Could the Civil Service Restrain a Populist, Anti-Democratic Government?

I recently recommended Ben Yong’s 16 October UCL lecture about the civil service’s constitutional foundations and how it might respond to a populist government¹.

The video is now up on YouTube. Here is a slightly shortened and lightly edited transcript. It is fascinating and worrying reading.

---

We face a moment where faith in democracy and established institutions seems to be at a particularly low ebb.

Democratic backsliding - the weakening or evisceration of institutions designed to sustain democracy - seems to be on the rise.

And intermediary institutions like the bureaucracy are seen as obstructing populist politicians who portray themselves as the direct representatives of the people.

So here’s my argument:

The doctrine of ministerial responsibility is the dominant constitutional principle underpinning the work of the civil service. It appeals to public lawyers, it appeals to politicians and the public, because it appears to reconcile the role of the civil service within the traditional framework of the separation of powers and parliamentary sovereignty; and it follows that the primary duty of civil servants is to support the incumbent government.

But ministerial responsibility leaves little or no space for the possibility that the civil service might have duties to the continuity of the state which are separate from its duty to support incumbent governments. And that, coupled with the deliberate ambiguity of the law of the civil service, seems to me to leave a civil service vulnerable to a government intent on democratic backsliding.

Francis Fukuyama has talked about the tricky political question of bureaucratic autonomy in situations where the political principal is manifestly incompetent, or corrupt, or is acting in something other than the public interest: democratic backsliding. This includes the elimination of institutions which sustain an existing democracy, it involves unprincipled political principals:- politicians deliberately undermining key aspects of liberal democracy including the electoral process and established institutions that serve to check or at least stabilise executive power.

Democratic backsliding is particularly pernicious because it is legitimated by the very institutions that democracy promoters have prioritised:- elections, majority voting through legislatures, and the laws that are passed by majorities.

And backsliding is often linked to populism which divides society into a pure people and a corrupt elite. Populist governments often push to undermine and remove institutions which frustrate “the will of the people”. Those intermediary institutions which check or at least stabilise executive power. And one of those institutions is the bureaucracy.

The most stark and disquieting example of backsliding is happening right now in the US. Their changes have served to foreshadow changes in other so-called established western democracies. And their attacks on key institutions have advanced with quite alarming rapidity and what’s been striking in both the first and the second Trump administrations have been the assaults on the federal civil service.

But there have also been some concerns about the direction of travel in the UK with various institutions subtly and not so subtly undermined - and attacked in recent years. These include attacks on judges - “enemies of the people” - after the Miller 1 judgement. Just a couple of weeks ago there was an opposition politician who called for root and branch changes to the appointment and removal procedures for judges that make bad decisions about immigration. There was the attempt to prorogue parliament, changes to the Election Commission, and sustained criticism of the civil service either because it’s been seen as obstructionist or as ineffective in delivering government policy - or both. It has seen as part of ‘the Blob’, the so-called deep state (what we used to call just ‘the state’!) and it’s set against democratically elected politicians.

Now we don’t have to take a firm position on whether or not democratic backsliding is happening in the UK but we can note that it is happening elsewhere and it might be worth us taking a kind of stock take. How would the civil service respond to a populist government intent on democratic backsliding?

Some scholars now portray the bureaucracy as a check on the passions or the whims of charismatic politicians. Some suggest that the bureaucracy should not simply resist unprincipled politicians but actually under certain circumstances actively protect democracy. Are there resources within public law that civil servants might draw upon?

One problem is that the civil service has been ignored in public law. This is because of the doctrine that elected officials politicians decide and unelected officials just advise and implement: ministerial responsibility. It’s a kind of transmission belt theory and on this view the civil service is merely the agent of a set of political principals:- We elect politicians. Politicians decide; officials simply implement or obey.

This is a neat chain of delegation. Ministerial responsibility then serves as the kind of description of practice but also as a normative justification for how the bureaucracy is controlled.

But the surface simplicity of ministerial responsibility really hides the complexity of what the civil service is and what it does. The idea that civil servants have no discretion or role in the policy process is just unrealistic

A ministerial decision, if there is one, is just one stage amongst many in the policy making process and at almost every stage of this process officials are exercising discretion. Instructions from ministers, for instance, are rarely specific enough to guide officials. There can be cross cutting issues which mean discussion between departments - and the detail isn’t just detail. It can determine the entire policy and how it’s perceived and how it works in practice.

Ministerial responsibility is a bit like the civil service analogue of parliamentary sovereignty in UK constitutional law. Once parliament has spoken that’s the end of the matter. And, similarly, the democratic credentials of ministers are such that the determination of the public interest is decisive. The civil service can’t look outside the incumbent government to seek an independent understanding of the public interest. Their only role is to implement the will of ministers.

And the civil service is not really thought of these days as having expertise as well. There’s a lot of outsourcing going on. And it’s often criticised for its sluggishness and its insistence on legality know that stopping and obstructing policy from happening

So does the internal law of the civil service offer anything which might help?

Well, it reflects its constitutional underpinnings. It’s driven by an insistence upon responsiveness to ministers. Examined more closely though, this internal law seems to have various qualifications and ambiguities.

The current law of the civil service is a mixture of statute, convention, prerogative, codes, and internal guidance. I’ll just look at a couple.

In 2010 the management of the civil service was largely put on a statutory footing through the Constitutional Reform and Governance Act - CRAG. CRAG however is in practice very bare bones codification. It just states that the civil service is managed by a minister of the civil service, that civil servants are appointed on merit, that there should be a civil service code which form part of the terms and conditions of officials’ service, and that must be published - and there are minimal requirements for code.

The Civil Service Code on the other hand adds a little bit of flesh to the impoverished provisions of crag but it has quite an odd legal status because it’s an instrument issued under CRAG. There are three relevant sections here.

The first section states that civil servants are accountable to ministers who are in turn accountable to parliament. So again what’s emphasised is this overarching doctrine of ministerial responsibility.

And in the recent FDA [Rwanda] case Mr Justice Chamberlain said that the code’s provisions need to be interpreted within its constitutional context of which the key rule was ministerial responsibility.

The second section sets out the core values of the civil Service: integrity, honesty, objectivity and impartiality

The third section, entitled political impartiality, is maybe the most nuanced of the three sections.

First the section is a directive of self-denial that emphasises the subordination of the officials own political views to those of the politician.

As Max Weber said, when an official receives an order his honour lies in his ability to carry it out exactly as if it corresponded to his own convictions even if the order seems wrong to him. Without this discipline and self-denial the entire apparatus would collapse.

Second, the section is an instruction of responsiveness - the readiness of civil servants to do what the incumbent government wants. It’s the primary duty of the civil service.

That said, responsiveness can be complicated. To begin with responsiveness to whom? Is it to the department, or to the minister of the department, or is it to the government as a whole?

Moreover responsiveness isn’t just about acceding to direct instructions. It may mean looking past the ministers immediate wants and needs and thinking about those wants and needs which are temporarily eclipsed in the moment. In other words, responsiveness can be about thinking about the short term and the long term needs of the government or the minister. So again the civil service has a little more discretion than is suggested by ministerial responsibility

Finally the section qualifies the primary obligation of responsiveness in two ways. The first provision states that officials must be responsive but in a way that maintains political impartiality and is consistent with the civil service’s core values. For instance, under integrity, civil servants have to comply with the law. Under objectivity, civil servants are expected to provide advice which is evidence based. They can’t just make things up. And these values are really a qualification on the primary principle of responsiveness.

The second provision points to another limit on responsiveness - the principle of constrained partisanship. Officials have to be responsive to present day ministers but their actions should not affect any relationship that they might have with future ministers. So there’s a limit to the degree of support they can offer to any one government. Too much support for an incumbent government might undermine confidence that the civil service could support a future government as well.

And these restrictions in the civil service code are enforced by provisions in the Ministerial Code. The Ministerial Code states that ministers have a duty to give fair consideration to the advice that’s given by civil servants - and they have to uphold political impartiality - and they shouldn’t ask civil servants to act in a way which would conflict with the civil service code. But Ministerial compliance with the Ministerial Code is essentially just convention. It is not in statute.

So the internal law of the civil service makes it clear that the primary duty of civil servants is to be responsive to the incumbent government. Much of what is popularly portrayed as ‘speaking truth to power’ or ‘free and frank advice’ is really just being responsive to the entire needs of the minister, not just in the short term but also in the medium and long term.

Responsiveness is therefore qualified by these principles of public integrity and constrained partisanship which both demand a distance from the incumbent government. But if Justice Chamberlain in FDA was right, the Civil Service Code still needs to be interpreted in line with the core constitutional principle of ministerial responsibility so that any tension between those principles of integrity and of constrained partisanship are likely to be resolved in favour of ministers.

OK - so the internal law of the civil service largely focuses on responsiveness.

It is not a strong basis for a civil service facing a populist government and yet there remains this kind of nagging belief that the civil service and its resources are not entirely at ministers’ disposal. It’s often expressed as the civil service having a separate constitutional personality or having a higher duty to the Crown which is something separate from the incumbent government. And all that comes out in various ways.

So, for instance, whenever a general election approaches there can be pre-election contacts between the civil service and key opposition parties. And there’s purdah, which means that the government should be very careful about how it uses the government publicity machine for its own electoral advantage and the civil service has to police that. There is the accounting officer regime in which (usually) Permanent Secretaries are personally responsible for the management of public money and their department. And, at an every day level, the separateness of the civil service might be seen in the principles of public integrity and constrained partisanship . All of these practices demand some distance from the incumbent government and reinforce the sense that the civil service is somehow separate from its temporary masters.

This isn’t surprising. Institutions will build up understandings and values which seemed to persevere and outlast temporary office holders. This is what Hugh Heclo would call institutional thinking. It’s about valuing continuity, the long term over the short term. It’s about thinking of one’s role as a profession, as an office with reference to ethical standards and fiduciary responsibilities and a shared normative vision, and it means stewardship. It means not guarding but also investing in assets on behalf of those to whom the assets will be passed in the future. So just as lawyers don’t just have duties to their clients but also to the courts we might say that the civil service has duties not just to the incumbent government but also to something greater: the Crown.

CH Sisson who was a poet and a civil servant once said:

Our large bureaucracy is a constitutional nonentity. Our suspicion that it is nonetheless not absolutely nothing remains and indeed it is something.

There’s a triple negative there ... and Sisson was a bit of a joker. Here’s something else he wrote:

Here lies a civil servant. He was civil to everyone, and servant to the devil.

But you know he did later argue that the civil service did ultimately serve something other than the incumbent governments and he called it the Crown and from time to time this argument has been run and Helen McNamara’s comment to Cummings and Johnson was really just the most recent example²

But there are problems with seeing the civil service is having a separate constitutional personality or having a higher duty to the Crown.

The first is that the narrowness of the (above) examples that are used to demonstrate the separate personality. Many of them operate around the time of government transition and so are really limited in time and scope - and a lot of them actually require prime ministerial authorisation. They don’t really represent what the civil service does on a regular basis.

The second problem is it’s not really clear what the Crown is. The Crown in UK law is many things: a symbol, a person, the government, the state.

Even if we could agree that the Crown was separate from the incumbent government, it’s quite difficult to determine or work out under what conditions that would actually matter, what the Crown’s interests might be and how those interests could be kept separate from the incumbent government.

The argument that the civil service has a separate personality was run in the recent FDA case which concerned The Safety of Rwanda Asylum and Immigration Act. Given that the Act in affect breached international law the question arose whether compliance with that would be in conflict with civil service code. Chamberlain J was sceptical and he didn’t really answer the point because he said he didn’t know what a separate constitutional personality would entail and again he preferred to answer the issue by reference to key propositions within the constitution - of which the most important was that ministers decide government policy. The role of officials is to advise and implement policy.

The final problem is that successive ministers and senior civil servants have actually worked to downplay any sense of a separate personality or a higher duty. The classic statement of this is the 1985 Armstrong memorandum (in the online library within www.civilservant.org.uk).

This was issued by Cabinet Secretary Lord Armstrong following the prosecution of Clive Ponting who was a civil servant at the time who leaked government information.

Lord Armstrong was prompted to draft a memorandum about the role of the civil service following the acquittal of Clive Ponting and he declared that civil servants are servants of the Crown for all practical purposes. The Crown in this context means the government of the day. And the civil service as such has no constitutional personality. So he was denying both the civil service had a separate constitutional personality or that the Crown was something separate from the incumbent government.

And what’s striking is how historically the civil service has largely worked to paper over any kind of tension between its duty of responsiveness in any separate institutional duties. So again think of Sissons’ triple negative. He was really not comfortable talking about any additional duties. There’s a studied ambiguity that goes on here and it’s a kind of prudential practice that the civil service uses to finesse its fragile role within government.

This is the IfG’s Catherine Haddon:

Rather than wanting more clarity and principles, officials are far more likely to defend the necessary shade in them. The principle exists but doing the job is something else.

Silence and ambiguity allow for flexibility because neither ministers or officials really benefit from conflicts with the other. Ministers need officials to do, to implement, to get things done; and officials have no legitimacy without ministers. The glue that holds that relationship together is mutual trust and explicit declarations of some higher loyalties would just weaken that trust. And a fight against ministers would be very harmful to the civil service. The civil service would lose not just with ministers of the incumbent government but with those from a future government. So institutional thinking is crowded out by this sense of responsiveness.

This prudential practice of silence has costs. It means a lack of certainty and clarity over the rules and it emphasises expediency over principle. And it presumes that the civil service has the capability to continue to manage that ambiguity and that ministers present and future are actually willing to tolerate ambiguity.

So how would this civil service respond to a populist government that came to power with an agenda to engage in democratic backsliding?

Let’s call this populist government Transform and assume that it has a small majority in the Commons - and it shares a similar agenda to the Trump administration: hostility to expertise and diversity and intolerance to established institutions.

I’m going to ignore substantial positive policy change like leaving the ECHR and just focus on the impact on the civil service.

The executive has traditionally chosen to manage the civil service by non-legislative means, even leaving CRAG aside, because it allows for flexibility. But what seems like a strength becomes a problem if you get a government that is intent on democratic backsliding. And there are lots of non-legislative changes that could be done.

There could be the alteration of structures and the redistribution of resources, personnel changes, the overhaul of bureaucratic norms ... and I’ll just go through a couple of them. Most of these could be fairly easily done I think most would be legal and it would be difficult to describe them as undemocratic. That’s the problem with democratic backsliding. A lot of the changes could be advertised as making things more democratic.

Government departments can be relatively easy modified or even abolished - very costly but it can be done with limited parliamentary scrutiny - Transform could carry out wide scale redundancies. By statute civil servants have the same rights as private employees but you know during the coalition they got rid of 100,000 civil servants.

Officials can be shifted around, redeployed, sidelined or constructively dismissed. (We can think about Sir Tom Scholar the then Permanent Secretary to the Treasury under Liz Truss and Kwasi Kwarteng.) Whatever the formalities, if a minister doesn’t want an official the official has to leave.

The number of special advisers could be drastically increased because there’s no statutory cap on the number and while under CRAG spads can’t order officials do anything if you get large numbers they could probably end up controlling access to the minister.

Transform could also introduce outsiders through the exceptional route for a time limited period under CRAG.

Guidance and the orientation of civil servants could also be changed with relative ease. Esther McVeigh who was previously a Cabinet Office minister used new guidance that stopped all external spending on equality diversity and inclusion activity. Suella Braverman published changes to the legal guidance on legal risk because she was worried that government lawyers had been too cautious and blocking policy.

Transform ministers could simply sideline or just ignore the work of civil servants. I mean, the odd rejection is fine and legitimate but what about blanket rejection? Ministerial acceptance of the work of the civil service is largely taken for granted but a lot of it is just convention.

How would officials respond to such changes?

Everything is going to depend on the concrete particulars and but there would be a kind of spectrum of responses.

The FDA case established that civil servants could refuse to comply with actions which are illegal under domestic law. But the more likely or difficult example is going to be where you get a Transform minister giving a completely legal instruction which is unambiguously bad in the sense of damaging the government in the short or long term.

So what would happen?

Well, one answer is responsiveness. That will be the default response. Officials will just attept to clarify what was being said and flesh out what the minister wanted. They would generally try to implement the policy perhaps in the belief that as the policy’s weaknesses would be exposed and remedied.

Depending on the nature of the instruction, it might end up getting being escalated to the law officers if there’s an issue of legality. If there was a conflict with the civil service code it might end up ultimately with the Civil Service Commission. The Permanent Secretary as accounting officer might have a say in this if it involved public money.

But all of these are quite rare; and their effectiveness unclear.

On the other hand, where there’s a new government it’s not uncommon for the civil service to be over responsive to prove they can work with their new masters. And the civil service doesn’t just work on direct instructions. They get cues and steers - and over time they might get the steer that certain kinds of considerations are not really appropriate and over time these might be left out of advice.

A much more serious step would be guerilla government which is bureaucratic shirking, leaking and whistleblowing - but all these actions go against the grain of the civil service - and there’s a specific provision in the civil service code saying that officials shouldn’t frustrate the implementation of policies.

The most formal means of resistance is resignation. Elizabeth Wilmhurst resigned for instance from the FCO over the advice given about the legality of the invasion of Iraq. Our Chair tonight, Jonathan Jones, resigned over the Internal Market Bill. In America at the centre for Disease Control a lot of officials resigned after the new head was fired.

All of these things can happen but guerrilla government and resignation present two dilemmas.

First, when is enough, enough? These matters happen in real time and it’s difficult to tell at what point compliance becomes complicity and when it is legitimate to resist. There are no simple measures that officials can use to define the threshold at which they can say “No!”.

Second, should one stay to ameliorate the worst of what is to come. As the good officials leave, less principled officials may start to fill the gap³. Democratic backsliding can lead to bureaucratic backsliding.

So, in short, if there were a democratically elected populist government, it’s quite unclear what the civil service would be able to do at all. Its normative framework is not resistance but responsiveness, and institutional values can only take officials so far

So what can be done?

That depends on what we think the problem is and its severity.

If we think the problem is a distant threat or just about improving civil service efficiency then just tinker around the edges.

But I think the problem is a little more fundamental, a little more urgent. What I’ve been describing is the absence of any strong normative or legal basis for the work of the civil service itself - and linked to that the potential erosion of permanence, impartiality and objectivity.

Harkening back to a golden age is a perennial temptation in this area, but I don’t know that we can continue with the status quo anymore. Even if those who are currently intent on transforming the state in the UK seem quite amateurish we can expect them to learn and evolve. You can think about the difference between the first Trump administration and the second Trump administration. There’s limited clarity about the role of civil servants’ relationship with ministers and that leaves the civil service in a vulnerable position for ministers intent on democratic backsliding.

Now the civil service will almost certainly say nothing. There may be some kind of preparation behind the scenes but it will be behind the scenes. The best thing people might do is maybe departments focusing on strengthening integrity.

The question of the appropriate relationship between ministers and officials is really a political one and resolution of this really requires consensus because unilateral reform of the civil service by any government can easily be undone by a future government with a government majority.

A more formal activity would be a Commission. Both the House of Commons Public Administration Committee and the UK Governance Project both suggested establishing a Commission to look into the appropriate model for the civil service in the 21st century. Sounds great! But the obvious problem is that there is little incentive for ministers to establish a Commission that might recommend reducing ministerial control over the civil service. And there’s no incentive for politicians who think they might become future ministers to pay any attention to a Commission either. And there are similar problems with putting things into statute.

Conclusion:-

The civil service has a lot more discretion than traditional public law theory allows. At best it’s a quasi-independent actor within the UK constitution. But that doesn’t make the civil service some kind of constitutional guardian or bulwark. While officials can speak truth to power there is, as Janet McLean has said, very little constitutional scaffolding. The ability of the civil service to push back and its permanence is largely based on the convention and ministerial forbearance.

And what legitimacy officials have is derivative on:

· elected ministers in the first instance,

· on parliament, when enforcing a statute,

· and on the courts in adhering to legality.

And the civil service itself hasn’t helped itself by working to obscure any clear answers.

The history of the relationship between the civil service and ministers is one of trust and willingness by both sides to reach some kind of accommodation. That accommodation is fragile and contingent and too much depends on ministers who have largely avoided defining a role the civil service in the constitution. The civil service ultimately exists as part of an ecosystem which underpins and supports democratic governance but the existence of this ecosystem and its various constituent parts should not be taken for granted.

Hugh Heclo was right. Institutions matter, and the danger in our time is not too much but too little institutional thinking.

---

Postscript

There was a recent Bluesky debate about whether a written constitution might constrain Transform Reform if it were ever to gain power. Robert Saunders argued as follows:

There is a serious, principled case for a codified constitution. But botching together a constitution to stop Farage - a constitution that could not command widespread support and that was widely seen as rigged - would deepen our problems, not resolve them.

A good constitution lifts the rules of the political game above the control of any single party or government. But it can only do that if it commands deeper loyalty than any one administration. A constitution explicitly devised to constrain one party, just before an election, won’t achieve that.

Any attempt to codify a constitution now would face open war from Reform, the Tories, the right-wing press, X & parts of Labour. No citizens assembly could function amidst such partisan fury. Its advisors, experts & reports would be trashed. Its chances of winning a referendum would be almost nil.

What would this constitution include? The monarchy? The House of Lords? The ECHR? First Past the Post? Devolution? All these things would be highly contentious - esp. if they then became harder to change. We always tend to assume a constitution would say what we want it to. That’s a mistake.

How would a new constitution be entrenched? There is no mechanism by which one parliament can bind its successors. A constitution with the weight of history behind it, or overwhelming public support, would exert a potent moral constraint. But this constitution would have neither of those benefits

Constitutions are only as strong as the willingness of institutions, parties & - above all - voters to uphold them. A constitution cobbled together in a contested process, that lacked broad public support, in the absence of agreement on big constitutional questions, would be a high road to disaster

If we want to rebuild the defences against authoritarianism (as we should) we’d be better off using the instruments we already have. Dismantling Henry VIII powers, getting dirty money out of elections, even changing the electoral system could all be done more easily than writing a new constitution.

If there’s one thing that’s definitely worse than *not* having a codified constitution, it’s having a constitution that cannot command support, that’s widely seen as rigged or that codifies the wrong things. Such a constitution would not constrain authoritarianism. It would turbo-charge it.

---

Martin Stanley

1. Ben Yong is a public lawyer and Associate Professor at Durham University. His research focuses on the work of the executive and Parliament, the role of officials, and the maintenance of ideals in the grubbiness of organisation.

2. “The Benn Act” required Boris Johnson to seek an extension to Brexit by a particular deadline. Dominic Cummings suggested they just ignore the deadline and, when he was told that the police might come for them for non-compliance, Cummings and Johnson mused aloud that they could barricade themselves in No.10 and the police would have to drag them out, so creating a perfect photo opportunity showing that Johnson’s determination to get Brexit done. Senior official Helen McNamara replied that “the police don’t work for you in that situation, Dom, they work for us. It’s not your building, these aren’t your people. The police work for the Queen. We all work for the Queen.” And the idea was quietly dropped.

3. These issues and examples are discussed in more depth in my ‘Speaking Truth to Power’.