegated to lower positions or dismissed (Kucinskas and Zylan Reference Kucinskas and Zylan2023).
Finally, therefore, even if civil servants recognise autocratisation and have access to legitimate instruments, they might not defend democracy: either for fear of repercussions, or because they see it not as their (professional) responsibility.
Empirical approach
To study the role of civil servants in the defence of democracy and their capability, instruments, and willingness, I conducted 29 in-depth semi-structured elite interviews (Lilleker Reference Lilleker2003; Littig Reference Littig, Bogner, Littig and Menz2009; Bogner, Littig and Menz et al. Reference Bogner, Littig and Menz2018) with civil servants at different career levels and across departments and agencies of the Dutch national government. The interviews were held in June to August 2024, and took between 45 minutes and 2 hours and 10 minutes each. The interview location was up to the participant, usually in their offices, meeting rooms, or the office-restaurant; five of the interviews were conducted online or elsewhere. One interview included two participants at the same time. I interviewed civil servants across 6 departments and 7 organisations that are organisationally part of a department but functionally independent. Of the participants, 16 hold non-managerial positions, 6 are middle-management, and 7 are upper-management. Participants contribute to different parts of the policy cycle, including research, consultation, legal affairs, and policy writing. As is common in qualitative research, the aim is not representativeness of the sample, but to capture the broadest diversity of experiences with autocratisation and democratic defence possible (Ritchie, Lewis, Elam et al. Reference Ritchie, Lewis, Elam, Tennant, Rahim, Ritchie, Lewis, McNaughton Nicholls and Ormston2014). I continued recruiting new participants until I reached theoretical saturation (Saunders, Sim, Kingstone et al. Reference Saunders, Sim, Kingstone, Baker, Waterfield, Bartlam, Burroughs and Jinks2018), evidenced by substantive overlap in participants’ ideas on and experiences with the role of civil servants in the defence of democracy.
Participants were contacted directly. The first few participants were selected based on their (publicly available) profile and area of work. Subsequent participants were selected using snowball sampling (Parker, Scott and Geddes Reference Parker, Scott and Geddes2019), specifically asking participants for references to colleagues who might disagree with them and who might think differently about their role in the defence of democracy. Snowball sampling likely results in a sample that is slightly biased towards civil servants who are willing or considering defending democracy, or have had discussions with colleagues about it, probably constituting a sample of the most likely administrative democratic defenders. Even so, as I show below, these most likely democratic defenders struggle to resolve the dilemmas, leading me to suspect they might be even harder for other civil servants. In other words, while this paper might overestimate the occurrence of democratic defence, it likely underestimates the salience of the dilemmas that civil servants face. Non-interviewed civil servants might err more on the side of caution (and ‘discomfort’, as outlined below) than the interviewed civil servants. In addition, this paper relies solely on interview data; further research is needed to triangulate its findings.
The interviews were recorded with the participants’ consent, and transcribed verbatim. The resulting quotes were translated by the author. The interviews were semi-structured, based on a general topic list (an introductory conversation about the participant’s work, followed by a more substantive conversation about their views on administrative neutrality, administrative activism, their democratic commitment, and their role in democratic resilience) and based on participant-specific examples and themes. Following an abductive approach, I analysed the interviews during and after the interview period, enabling me to ask later participants to reflect on comments by earlier participants. Additionally, this allowed me to discuss my interpretation with some participants to refine it as the interviews took place.
To ensure anonymity, I only refer to participants with pseudonyms that in no way reflect any participant characteristics (denoted with an asterisk, *) or as the general ‘civil servants in the Dutch national government’. I omit specific cases or workloads in the quotes to minimise participants’ risk of exposure and use square brackets to generalise sensitive examples.
All participants have been invited to read and reflect on the manuscript, but the interpretation ultimately remains my own. See the Online Appendix for the general interview set-up, the sensitising concepts used in the analysis of the transcripts, and reflections on reflexivity and positionality.
The Dutch case
I study civil servants in the national administration of the Netherlands. The Netherlands is a typical case (Gerring Reference Gerring2007; Seawright and Gerring Reference Seawright and Gerring2008) of liberal democracies: it has been a relatively stable liberal democracy since the end of the Second World War (Angiolillo, Lundstedt, Nord et al. Reference Angiolillo, Lundstedt, Nord and Lindberg2024; V-Dem 2024), and has not seen real threats to democracy similar to those in the paradigmatic cases of democratic recession. Yet, as in other liberal democracies, trust in the government has been declining over the past years, and there appears to be growing ambivalence towards democracy (Verwey-Jonker Instituut 2021; Voogd, Jacobs, Spierings et al. Reference Voogd, Jacobs, Spierings and Lubbers2023). In addition, Dutch democracy has seen some ‘stress tests’ where the normative guardrails (Levitsky and Ziblatt Reference Levitsky and Ziblatt2018) of democracy were challenged. One of these stress tests exemplifies the role of the civil service as a potential democratic defender.
During the Algemene Politieke Beschouwingen in 2024 (18 and 19 September 2024, the General Political Debates, the debate of the Dutch lower house (Tweede Kamer) at the start of the parliamentary year), the government proposed to use a state of emergency (staatsnoodrecht, Articles 110 and 111 of the Aliens Act) to solve asylum-problems in the Netherlands rather than submit emergency legislation. A state of emergency would (temporarily) sideline parliament, while emergency legislation would require parliament to pass new laws following regular but fast-tracked parliamentary procedures.Footnote 6 While it is by no means certain that rushed procedures or states of emergency would deterministically lead to autocratisation, the use of these legislative options in a situation that did not necessarily demand it can be considered a ‘stress test’ of democratic guardrails. Specifically, following Levitsky and Ziblatt (Reference Levitsky and Ziblatt2018), the governments’ proposals signal weak commitment to the democratic rules of the game, where alternative and less intrusive legislative options were still available. It has since become clear, however, that the Schoof-government is not going to call a state of emergency on asylum and is working on proposing measures following the standard procedure.Footnote 7
The role of civil servants during these Algemene Politieke Beschouwingen is noteworthy: the opposition (later joined by coalition party New Social Contract, Nieuw Sociaal Contract) pushed the government during the debate to publish civil servants’ internal advice about the proposed state of emergency. Then prime-minister and former top-civil servant Dick Schoof agreed to publish these internal documents, first partly redacted,Footnote 8 later in full.Footnote 9 In these documents, civil servants from different departments advised strongly that the state of emergency was ‘democratically and legallyFootnote 10 unacceptable’. Opposition parliamentarians used these administrative advices to accuse the government of turning into an ‘autocratic government that resists any form of opposition [literally: tegenmacht, counterpower]’.Footnote 11 In response, Geert Wilders, the leader of the PVV, the largest party in parliament, said he did not have ‘a backbone like a banana,’Footnote 12 insisting on the state of emergency. His co-partisan and Minister of asylum and migration, Marjolijn Faber, similarly said she prefers a state of emergency: ‘I value civil servants, but they have an advisory role. It’s up to the minister and the government to make the final decisions’.Footnote 13
These Algemene Politieke Beschouwingen, as a stress-test of Dutch democracy, show two aspects that are relevant to this study. First, even in a liberal democracy like the Netherlands, threats to democracy do occur, requiring constant democratic defence (which I will return to in the ‘Conclusion’ below). Second, and more salient, it shows that civil servants have taken a role as democratic defenders upon themselves. Importantly, civil servants’ arguments were used, almost verbatim, in parliamentary debates by the opposition (and later joined by coalition party NSC). As such, while autocratisation is comparatively unlikely in the Netherlands, there are at least ‘stress tests’ that require responses, in this case, by civil servants. The broader case of administrative democratic defence in the Netherlands is therefore a typical case of liberal democracies where challenges against democracy (and its resilience) exist.
In their own words: how civil servants navigate the dilemmas of administrative democratic defence
To show under what circumstances administrative democratic defence unfolds, I now turn to civil servants themselves. Through the interviews, it became clear that there is no single answer to the question of under what circumstances administrative democratic defence unfolds. The choice to defend democracy is ultimately one shaped by how individual civil servants perceive their capability to recognise autocratisation, their access to legitimate instruments, and their willingness to use those instruments. Other contextual dimensions notwithstanding, these three dimensions emerged as the most dominant. I therefore let civil servants formulate and navigate the dilemmas they face in their own words, deliberately including longer quotes to convey the full complexity and emotional weight of these dilemmas, as well as civil servants’ ongoing reflection on their position, knowledge, limited support, and moments of frustration.
It is important to note here that all examples and quotes refer to instances and examples that civil servants self-describe as autocratisation and administrative democratic defence. The examples given go beyond the asylum case discussed above, and they involve ministers, politicians, and policies during and before the 2023–2025 time period.
The dilemma around capability: expertise or endgame?
Whether civil servants think they have the capability to recognise autocratisation depends on the balance between their expertise and the endgame of policy-making. First of all, civil servants recognise they play a role in democratic resilience:
Well, we are not the fourth power,Footnote 14 but we are the experts. The expertise [literally: vakmanschap, craftwork] and knowledge on [democracy] resides with us – it should be with us. (author’s interview with Alicia*).
I believe the civil service definitely plays a role in [democratic resilience]. Simply by having the expertise to thoroughly understand your topic and present the pros and cons to a politician. Yes, that’s certainly a part of democratic resilience. (author’s interview with Susan*).
Alicia’s* and Susan’s* role-perception is rooted in their expertise, not in their personal opinion. To disentangle personal preferences from administrative duties and expertise, civil servants are trained to present both arguments in favour and against their own and the politicians’ positions. If a clash between civil servants and politicians does occur,
[civil servants] should always provide sound arguments, along with the right reasoning. It’s essential to know what you’re talking about. (author’s interview with Luna*)
But I can see how [an advice] might be perceived as lacking neutrality, especially if it contradicts political will. However, if the advice is grounded in professional knowledge, then that should be the guidance offered. I believe it’s crucial for the civil service to operate this way in a democracy. (author’s interview with Colin*)
These four quotes show how civil servants think in what way they should have the capability to recognise autocratisation: it should be guided by professional knowledge and a critical weighing of arguments in favour and against. Against this self-perceived expertise, however, civil servants struggle with what Fred* called the ‘endgame’. That is, within the full process of policy-making, a single civil servant is just a cog in the machine, unable to oversee all the consequences of a single policy:
Most individual civil servants cannot fully gauge the endgame, the consequences [of a policy]. You shouldn’t step into a role where you can’t oversee all the outcomes. Instead, use your position, knowledge, and expertise within the frameworks you have. (author’s interview with Fred*)
Applying this problematic endgame to administrative democratic defence, there are two dominant gaps in the ‘knowledge and framework’ that civil servants have. First, civil servants recognise that autocratisation occurs incrementally (cf. Maerz et al. Reference Maerz, Lührmann, Hellmeier, Grahn and Lindberg2020), making it even harder to gauge the consequences of a single policy. Zachary* illustrates how they realise years after the fact that they could have intervened earlier, and Dennis* explains it in more general terms:
In the end, after years, [the policy] culminated in a total clusterfuck [sic]. But when you look back at the timeline, all single steps seem logical – everything makes sense, until it went horribly wrong in the end. Then you reflect: ‘when should I have seen it coming?’ That’s difficult. Yes, in hindsight, I could say ‘that might have been a questionable aspect’. But did I really perceive it that way back then? When do you actually see it? [Here, Zachary* puts their head in their hands, falls silent for a moment, and says from behind their hands:] I don’t know. (author’s interview with Zachary*)
The issue with taking very small steps is that there’s always a justification for them, a reason why they’re not seen as anti-rule of law or anti-democratic. It often comes down to a weighing of interests. That’s the challenge: you can rarely assert with certainty that a proposal is so contrary to the rule of law or democracy that it should be outright rejected. (author’s interview with Dennis*)
Second, there is no shared understanding or professional framework of what ‘democracy under pressure’ means. Justin* calls this a ‘democratic criterion’, which is different from the ‘legal criteria which are a lot more rigid. The democratic criterion is… well it is more dependent on interpretation and weighing’. Ernie* pointed out that the word democracy occurs only once in the Dutch constitution, and in a vague and non-committal way at that. The full preamble of the DutchFootnote 15 constitution reads: ‘The Constitution protects the fundamental rights and the democratic rule of law’. Yet nowhere is the ‘democratic’ in ‘democratic rule of law’ further defined. In other words, there are only limited professional guidelines or shared frameworks to formulate an argument that a proposal harms democracy. This makes administrative democratic defence different from opposition to unconstitutional proposals or advice against controversial policies, where more shared, professional guidelines and ‘rigid criteria’ exist. As such, civil servants note:
We’ve made these [democratic] dilemmas very personal. There aren’t any shared frameworks to reference, which makes it easy for people to criticise you for basing your advice on personal opinions. (author’s interview with Seamus*)
What does [saying democracy is under pressure] actually mean? [Democracy] is such a big concept… what do you really mean when you say [it is under pressure]? It’s difficult to articulate. It raises questions like: what does that really mean? What can we do about it? It doesn’t relate much to your day-to-day responsibilities. (author’s interview with Colin*)
In sum, when it comes to the first necessary circumstance for administrative democratic defence – do civil servants have the capability to recognise autocratisation? – the answer is ‘yes, but…’. Yes, civil servants believe they have (and should have) the expertise and knowledge to recognise autocratisation, not only if they are closer to the core of democracy, but also if their topic of expertise is further removed from it. Their craftwork and topical knowledge should allow them to understand if and when procedural rules or ‘soft guardrails of democracy’ (Levitsky and Ziblatt, Reference Levitsky and Ziblatt2018) are tarnished. But they struggle with the endgame of policy-making, the idea that they are not likely to be able to oversee all outcomes of a policy proposal, on two counts: autocratisation is so incremental that individual proposals are easily legitimated or overlooked and the lack of a ‘rigid democratic criterion’ (as Justin* puts it). In practice, this means that civil servants are sometimes hesitant to speak out in time against proposals that, after the fact, clearly harmed democracy (in Zachary’s* case) or that are legitimised in some way (as Dennis* puts it). And even if civil servants think they recognise autocratisation, they have a difficult time formulating that conclusion based on their professional knowledge (as Colin* puts it) or ‘the right reasoning’ (as Luna* puts it).
The dilemma around legitimate instruments: internal or external?
Civil servants face a further dilemma about which instruments to use. The clearest distinction that emerged from the interviews is the distinction between the internal voicing of dissent (such as voicing your concerns to your supervisors) or the external flagging of issues (such as leaking to the media, striking, or issuing public statements).
For all the civil servants I spoke to, the instrument they perceive to be most legitimate is always the internal voicing of dissent. This entails their formal advice to politicians, but also informal discussions with colleagues and managers. Lee* exemplifies this use of internal instruments. They encountered a proposal
that was simply not possible – constitutionally speaking. It truly was not permissible. However, it had been done in other departments: it was possible, but it really was not permissible. I consulted colleagues who are well-versed in constitutional matters, so my arguments were sound. In every meeting with the minister, I kept reiterating, ‘this is not permissible for all these reasons’. But the questions kept coming. What helped me significantly was when I called the assistant to the minister and said, ‘what the minister wants is simply unconstitutional’. That’s when it clicked, that they really shouldn’t be doing it. In subsequent meetings with the minister, the same question arose, and I wrote my advice down in response. I articulated, in very strong terms, why it was not possible. Then it finally clicked for everyone. (author’s interview with Lee*)
Lee* was ‘proud’ of the way they handled this, and colleagues at the other departments thought it was ‘good, very good’ what Lee* did. Looking through the steps they took, they are committed to the internal instruments, even though it is clear that the policy was fundamentally wrong to Lee*: seeking strong legal argumentation from colleagues, consulting with their supervisor, and answering questions in meetings. However, in the first instance, the minister kept pushing to implement the policy because it had been done in other departments. Even when it apparently ‘clicked’ for the assistant to the minister, yet another meeting was needed, and only when the advice was written down did it result in the proposal being withdrawn. It is clear Lee* followed procedures they thought were legitimate. At the same time, these instruments were highly inefficient (and barely effective).
In some interviews, I asked participants explicitly to reflect on the choice between internal and external instruments. In response, Michael* said:
You sometimes start to wonder – should you drop a hint somewhere? Maybe to someone in the media… Those thoughts do come up. I’m very much against sharing confidential information, but when you see miserable decision-making that the media isn’t picking up on, I think it wouldn’t be all that surprising if a civil servant had coffee with a journalist. I wouldn’t find it strange, even if it’s just to say, ‘You might want to dig into this [topic], because something is off’. I’m not saying I would do it myself, but… (author’s interview with Michael*)
For Michael* the external instruments are not necessarily legitimate options in the day-to-day toolkit of administrative democratic defenders. Instead, ‘having a coffee with a journalist’ and ‘dropping a hint somewhere’ should only occur when there is ‘miserable decision-making’ that might otherwise go unnoticed. That is, in the absence of effective democratic defence through legitimate instruments, some illegitimate instruments could be considered. Other civil servants, however, think that civil servants should engage in the public debate about democracy, almost as standard practice:
Well, we engage with the media more frequently these days. If you typically only approach the media when [the politicians] aren’t listening, it quickly earns you the label of ‘activist group’. That’s why we now go to the media more often, not just when [the politicians] fail to listen. (author’s interview with George*)
I think [being scared to openly talk about our advices] creates an unnecessary tension. We should be able to speak frankly and freely about [our advice about democracy]; there’s nothing strange or awkward about it. Honestly, I don’t understand it. As civil servants, it is what we should be thinking about – it’s our right to exist. (author’s interview with Luna*)
In contrast, Justin* draws a clear line where civil servants’ responsibility ends:
In the end, the decision isn’t yours. You can be very firm, saying repeatedly, ‘don’t do it, don’t do it’ But ultimately, if the decision-makers say, ‘we’re going ahead,’ then as a civil servant, it’s not your place to decide what is allowed and what is not. (author’s interview with Justin*)
As such, Justin* believes civil servants should not voice their critiques externally,
with [their] government ID around [their] neck. I would make a clear distinction there. Otherwise, it could seem like I’m using my position to settle personal scores. (author’s interview with Justin*)
These quotes highlight how civil servants struggle with the legitimacy of different instruments, from internal advice to cautious external signalling. Yet, all of these options still assume that civil servants remain at their jobs, attempting to defend democracy from within. This raises the question: what happens when civil servants feel that neither internal nor external instruments are efficient instruments to defend democracy? At that point, the most drastic instrument available is resignation (cf. Hirschman Reference Hirschman1972 on voice and exit). Are civil servants willing to quit their job in defence of democracy? Hannah*, Susan*, George*, and Ernie* have different perspectives on this:
At the same time, when you leave [your job], you lose all your influence. I think I’ll keep asking questions – I hope I will. But I imagine I would stay at first. (author’s interview with Hannah*)
We’re in a position where we can still exert some influence, because politicians can be open to certain arguments. I also think it’s important to stay and show that. The danger is, if everyone leaves and new people come in who are happy to work for [a minister], things could end up even worse. So, part of the reason for staying is to keep things in check. The civil service remains, while politicians come and go. (author’s interview with Susan*)
It’s true that up to a certain point, you can say [some policies are] not your responsibility. But eventually, you have to ask yourself, ‘has this gone too far? Do I still want to be doing [this job]? How do I explain my actions to my children?’ That’s a personal choice. For many civil servants, we are not there yet, but when we get there, you need to know where your own moral line is. (author’s interview with George*)
If we keep acquiescing, keep going along, might people eventually ask in fifteen years, ‘Who were those bystanders?’ And that’s what I’m afraid of. (author’s interview with Ernie*)
Hannah* emphasises the power of ‘asking questions’ and Susan* expects that civil servants can convince or outlast ministers who make undemocratic proposals. George* and Ernie*, on the other hand, appear to want to draw a ‘moral line’ (as George* puts it) somewhere. They do not want to be a ‘bystander’ (as Ernie* puts it) or complicit in autocratisation through acquiescence (cf. Alon-Barkat, Gilad, Kosti et al. Reference Alon-Barkat, Gilad, Kosti and Shpaizman2025, p. 14).
This discussion around the second necessary condition for administrative democratic defence – do civil servants have access to instruments to counter autocratisation? – highlights how important it is for civil servants that the instruments they access are both legitimate and effective. Perceived legitimacy is highly personal. There appears to be a broad consensus that ‘loyal contradiction’ or ‘loyal criticism’ (van der Meer and Dijkstra Reference van der Meer and Dijkstra2021) is legitimate as long as it remains internal, but it becomes more contentious when civil servants speak out publicly. The ultimate external instrument, quitting your job with a ‘noisy exit’ (Kucinskas and Zylan Reference Kucinskas and Zylan2023, p. 1770), does not often appeal to civil servants. Some explain that they want to remain in a position ‘where they can still exert some influence’ (as Susan* puts it). Even Lee*, for whom the proposal ‘really was not permissible’ on a fundamental level, admitted in the end they would probably have asked for a different case-load but not quit their job. There are grey areas around the legitimacy of available instruments, putting the burden on individual civil servants to make such personal and moral choices.
The dilemma around willingness: discomfort or audacity?
This personal burden manifests itself most clearly in the final dilemma civil servants face: given everything they know and given the instruments at their disposal, do they want to defend democracy – potentially at the risk of personal and professional costs? The first obvious costs are disciplinary or criminal sanctions when they choose an instrument, such as leaking, that is officially forbidden. But also, when they want to choose legitimate instruments, civil servants face risks that might make them reconsider their democratic defence. Both Hannah* and Zachary* are acutely aware of these risks to civil servants who appear too critical (not just those who overstep formal rules and regulations):
At a certain point, I have to be honest: my mortgage needs to be paid. It will affect my home life. The protection offered isn’t sufficient if you truly want to speak out. (author’s interview with Hannah*)
Look, I can confidently say that I would definitely [speak out]. But I also have a mortgage and children in university, you know? It’s easy to be outspoken, but you have to wait and see how things actually play out. (author’s interview with Zachary*)
The final dilemma around willingness, therefore, manifests itself most clearly in the contrast between administrative discomfort and administrative audacity. Administrative discomfort is the realisation that it is not up to the civil servant to make the final decision. Justin* and George* illustrated this already, but Fred* puts it most clearly:
There are many moments of discomfort. Yes, decisions are being made (including ones that you’ve been working on yourself) that you believe are, quite frankly, bad policies. However, at a certain point, there’s nothing more you can do about it. So, you have to learn to let it go. (author’s interview with Fred*)
Fred* suggests that civil servants should suffer this discomfort, ‘let [their concerns] go’, and accept that ‘bad policies’ are implemented against their own preference and advice. Dean* on the other hand, explains that civil servants should be steadfast. Dean* calls this audacity: going against (explicit) political will to draw a line in the sand:
Let me share a personal example. [A minister] once said to me, ‘I want [example]’. I replied, ‘Well, that’s your responsibility. You’re the one who has to sign off on [the policy], not me’. However, they then devised a ruse to get me involved. After considering it, I decided, ‘I’m not going to do that. I stand by my work and my advice’. The minister responded, ‘But I’m your boss, so you will do [this]’. I said, ‘That’s not how the world works’. The minister remarked, in turn, that they were going to raise [my behaviour] at the ministerial council. Later, the phone rang. It was the minister saying, ‘I spoke with my colleagues, and they all think you need to [do it]’. I replied, ‘In that case, I’ll come in tomorrow to resign, because I won’t do it’. After the discussion in the ministerial council, the minister dropped the case. In [another situation], I texted [my supervisor], asking if I was too inflexible [towards a minister]. The reply was, ‘the minister thought you had some audacity’. That made me realise how civil servants can face pressure. But at some point, you have to draw a line. Apparently, they respected my stance because they never followed up with me. But it was challenging. From my experience, you can earn respect as a civil servant if you occasionally take a stand. (author’s interview with Dean*)
Dean* (and Lee* before) clearly recognised policies they thought needed to be stopped – not because of their personal policy preference, but because of core, fundamental considerations about democracy and the rule of law. Dean* was willing to quit their job when forced to, but as long as they had their position, they would not sign off on a proposal that harmed democracy. Yet, it takes considerable audacity to speak out against the political leadership. In Dean’s* case, the minister pulled rank (‘I am your boss’) and personally called Dean* in an effort to pressure them again after a discussion with the entire ministerial council.
Many civil servants recognise they have (theoretical) space for internal ‘loyal criticism’, but
there’s a significant difference between having the space to criticise and actually taking that space. I don’t think any civil servant would deny that the opportunities to criticise exist because they certainly do. However, it takes a strong person who is very sure of themselves to seize that opportunity in the moment. (author’s interview with Angelina*)
What Angelina* points to, and Terry* and Luna* make more explicit, is the risk of self-censorship.
We handle a lot of dossiers, so we must constantly choose where to invest our energy and resources. That capacity is always a factor in your judgment, and it carries the risk of using it as a reason to overlook certain, more politically complicated dossiers. I do realize that is a risk. (author’s interview with Terry*)
I can assure you that discussing these challenging topics is far from guaranteed within organisations. Civil servants are incredibly busy, often facing tight deadlines. They simply don’t have the time to engage in discussions [about democracy]. Unfortunately. (author’s interview with Luna*)
Whether it is a conscious choice to not engage in ‘politically complicated dossiers’ (as Terry* puts it) or an unintended consequence of limited capacity or high workload, both hinder the audacity that is needed for administrative democratic defence.
Finally, many civil servants explain how support among colleagues can strengthen their willingness. Lee* and Dean* relied on their managers for backing. Angelina* connects it to the problems of recognition, and Fred* links it to a broader category of democratic defenders:
It has to be a collective effort. Truly getting to the bottom of things, as you’d like to, and understanding all the intricate details, is unfeasible. The entire process is almost beyond comprehension, so you can’t tackle it alone. (author’s interview with Angelina*)
Ultimately, it’s all about interplay. Democracy isn’t solely about the minister and parliament; it also involves the courts, the Ombudsman, and various other roles and responsibilities. (author’s interview with Fred*)
Taking all this together, I argue that the dilemma around willingness is the hardest to resolve. Civil servants’ willingness to defend democracy is not simply a matter of professional knowledge (capability to recognise), nor of institutional rules (access to instruments) – although those dimensions can benefit and hinder administrative democratic defence. Often, the choice comes down to a single (group of) civil servants under conditions of uncertainty and risk. Some civil servants, like Fred*, accept discomfort as part of the role; others, like Dean* and Lee*, demonstrate audacity by drawing firm lines against political pressure. Most, however, operate in the grey zone where silence, self-censorship, and pragmatic compromise can feel like the safer path. Whether willingness translates into democratic defence, therefore, depends less on the formal existence of space for criticism than on whether civil servants are willing to seize it.
There are some guidelines about capability and legitimate instruments: a government programme on ‘Ethics and Dialogue’;Footnote 16 a book authored by a civil servant and published by the government titled ‘Power and Courage’;Footnote 17 and new standards that civil servants should be ‘test[ed …] especially for awareness of their functioning under the rule of law’.Footnote 18 This should help civil servants make the distinction between activism and craftwork and enable civil servants to organise some form of collective counterpower when fundamental issues are at stake. However, these programmes do not give sufficient guidelines for civil servants ‘in the moment’, especially when autocratisation occurs incrementally (as Zachary* describes), or is justified by some weighing of interests (as Dennis* describes). When administrative democratic defence is audacious, rather than standard and expected, when there is a real risk of repression, and when it is a comfortable way out to self-censor, it becomes increasingly hard to keep recognising autocratisation. The willingness-requirement, therefore, likely determines more than capability and instruments whether administrative democratic defence takes place. There are some benchmarks to guide a civil servant’s perception of their capability and instruments, but hardly any, except their personal motivation and dedication, about willingness.
Discussion and theoretical implications
The theoretical and empirical discussions above show the importance of capability, instruments, and willingness. The abductive approach in this paper also allows for further theorisation to better understand (administrative) democratic defence. Three tensions require resolving, each with implications for the study of autocratisation and democratic defence:
First, I have presented the three dilemmas in relative independence of each other. However, especially in the discussion around willingness, it became clear that the dilemmas are at least partially interdependent. Further research is needed to establish this interdependence, but the evidence suggests that the factors driving administrative democratic defence (expertise, access to legitimate instruments, audacity) are mutually reinforcing, as are the factors that hinder it (endgame, limited access, discomfort). For instance, a civil servant confident in their expertise and aware of legitimate instruments might be more likely to develop the audacity required to defend democracy, while the absence of these factors can create a downward spiral in which discomfort feeds on lack of capability and inadequate instruments.
The necessity of the three requirements also implies that autocratisation (or at least its potential) might start even earlier than we typically study, as I expect that would-be autocrats actively undermine the potential for democratic defence by targeting one or more of the three requirements. The repression of dissent and critique is a well-established hallmark of autocratisation (Gerschewski Reference Gerschewski2013; van Lit, van Ham and Meijers Reference van Lit, van Ham and Meijers2023), affecting defenders’ willingness. However, would-be autocrats might also erode defenders’ capability and access to instruments, for instance by curtailing training programmes or by restricting channels for ‘voice’ (Hirschman Reference Hirschman1972) and dissent. In this way, would-be autocrats can pre-emptively disable democratic defence through seemingly technical or procedural changes that are hard to understand as autocratisation. This suggests that autocratisation – and the opportunities for defence! – may begin well before overt repression occurs.
Second, much of the literature focuses on the macro-level structures that enable and hinder autocratisation, which is often termed democratic resilience: the ability of a system to ‘bounce back’ from autocratisation (Merkel and Lührmann Reference Merkel and Lührmann2021). While such work generates theoretical expectations about the actions of agents working within institutional structures, the interviews show there are large grey areas that institutional explanations cannot easily resolve. As Angelina* reflected, even when formal space exists for criticism, civil servants may not be audacious enough to use it. Similar grey areas exist regarding which instruments are legitimate and how the ‘democratic criterion’ should be applied. Hence, we need to understand the occurrence of elite democratic defence on a micro-level, asking what motivates and what stops specific actors from acting – thanks to or despite structural constraints. The three requirements of capability, instruments, and willingness provide a first step in that direction.
The theoretical implication is that these requirements also explain why democratic defence sometimes fails to occur, even when autocratisation is ongoing:Footnote 19 each requirement can become a bottleneck. A comparative study of these characteristics across cases of incumbent-led democratic recession could reveal how the sequencing of failures in capability, instruments, or willingness contributes to the broader process of autocratisation itself.
Third, while it is outside the scope of this paper to fully develop a normative argument about the desirability of administrative democratic defence (but see Bauer Reference Bauer2024; Guy Peters and Pierre Reference Guy Peters and Pierre2022; Yesilkagit, Bauer, Guy Peters et al. Reference Yesilkagit, Bauer, Guy Peters and Pierre2024), there are clear normative trade-offs. As Justin’s* notes, the primacy and responsibility of politicians over civil servants must be respected. Moreover, administrative democratic defence carries risks of policy freeze and overreach if the civil service obstructs political proposals too often, as well as risks of ideological bias within the bureaucracy. Yet these risks must be balanced against the dangers of inaction in the face of democratic erosion. Alicia* captured these tensions well:
There’s a reflex that has persisted for too long: ‘democracy belongs to someone else. Everything is fine in the Netherlands, and if it isn’t, we can rely on Strasbourg or Brussels’. However, everyone has a responsibility [for democracy]. I also believe that judges and the media play significant roles in this. As civil servants, we have a duty to keep the media informed and up to date. But, ultimately, it’s the political parties that serve as the vehicle to bring these issues into the debates. And this is something we consider [in the department] as well. Parliament needs to be positioned to engage [with democratic resilience] effectively. (author’s interview with Alicia*)
This suggests that other democratic defenders face similar dilemmas of capability, instruments, and willingness, though they likely resolve them differently: courts likely have more effective instruments, independent watchdogs likely face less discomfort, and politicians are likely to struggle less with the ‘democratic criterion’ due to their electoral mandate. The role of civil servants in this interplay of actors (cf. Tomini, Gibril and Bochev Reference Tomini, Gibril and Bochev2023) depends thus on the normative view: it might be to actively defend democracy, or it might be to play an important information-providing role (to a plethora of other actors, including the incumbent would-be autocrat).
On the reverse side, the framework points to several ‘quick wins’ for democratic defence. Continued training on the ‘democratic criterion’ and, more specifically, an awareness of how small, incremental actions can constitute autocratisation would bolster defenders’ capability: if not for a full-fledged democratic defence, then at least for close scrutiny and early warning. Institutionalised instruments dedicated specifically to the defence of democracy (rather than generic mechanisms for critique or administrative advice) could move democratic defence away from an ad hoc, personalised act toward a more routine practice. While concerns about overreach and policy freeze remain, in light of incumbent-led democratic recession, a sustained societal and political debate about the potential autocratic consequences of proposed policies – based on the expertise of a variety of actors – is both necessary and likely to strengthen defenders’ willingness.
Conclusion: serving the minister or serving democracy?
In this paper, I sought to explain under what conditions administrative democratic defence – actions undertaken by civil servants to stop, slow down, or halt proposals that harm democracy – unfolds in a liberal democracy. Based on an abductive analysis of 29 elite interviews with Dutch national-level civil servants, I find there are three dilemmas that need to be resolved before civil servants engage in democratic defence. First, civil servants talk about their capability to recognise autocratisation. They feel they should have the expertise to do so, yet they struggle with the endgame of policy-making, in which they can neither always see the outcome of individual policies during incremental autocratisation nor judge them adequately along a ‘democratic criterion’. Second, civil servants talk about their access to legitimate instruments that might halt autocratisation. On the one hand, they know they can legitimately voice a critique internally in meetings with colleagues, supervisors, or the political leadership. On the other hand, internal instruments are not always effective, making some civil servants consider external instruments such as talking to the media. Third, civil servants talk about their willingness to defend democracy. On the one hand, they know it is part of their job to suffer some discomfort and accept policies that go against their expert advice. On the other hand, they sometimes want to draw a line in the sand and be audacious in the face of potential personal and professional risks. Overall, these three dilemmas show that administrative democratic defence, while straightforward in theory, is surrounded by uncertainties in practice. More importantly, these uncertainties appear to be down to individual civil servants to solve, in the absence of clear institutional rules, expectations, a ‘democratic criterion’, and guidelines as hypothesised in the ‘Discussion’. This is most clear in the dilemma around willingness, where civil servants are faced with political pressure from not just one minister (in the case of Lee*) but the entire ministerial council (in the case of Dean*).
The case of the Netherlands, as a typical liberal democracy, shows two additional things. One: even in a liberal democracy, politicians ask civil servants to implement proposals that are deemed undemocratic. This implies two: democratic defence itself should be an ongoing process – even in liberal democracies. Especially since autocratisation occurs incrementally (Maerz et al. Reference Maerz, Lührmann, Hellmeier, Grahn and Lindberg2020), even seemingly innocuous threats should be noted, if not explicitly counted. This is where civil servants come in, as they are likely to be part of the incremental process: either as targets of (potential) autocratisation or as accomplices. The reason Dutch democracy is resilient until now is not just because there are strong institutions: autocratisation attempts occur within those institutions regardless. It is the actors who hold positions in those institutions, the ‘passionate defenders of democracy’ (author’s interview with Justin*), the ‘guardians of the rule of law’ (author’s interview with Luna*), and the ‘guardians of democracy’ (author’s interview with Katie*) who make liberal democracy resilient. However, the Dutch civil service is not necessarily a typical civil service: it lacks political appointments (except at the top level; van Dorp and ’t Hart Reference van Dorp and ’t Hart2019), making it less susceptible to politicisation. The United States’ (Guy Peters and Pierre Reference Guy Peters and Pierre2004) and Hungary’s (Meyer-Sahling Reference Meyer-Sahling2008) civil service are more politicised – and likely even more so since 2011 (Levitsky and Ziblatt Reference Levitsky and Ziblatt2024) – whereas Belgium’s (Dierickx Reference Dierickx2004) and the United Kingdom’s (Sausman and Locke Reference Sausman and Locke2004) appear less so. The level of politicisation thus, might be an important contextual circumstance factor shaping administrative democratic defence (Guy Peters and Pierre Reference Guy Peters and Pierre2022). I suspect that it affects primarily the willingness of civil servants to defend democracy, as speaking out becomes more audacious in politicised contexts. The findings in this paper are therefore most likely to travel to contexts where politicisation of the civil service is minimal. In contexts where the civil service has already been co-opted, or autocratisation is more advanced, legitimate instruments are typically more restricted, and repression more common. The framework of capability, instruments, and willingness suggests that administrative democratic defence becomes progressively harder as autocratisation deepens and the dilemmas become more difficult to resolve. In sum, I expect that the three dilemmas travel across contexts, though not necessarily in the ways in which they are resolved.
Finally, of course, civil servants are not the only defenders of democracy. As explored in the ‘Discussion’ above, there are many other potential democratic defenders, many of whom, such as judges, may be more normatively acceptable than civil servants. Others, such as elites within the would-be autocrat’s party, might be more effective. However, this study of administrative democratic defence shows that even if other democratic defenders should be able to defend democracy, it is not a given that they actually do, as they likely still face risks and uncertainties. Nonetheless, my focus on civil servants might overemphasise the relevance of some dilemmas compared to other actors. Since elites play a pivotal role in defending democracy against incumbent-led autocratisation (cf. van Lit, van Ham and Meijers Reference van Lit, van Ham and Meijers2023), future research should look into how other actors perceive their role and shape their responses.
Importantly, the difficulties of democratic defence do in no way preclude it from occurring. I have shown examples when civil servants do defend democracy, even with ‘clammy hands’ (author’s interview with Alicia*). And it is likely multiple more defences have occurred, outside the view of the public eye and outside the purview of my interviews. However, it is also likely that some autocratisation attempts have not been recognised or not been acted against, because the dilemmas at that point in time were resolved differently. That is why the interplay of actors, each with different expertise, different instruments, and different moral grounding, is necessary in the defence of democracy: to overcome any gaps that a single defender might leave. As such, democratic defence hinges on multiple actors, to be aware and signal whether an incumbent is up to something.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1475676526100851
Data availability statement
Neither interview recordings nor the transcripts will be made publicly available due to the sensitive nature of the contents. It proved impossible to sufficiently pseudonymise the transcripts. The interview recordings and the transcripts will be saved on secure servers at Radboud University for 10 years to ensure scientific integrity.
Acknowledgements
The author thanks participants of the Hot Spot Sustainable Democracy at Radboud University (March 21, 2024), of the workshop ‘Sounding the Alarm’ at the 2023 ECPR Joint Sessions of Workshops in Lüneburg (March 25 to March 28, 2024), of RMRD 3 at Antwerp University (June 17, 2024), of the workshop ‘Resilience against Extremism’ at the Vrije Universiteit Amsterdam (September 27, 2024), and of the PSC Feedback Workshop (November 11, 2024), as well as Carolien van Ham, and Maurits Meijers and the three reviewers for their helpful comments and valuable feedback. In addition, the author is grateful to the participants in this study.
Financial support
Open access funding provided by Radboud University Nijmegen.
Competing interests
The author declares no conflict of interest.
Ethical standards
The interview set-up has been approved by the Ethics Assessment Committee Faculty of Law and Nijmegen School of Management of Radboud University, under number EACLM-LT-037.
Open access
