Civil servants are individually and collectively approaching an ethical challenge. That would be dangerous territory at the best of times, but it is made doubly so by the fact that vanishingly few of them have spotted that there is a challenge at all. There is a resounding silence of leadership on the most fundamental challenge to the nature of the civil service in generations.

Brexit has dominated the politics of the UK for the three years since the vote to leave. It is self-evidently an important issue in its own right. But it has also raised constitutional questions of much wider significance, some of which have also been getting increasing attention. One which has so far remained in relative obscurity is the implications for the future of a non-political civil service. Brexit provides its context and in some ways its catalyst, but it is not fundamentally a brexit-dependent question and so is independent of views on brexit itself.

As I have discussed before, the ethical foundations of the civil service have some distinctive characteristics. Civil servants subordinate their personal political and policy preferences to the greater good of a wider political system. They will work tirelessly in support of goals they may not share and ministers whose party they did not vote for. They do so not because they are amoral or immoral but because they – and the constitutional settlement of which they are part – places high value on there being an effective and professional bureaucracy. So ministers decide – in practice as well as in theory – while civil servants analyse, advise and deliver.

That approach requires answers to three fundamentally important questions:

What makes the decisions which civil servants implement legitimate?

Where are the boundaries of that legitimacy and how can they be detected?

What should civil servants do if those boundaries are reached and crossed?

The initial answer to the first question is pretty straightforward: decisions are legitimate because they are made by ministers and ministers have democratic legitimacy as members of a government which enjoys the confidence of the House of Commons. For a very long time that has seemed to be a sufficient answer with little practical need to enquire further.

It has appeared sufficient because of one more element: an intangible, but very real, acceptance that the support of a majority of the Commons was a sufficient test of democratic legitimacy. In practice, it is now unknown for a party with a majority of seats in the Commons to have won them with the support of a majority of the electorate.1 But very clearly that has not created a crisis of democracy: the decisions of those governments have of course been politically controversial, but they have not in general been attacked for want of legitimacy.2  It’s not surprising that supporters of decisions and governments don’t go out of their way to question their validity. If anybody is going to do so, it is their opponents – and the fact that that generally hasn’t happened is a pretty strong indication of losers’ consent to the underlying system, if not necessarily to the specific outcomes that system generates.

That tacit agreement not to notice that there is a problem is now breaking down. There are two obvious drivers for that, and no doubt many more causal factors which could be identified. One is that the government’s response to the brexit referendum showed little sign of recognising the need for losers’ consent: ‘you lost, get over it’ may be satisfying in the moment but is hardly best calculated to broaden the perceived legitimacy of the decisions which followed. The other is a more direct break in the chain of democractic accountability and legitimacy: the deep confusion which has resulted from the tension between direct and representative democracy, exacerbated by the unintended consequences of the Fixed-term Parliaments Act 2011, resulting in a government with very uncertain parliamentary support.

So in short, the aftermath of the referendum has made the underlying problem of losers’ consent much more visible (even if not directly discussed in those terms), and the state of parliament has put further strain on the chain of democractic legitimacy, even in the weak form which has characterised the UK system for many years.

What then should the civil service do?

There is a simple answer, which is to carry on regardless. That is the answer still being assumed, based fundamentally on the idea that the government remains the government until it stops being the government and that for as long as it does so, it is not for the civil service to look behind the formalities of its continuing existence or to question its authority.

That position has some attractions: we don’t want to be in a world where the civil service takes it on itself to decide whether it likes a government enough to be prepared to work for it. But there is also a profound weakness: if this is not enough attenuation of authority at least to require questions to be asked, what would be? And that brings us on to the second question, about where the boundaries of legitimacy should be drawn.

Where are the boundaries of that legitimacy and how can they be detected?

There is no shortage of examples, historical and modern, of states which have kept the forms of democratic government while edging towards authoritarianism. The difficulty is that when those forms fall away, it’s generally too late to do much about it. Before that point, though, there is inevitably judgement and ambiguity, with a very understandable temptation to see the continuity of what is legitimate and fail to see the discontinuity to what is illegitimate.

I do not assert that there is a single objective test of whether we at, approaching, or beyond that boundary. Nor is it the point of this post to assert that any such test has or has not been met. The assertion here is instead that it is an ethical imperative for civil servants to be looking for that boundary and to avoid complicity in crossing it.

There is though a precautionary principle which is relevant to making the judgement. The dominant political myth in the UK is that its political system is inherently stable, bending and adapting to changing times, but never breaking. There are some pretty obvious perspectives from which that has never been true, of course, but that hasn’t challenged (and, to a remarkable degree, still doesn’t challenge) the myth. If that dominant myth were well founded, all of this would matter much less, we could safely treat it as part of the routine ebb and flow of politics, with which the civil service is entirely comfortable. But if the wider political system is more brittle than that dominant myth allows, we should be much more worried, and at the very least looking out for signs that we may be going beyond the point at which everything just springs back to normal.3 It is true that the UK has not had some of the radical discontinuities of government and constitution which many other countries have had to go through.4 But while that may be an indication that the elastic limit has not been reached, it cannot be evidence that the limit does not exist.

In their study of How Democracies Die Steven Levitsky and Daniel Ziblatt put forward four behavioural warning signs for recognising authoritarian leaders:

We should worry when a politician 1) rejects, in words or action, the democratic rules of the game, 2) denies the legitimacy of opponents, 3) tolerates or encourages violence, or 4) indicates a willingness to curtail the civil liberties of opponents, including the media.

And they stress that:

A politician who meets even one of these criteria is cause for concern.

The European Union (Withdrawal Agreement) Bill provides the basis for a thought experiment.5 There is no suggestion from the government that it should do anything other than go through all the normal parliamentary stages in both houses. There is in that formal sense nothing at all unusual about it. But the proposal is that a complex bill with very substantial constitutional implications should go through all its Commons stages in three days, starting only hours after the bill was published, in dramatic contrast to the time given to past bills of equivalent significant and complexity.6 That won’t happen, of course, without the consent of MPs, but irrespective of that, it’s reasonable to ask whether a bill passed in that way enjoys the same depth of democratic legitimacy as one given time for reflection and fuller debate. Again, it is not the purpose of this post to answer that question, but to suggest that it is one which needs to be asked, because, as Levitsky and Ziblatt observe:

there is no single moment—no coup, declaration of martial law, or suspension of the constitution—in which the regime obviously “crosses the line” into dictatorship, nothing may set off society’s alarm bells. Those who denounce government abuse may be dismissed as exaggerating or crying wolf. Democracy’s erosion is, for many, almost imperceptible.

If we accept that there is a boundary, however imprecise and hard to discern, and that it might have been crossed or might be crossed, we need to go to the third question.

What should civil servants do if those boundaries are reached and crossed?

In principle the answer to that is simple. At the point any civil servant judges that the democratic legitimacy of ministers has broken down, they must also accept that their ethical authority has also broken down. Whatever a civil servant does after that, they do as an independent moral agent, personally responsible for their decisions and actions. They may nevertheless choose to continue, accepting that responsibility. Or they may choose to walk away.

Again, my point here is not to dictate a course of action. It is to argue that whatever course is chosen, that choice should be deliberate and conscious. There will always be tempting arguments to stay a little longer, to do a little more, to believe that the spring is still elastic. The right course of action will probably be fully clear only in hindsight, and not necessarily even then. So these are hard choices in difficult circumstances. We should be reluctant to jump quickly to criticise how those choices are made. But civil servants need to recognise their responsibility to have these issues in mind and the civil service needs to be much more ready to support them in doing so.

The institution, of course will remain. Authoritarian governments have civil services, just as democratic ones do. But the surface form hides a profound difference. In such a civil service, loyalty is ultimately to the holders of power, not to the idea of good government, and the consequences are very different. Those who choose to be part of them are choosing to accept those consequences. The standard you walk past is the standard you accept – and that applies doubly to the standard you sit down with.

Perhaps all this is unnecessary fear mongering. Perhaps the political and democratic institutions of the UK are not only not at risk of immediate harm, but not even close to danger. I hope that proves to be the case. But there is a better chance of avoiding the dangers if we are prepared to recognise the risk.

  1. Nor is it self-evident that the support for a coalition is the aggregate of the support for the parties making up the coalition, but we don’t need to pursue that further here.
  2. The electoral system is itself a matter of political controversy, but the argument that it could be improved has never been an argument that decisions made under the present system are intrinsically invalid.
  3. Or, to stretch the analogy, when the elastic limit is reached.
  4. Though it is also true – and frequently overlooked – that the UK has only existed in its current form since 1922.
  5. This post was written after the Bill was published, but before second reading and, critically, before the debate on the timetable motion.
  6. Both the Constitution Unit and the Institute for Government have set out some striking comparisons.


  1. Your third paragraph certainly feels like comfortable constitutional territory, but I wonder if we should challenge those assumptions a bit more.

    I can’t be sure of course, but I wonder if some of those pushing ministers in authoritarian directions feel that an administration comprised of people who are doing things without believing in them is part of the problem, rather than a constitutional asset. cf, America, where of course things are a bit different.

    Civil service recruitment and PR tend to emphasise ‘work on things that matter’, to paraphrase, rather than an appeal to Northcote-Trevelyanism. I think the suspicion emerges between ministers and officials then (probably under any government, but I only have direct experience of two or three) that there’s a statist, liberal bias to the implementation of policy regardless of what manifestos declare and voters decide (admittedly in their shaky plurality).

    Perhaps we’re moving inexorably and not necessarily for the worse, towards a constitution with stricter boundaries, checks on executive power (hi there FTPA) and separation of powers more generally.

    There’s certainly a level of bureaucracy which needs to remain expert and impartial, but that doesn’t need to extend to most of the SCS, I’d argue. If ministers had the officials they trusted, with policies they could see taking effect, and our constitutional settlement was on the rules rather than the etiquette, then perhaps we’d see less of the frustration and theatre we see with Brexit, and ministers would complain less that the levers of power are ineffective. As long as it’s on civil servants to blunt the baser instincts of ministers for the good of society, the more confused the picture looks to those who are asked to vote on it.

    Still, I could be horrifically wrong, of course.

    1. I think your challenge to third paragraph is a good one, but slightly separate from the point I was trying to make. Clearly the Whitehall model of how a civil service should work is not to the only one consistent with democratic government, as is evident from many other countries. But quite a lot of the subsequent argument still stands (I think): the fact that senior officials may be political appointees doesn’t fundamentally change the question of whether they government they serve has sufficient democratic authority, and indeed the risk of being over optimistic about the level of that authority may perhaps be the greater because there is even less reason for such appointees to call it out.

Comments are closed.