Going for gold

Thoughts on civil service reform

When I was a child I took a dim view of the civil service. I didn’t really know what it did but it seemed to yield the most boring contestants on Going for Gold and Blockbusters. The grey-faced ones who, when introduced by Henry or Bob, could only explain their job with those two words: “civil servant” in a laborious monotone. Later, a good friend of mine from university joined the civil service and I remember thinking she had chosen the most boring possible path for herself. I meanwhile, had become a property journalist and spent my days interviewing chartered surveyors. (The point of this story is that I was a fool).

These days it seems everyone has a dim view of the civil service. Keir Starmer’s angry about reckless security procedures as well as frustrated by excessive caution and a blocker mentality. Danny Kruger makes videos of himself wandering around Whitehall pointing at buildings and telling us how many people Reform UK will sack. Kemi Badenoch has conceded that 10% of civil servants are good, but says 10% should be in prison. Liz Truss still believes she was taken out by the deep state. Ed Davey says he was lied to about Post Offices.

But most proposals I see for civil service reform rest on magical thinking. That mass redundancies will make the state more effective. That shouting at people to obey you is a useful motivational tool. That people who come into the office are always highly productive. That all we need is an injection of genius-level talent at the top. That there are levers to pull to make things happen.

I am, by instinct, a reformer. I believe the civil service needs to change, and radically. But serious proposals need to engage with some immutable realities. I’ve got eight; and then I’ll put forward some ideas for reform that might actually work.

1. People have to want to work for you

This is the most important. Most politicians act as if the problem with the civil service is that it doesn’t pick the right people. But picking people is easy if the right people apply and stick around. Talented people have options. Political leaders need to take responsibility for making the civil service a place that talented people want to be. You cannot simultaneously tell people the civil service is a nest of incompetents and expect the competent to queue up to join.

2. Selection effects matter

People are not randomly sorted into professions and careers. Every organisation has characteristics, and those characteristics affect who works there: this isn’t the deep state it’s human behaviour. Universities are not filled with rapacious capitalists because the pay isn’t great and there’s no commercial upside. Police officers are (on average) more interested in law enforcement than nursery school teachers. Many of the civil service’s characteristics are intrinsic to the kind of work that the civil service does. You can shift this a bit but most private-sectory risk-takers want to work in a risky private-sector environment.

3. Organisational culture is sticky

Culture is the accumulated residue of thousands of decisions, incentives, stories and norms, laid down over years and encoded in the behaviour of people at every level of the organisation. It is not changed by announcement and organograms. This is even true in small organisations: at Demos I changed almost all the staff, all the trustees and changed offices but there was still an intrinsic Demos-y ness about the place, for good and ill.

4. Senior tenures are short, and short tenures produce bad incentives

Linked to this, is that the average tenure of a permanent secretary is under five years. Ministers last, on average, less than two. Special advisers turn over faster still. If you have two years in a role, the correct strategy is not to fix the system. It is to find something you can credibly claim credit for within two years, and make sure the claim is visible. This produces a very particular kind of activity: announcements, reviews, pilots, new units, reorganisations - all the things that can be initiated in a short tenure and attributed to the person who initiated them. It systematically underproduces cultural change, capability building, the steady unglamorous work of making large systems function better. The rational response of the people below this high-churn leadership layer is to consent and evade. You nod. You produce the required documentation. You wait. The minister moves on and the initiative quietly dies, and you were never going to be rewarded for taking a risk on something that might fail anyway. Short tenures at the top produce a form of learned helplessness in the middle - and learned helplessness, once established, is very difficult to treat.

5. No organisation relies only on geniuses

There is a fantasy version of civil service reform in which the mediocre are swept out and replaced with exceptional people, and the exceptional people then do exceptional things. This fantasy misunderstands how organisations work. The most effective organisations are not the ones with the sparkliest stars; they are the ones that make it possible for ordinary people to do their jobs well. That requires clear processes, good management, psychological safety, and a reasonable relationship between effort and reward. None of these are glamorous. All of them matter more than the quality of the top five percent.

6. Initiating ideas is a lot easier than delivering them

This is particularly true when it comes to the pathway from genius idea to practical change. Every change needs to be initiated, but initiating things can easily become a pathology of its own. Politicians and special advisers don’t just have short tenures: they have short attention spans. Politics selects for people who like pace, thrive on adrenalin and score a zero in the “completer finisher” box on psychological assessments. If these people want to actually deliver things, they do not need more people like them around them: they need people who think and behave differently.

7. Government is not the private sector (and it’s definitely not a start-up)

The private sector comparison is irresistible to reformers and almost always misleading. Businesses can choose their customers. They can exit unprofitable markets. They can let failing products die. Government cannot. It is responsible for everybody - including, and especially, the people that no profitable business would touch. It operates under a degree of public scrutiny and democratic accountability that has no private sector equivalent. It cannot offer equity or uncapped bonuses. Its “shareholders” vote every five years and have radically divergent views on what the organisation should be doing. Importing private sector management techniques wholesale ignores all of this. Some things transfer well: a focus on user needs, clearer accountability, better data. Others transfer badly or not at all.

And the start-up comparison is not even wrong - it is a different category of error. Start-ups are defined by their freedom from the constraints that define government. Using one as a model for the other is like studying cheetahs to improve the performance of container ships.

8. Running large organisations is materially different from running small ones

In part because of this. The skills that make someone brilliant at building something from scratch are not the skills required to run it at scale. This is one of the best-documented findings in organisational psychology and one of the least applied lessons in politics. Founders and political leaders share a particular profile: high vision, high risk tolerance, strong personal authority, the ability to move fast and demand the same from others. These qualities are not just valuable but essential at the initiation stage. They become actively harmful at scale, where the job is not to inspire but to coordinate - to get hundreds of thousands of people with different motivations, different capabilities and different ideas about what matters, working together well enough to deliver things.

Don’t despair

OK: that’s 1000 words of why it’s hard. Do we just give up? No. We stop bashing the civil service for not being like politicians, and instead invest time and energy in building it to become what politicians actually need. Suggestions:

1. Secure cross party consent. This is often the policy refuge of the scoundrel, so I mention it with trepidation, but I can’t see a way forward without it. Every incumbent government’s effectiveness is undermined by the opposition’s threatening radical reforms or making accusations that the state is corrupt and failing. It affects their ability to attract and retain talent, and it distorts incentives. It is better to cede some control over the future of the civil service.

And if you expect to be the incoming government: there is so much more to be gained from having an effective civil service than there is to be gained by political name-calling about the ‘deep’ or ‘lazy’ state. Those short term gains in opposition will be no comfort when you get into power and discover it wasn’t true, and that you have over-promised. If you honestly believe the only reason problems haven’t been solved is because you weren’t in charge, you need a therapist, not a job in Downing Street.

2. Reform of terms and conditions. It is really hard to restructure teams in the civil service. It is even harder to get rid of mediocre performers. The timescales for reform, and the scale of redundancy payments needed to downsize are too onerous. This is hard to change especially for the ‘stock’ of civil servants but it is far more important a problem than supposedly gold-plated pensions which receive far more public attention.

3. Get real about pay. Public servants are not (mostly) overpaid. In fact, we tend to underpay for critical functions, so we rely on external expertise in the shape of consultants that costs five times as much, and get outmanoeuvred into total dependence on them. The political dynamic that says we should pay less for the sake of ‘efficiency’ is nonsense. Salaries should go up, especially for specialist skills. We need to build internal, publicly governed, capacity for complex roles.

4. Turn the reform process upside down. This one’s a bit harder to explain. It’s grounded in what I’ve heard called the ‘new co’ model of delivery.

Here’s the playbook: set up a new unit, give it proximity to power, protection from the usual rules, and the best people you can find, and let it rip. This has become the default delivery model for ambitious reformers and it’s easy to see why. When tenures are short and the underlying system is resistant to change, carving out a protected space where things can actually happen is a rational response. Whole careers have been made by leading a “new co” team: it gives you brand and reputation as someone who knows how to make things happen.

But it means that fixing the underlying system remains permanently nobody’s job. It leaves the rest of the system not just unreformed but unreformable: stripped of the attention and access it needs to get anything done. It pulls the best talent away from the core, concentrating capability at the edges of the system while the centre continues to calcify. New co is where you make your name. Old co is where careers go to die.

Our new Cabinet Secretary Antonia Romeo has been brought in to fix the civil service. Will she? Or will she look for some demonstrable early wins by setting up some shiny new-cos? You tell me.

I don’t see how you fix the politics of this. Every minister (and perm sec) will want their shiny thing.

The best answer, I think, is reverse acquisition. In the private sector, large companies often buy small companies. But sometimes small, high-growth, effective - companies sometimes buy large ones. They bring their culture to bear on the assets and infrastructure of the large organisation.

What if the Behavioural Insights Team hadn’t spun out as a global consultancy but taken over the prevention functions of the Department of Health? What if the team that ran London 2012 hadn’t been disbanded, but had taken over DCMS? What if the high-performing passport office took over the constantly-failing immigration and nationality service? (I don’t mean just move the high-performing boss into a new job.)

I don’t know if these are the correct specific ideas but I’m proposing a mentality shift. There is energy and ambition when you start a new thing, with the freedom that a blank page affords. But too often we leave these things as shining examples at the edge of the system, useful to someone’s CV, and comfortable only for those inside the bubble of high-performance. Instead, let the edge cannibalise the centre, chewing up the dysfunctional bureaucracies and turning them into something digestible.