There aren't any levers

A dead metaphor is killing the government's ability to get things done

When I was a child, my mother (a doctor) used to enjoy complaining about medical inaccuracies in the plot of Casualty and I used to roll my eyes at her. These days, my children roll their eyes at me, because I love to complain about political inaccuracies on screen: Doctor Who, The Diplomat, Love Actually - it’s always the wrong people, the wrong rooms and the wrong protocols. And yet the one mistake I’ve never seen on TV is a Prime Ministerial office that contained levers for operating the government. That seems to be a mistake that only the real politicians make.

This week, in his evidence to the Liaison Committee, Keir Starmer mimed pulling a lever as he complained: “As Prime Minister, every time I go to pull a lever, there are a whole bunch of regulations, consultations, arms-length bodies that mean the action from pulling the lever to delivery is longer than I think it ought to be”.

He is not talking nonsense. There is a real problem with the effectiveness and responsiveness of the state. But the “levers of government” is a dead metaphor that is actively holding us back from solving that problem.

More than a metaphor
In an earlier draft, I had about 500 words here on the problem with the levers metaphor in a world of distributed power. Here’s the short version:

Change doesn’t happen in straight lines and is propagated (or not) by human actors with their own preferences and motivations. Delivery is disaggregated between departments, tiers of government, sectors, agencies, regulators and professions. The economy is not a thing but an aggregation of billions of decisions, each made on the basis of incentives, opportunities and desires. Citizens’ choices and relationships control far more of what happens in our country than any Prime Minister could ever dream. Culture eats levers for breakfast.

As I wrote, I could just hear the lever-bros over my shoulder saying: precisely. You’re coming up with excuses for impotence. That’s why we need more control at the centre. So I’ve put my 500 words at the end, for those who don’t get irritated by these observations. That way I can concentrate my argument on those who believe that centralising power will work to achieve their goals.

There are some lever-like bits of government. Ministers can change tax rates with precision, adjust benefit entitlements by age, income or household type, and move billions of pounds between accounts overnight. Politicians never complain about unresponsive levers when they are talking about these things. They work in terms of mechanical functioning. What they can’t guarantee is outcomes.

Changing a tax rate is a lever; getting the money in is subject to behaviour; generating growth is a system response. Adjusting welfare payments is a lever; getting people to claim them is subject to behaviour; improving children’s lives depends on personal regulation, institutional performance, relationships and time. The machinery can move the inputs with precision and still miss the outcome entirely.

The word ‘tools’ is a better one than lever. Government has tools, it’s just that they are not mechanically connected into that system like a lever is.

It is also true that many of those tools need improvements. If the way to govern well is to design, tweak, watch, learn and tweak again, then we must accelerate the pace of that learning and improvement cycle. Legislation should be faster, consultation could be accelerated, attitudes to risk should be changed, and most importantly of all, employment regulation for the public sector needs to be rethought, so that people can be moved, teams re-invented and functions redesigned without years of consultation.

They had a good idea, once
What makes all this more frustrating is that many people in and around this government have, at times, understood exactly this problem and articulated a better answer.

The idea of “mission-driven government” was precisely an acknowledgement that you cannot command outcomes you don’t directly control. You have to set direction, create alignment, mobilise actors who don’t report to you, and sustain effort over time. Missions work not by issuing instructions, but by shaping incentives, building coalitions, and giving people permission to move together even when formal authority is fragmented.

This wasn’t fluffy thinking. It drew on a real body of both academic theory and delivery practice. NESTA and Public Digital (an agency founded by the early leaders of the award-winning Government Digital Service) published The Radical How in 2024, UCL’s Institute for Innovation and Public Policy has been developing the thinking for years - this beginner’s guide from 2019 is a good place to start. Mission-thinking reflected lessons from industrial strategy, climate policy and health, where progress comes from coordination across government, markets, civil society and professions.

And to be fair, traces of this approach are still visible. Programmes like Test, Learn and Grow explicitly recognise that reform has to be iterative and grounded in practice, not announced fully formed from the centre. The No.10 Partnerships Unit exists because Downing Street knows it cannot deliver priorities without working with actors beyond the state. The new Office for the Impact Economy reflects an understanding that markets, social purpose and public goals can be aligned rather than held apart.

The tragedy is this thinking has been marginalised. Mission-shaped, system-level work is treated as an interesting supplement rather than the core operating model of government. It sits alongside the “real” business of targets, announcements and delivery plans, instead of replacing the mental model that produces them.

So when pressure mounts, as it inevitably does, the centre reaches back for the language of levers and control. Not because ministers don’t know better, but because the institutions of politics still reward the performance of command more than the practice of coordination. Mission-driven government becomes a side dish, when it should be the main course.

If we are serious about making the state more effective, this has to flip. Leading change in a distributed system is not a niche capability or an innovation add-on. It is the job.

Political gap
There is also a real political gap here, and it’s worth naming it honestly. A Prime Minister does not just want life in the UK to get better; they also want to be seen to have made it better. Democratic politics runs on attribution. “More prisons” is a far better vote winner than “Setting the conditions in which crime falls through sustained, collaborative action between the state, citizens, professions and the private sector”.

Voters understandably want to know who to credit or blame. The media wants a face, a decision, and a hero or a scapegoat. Mission-led, system-shaping work spreads both effort and credit so widely that it can feel politically thankless, even when it succeeds.

There is a psychological dimension too. Politics selects for people with unusually high agency, stamina and belief in their own capacity to make things happen. You fight your way to the top of a brutal system on the implicit promise that, once you get there, you will finally be able to act. To discover instead that power is conditional, negotiated and indirect can be dispiriting at best. The temptation to reach for command-and-control is not just political theatre; it’s a human response to frustration.

Liz Truss took this impulse to its logical extreme, arriving in office apparently expecting hegemonic power from day one, as if the problem all along had been insufficient will. The collision between that expectation and the reality of governing a complex, interdependent system was swift and destructive. But the underlying instinct is not unique to her. It is the ever-present risk when leaders mistake visibility for control, and authority for capability.

The challenge, then, is not to mock the desire for grip, but to redirect it, away from the fantasy of levers and towards the harder, less glamorous work of shaping systems in which many others can act. That is a different kind of power, and it demands a different story about what leadership actually is.

Effectiveness in the real world.

Intellectually, as I’ve said - the government already has the answer. They have a coherent theory of how to lead change, developed seriously inside and alongside the Labour movement, and which has been tested in practice at the margins of Whitehall. My first recommendation is therefore the simplest and hardest: mainstream the theory you already believe in. Stop treating system leadership as an innovation, and start treating it as what governing is.

Operationally, even that shift does require real, uncomfortable change. The Prime Minister is right about the lag between decision and delivery, but the causes are structural, not cultural alone. His friends (and mine) at the Future Governance Forum wrote him a paper about how to fix the centre of government: take it out of the drawer.

If you want to change employment practices, speed up legislation, or abolish arms-length bodies to bring decisions in house, it will take time. In the interim, government has more tools than it sometimes admits. Bring back the Contestable Policy Fund to enable Ministers to get formal advice from external sources. Use political appointments more deliberately to strengthen strategic capacity rather than just communications. And above all, be clear about what you are trying to achieve. When leaders are precise, even the sclerotic system of Whitehall is capable of responding.

Then there is the political question. The language of “missions” has run its course. It has become to associated with the early struggles of this government. If the aim now is to carry the public with you, the more durable frame is national strategy: clear choices about what the country is for, what it is trying to build, and where collective effort will be concentrated. I’ve written about this before, too.

So many of the building blocks are already there in the government’s own agenda. Clean, homegrown energy. Pride in place. Sovereign defence, security and technological capability. A serious attempt to get Britain healthy again. These are not slogans; they are strategic domains that require long-term coordination across the state, markets and society. Framed properly, they make sense to voters in a way that process language never will.

And this is where the Prime Minister’s most powerful tool becomes obvious. It is not a lever: it is a voice. The ability to set direction, to explain trade-offs, to legitimise patience, to share credit, and to tell a convincing story about how change actually happens. In a system where power is distributed, leadership is less about pulling harder and more about speaking clearly enough that others choose to move with you.

That is not a consolation prize for the absence of control. It is the real work of governing.

Addendum: my personal vendetta against the levers metaphor

1. Levers imply linearity. You pull, something moves. But modern government does not operate on a single axis of motion. It is a dense web of incentives, obligations, contracts, professional norms, statutory duties, judicial oversight and political risk. Change propagates unevenly, ricochets off institutional boundaries, and is often refracted by actors whose behaviour is shaped as much by fear and habit as by instruction.

2. Levers assume sole agency. One hand on the control, one decision-maker in charge. But the defining feature of the modern state is not centralised power, it is distributed responsibility - we are actively disaggregating power through devolution and this is a good thing. Delivery is and always will be fragmented across departments, regulators, local authorities, arm’s-length bodies, private providers, charities, professions and frontline workers with statutory autonomy. Many of the people whose actions matter most do not work for ministers, and legally cannot be ordered about by them.

3. Levers belong to machines, not to living systems. They make sense for steam engines and railway points. They make far less sense for institutions made of humans, relationships and trust. Government does not respond mechanically to force; it responds socially to signals. Culture eats levers for breakfast. A policy instruction that aligns with professional identity, organisational purpose and public legitimacy can move astonishingly fast. One that doesn’t will stall, however hard you pull.

4. The metaphor subtly reinforces a hierarchical structure. Levers sit at the top of the system. They flatter the idea that authority flows downwards and compliance flows upwards. But power is far more conditional than that. It flows sideways. It pools in unexpected places. It depends on consent, competence and credibility. Leaders who believe they are failing because they can’t find the right lever often miss the real task: building coalitions, reshaping incentives, changing narratives, and earning trust across institutional boundaries.

5. Levers also misrepresent the object being acted upon. Much of what governments care about most - growth, productivity, innovation, resilience - does not sit inside the state at all. The economy is not a machine owned by government; it is a vast, adaptive system of firms, workers, investors and consumers, most of whom will never take an instruction from a minister. Public policy shapes the context but it does not directly operate the economy - when leaders talk as if they are pulling levers to “deliver growth”, they risk mistaking influence for control.

6. And then there are citizens, the most consequential actors in almost every government ambition, and the least susceptible to being “pulled”. Whether children are well-fed, streets are safe, energy use falls, health improves or communities thrive depends not just on what the state does, but on how people behave: what they buy, how they travel, how they care for one another, how they respond to authority, and whether they believe collective effort is worthwhile. Levers are the wrong metaphor for a world in which consent, norms and everyday choices matter more than control.

7. Levers encourage a kind of performative frustration. If you believe that government is a machine with jammed controls, it is tempting to respond by pulling harder, complaining louder, or promising to rip the whole thing out and start again. That posture may be emotionally satisfying, and politically useful, but it doesn’t get us closer to effective action. It keeps us stuck in a cycle of expectation and disappointment, rather than forcing us to grapple with what governing actually is.