Plato compared democracy to a ship where passengers vote on navigation instead of deferring to the navigator.
Many democracies have accepted the premise that a state is naturally incompetent, so we should use the private sector as much as possible. Markets do some things far better than ministries, and competition can discipline cost.
But the premise went too far. Governments stopped asking which tasks the state should do itself and treated outsourcing as the default. Design, engineering, project management and even policy analysis moved to consultants, contractors and lawyers.
The result was a continual hollowing out. A state that doesn’t employ engineers cannot judge an engineering bid. A department that outsources its thinking cannot tell a good proposal from an expensive one.
The premise became self-fulfilling. Treat the state as incompetent, strip out its expertise, and it becomes incompetent. Its failures are then cited as proof that it always was.
But can it be fixed?
That’s the wrong question. It has to be. In a changing world, a democracy must be competent, because the cost of incompetence is enormous and compounds.
Democracy doesn’t produce competence automatically. Plenty of democracies are dysfunctional, and the sceptics identify real weaknesses. But competence in a democracy is achievable, and it comes mostly from institutional design.
Britain Used to Be Functional and Does Have Competence in Some Places
Britain once built at speed. The first stretch of the M1 opened in 1959 after roughly 19 months of construction. In the late 1960s the country completed homes at roughly twice today’s rate, and new towns like Milton Keynes rose from fields.
Today the picture is different. HS2’s estimated cost multiplied many times over, and its northern leg to Manchester was cancelled in 2023. Hinkley Point C, Britain’s first new nuclear plant in a generation, has run years late and billions over budget.
The planning application for the Lower Thames Crossing ran to roughly 360,000 pages. These are not isolated scandals; they are symptoms of a system that lost its capacity in four ways.
A discretionary planning system. Since 1947, almost every significant development has needed case-by-case permission. Countries like Japan set clear rules: if a project complies, it proceeds. In Britain every scheme is a fresh negotiation, and fear of legal challenge makes applications defensively enormous.
A hollowed-out state. Outsourcing and the Private Finance Initiative locked the public into long, costly contracts, and the 2018 collapse of Carillion exposed how dependent government had become. The state lost the engineers and project managers who could act as an intelligent client.
Churn. Britain has had more than a dozen housing ministers since 2010, and civil servants move posts frequently. Few people own a project from start to finish.
Centralised but weak. Power sits in Whitehall, yet councils bear the local costs of growth while capturing little of its benefit. The incentive to say no is strong; the incentive to say yes is weak.
Yet competence has not disappeared. The 2020 Vaccine Taskforce, a small team with a clear mission, real authority and genuine expertise, helped make the UK one of the first countries to vaccinate at scale. The 2012 Olympics opened on time. Britain did not lose the ability to be competent; it lost the institutions that let competence operate.
Separate the Machine from the Argument
The central fix is to separate state capacity from policy. Policy is what democracy debates; execution is what the state should simply do well.
And we need to fix both sides: the machine that delivers, and the democracy that steers it.
Fixing the machine starts with insulating it from the electoral cycle. At present our institutions are swayed by voting in different directions, so long-term projects lurch with every change of government. We need to separate them.
Britain already has a model that works. Since 1997 the government has set the Bank of England’s inflation target, and the Bank decides how to hit it. Elected politicians choose the destination; an accountable professional body steers.
The same principle could govern infrastructure. A permanent delivery body, with in-house engineers, standard designs and a rolling pipeline of projects, would build what Parliament decides, without relitigating each decision at every step.
Ideally, there should be very little to debate if the system works well. Much of today’s political argument is really about failure: why the trains are late, why the project overran. When the state is competent, the debates diminish and become narrower in scope, focused on genuine choices of value.
Proactive Democracy
The other side to fix is democracy itself. We currently have a false form of democracy: a vote every few years, with little impact. Citizens hand over a blank cheque, then watch from the sidelines until the next election.
A proactive democracy would reach out to people rather than wait for them, bringing them into the management of the state, each where they know most. Residents know their streets. Nurses and teachers know where services break. Engineers know what a tunnel costs, and ordinary citizens are well placed to weigh questions of value.
This already works. Ireland’s Citizens’ Assembly, a randomly selected group of citizens, deliberated on abortion and shaped the 2018 referendum. Taiwan has used online platforms to find consensus on contested questions such as regulating ride-hailing apps. Cities from Porto Alegre to Paris let residents decide how parts of their budgets are spent.
Technology makes this possible at scale for the first time. Digital tools can turn consultation from a periodic ritual into a continuous conversation. Estonia already runs most public services online, showing that a digital state can be both efficient and trusted.
Proactive democracy also helps the state build. Engagement that comes early, before plans are fixed, surfaces objections when they can still improve a project. It replaces the slow, adversarial pattern of consultation as a formality followed by litigation as the real veto.
This is not a replacement for elections. It is a way of making democracy part of how the country is run, rather than an occasional verdict on those who run it.
Testing Like a Tech Company
Governments routinely roll out national reforms that have never been tested. No serious technology company would bet its future that way.
Tech firms run A/B tests: try two versions, measure the results, keep what works. Democracies can do the same with places. US Justice Louis Brandeis called the states “laboratories of democracy” in 1932, and the idea deserves a revival.
There are promising precedents. Finland tested a basic income on about 2,000 unemployed people in 2017 and 2018. Britain’s Behavioural Insights Team ran randomised trials on tax reminder letters and found that small wording changes raised payments. When Auckland upzoned most of its residential land in 2016, studies later found that homebuilding rose substantially.
Britain could go further. Designate trial cities or zones with permission to try different planning rules, procurement models or ways of delivering services. Measure the results openly, scale what works and drop what fails.
This turns reform from an ideological gamble into a learning process. It also lowers the stakes of political argument: instead of fighting over whose theory is right, we can find out.
The Objections
This argument has serious critics, and they deserve an answer.
“There will always be plenty to debate.” True: many disagreements are about values, not competence, and some see today’s vetoes as hard-won protections. Competence does not end the argument; it clears away the noise of failure so the argument can focus on the choices that matter.
“Policy and execution cannot be cleanly separated.” How a railway is built involves value choices about routes, compensation and nature. The line is blurry, but the Bank of England shows it can be drawn deliberately and revisited.
“Delegation becomes technocracy.” Handing execution to experts risks bodies that answer to no one. The safeguard is a clear mandate set by elected politicians, public reporting and leaders who can be removed.
“Participation gets captured by the loud.” Open consultations attract the organised and the angry, not the typical citizen. Random selection, as in citizens’ assemblies, is the best answer yet found.
“Places are not test subjects.” Trials on real communities raise questions of consent and fairness. Trial zones should be opt-in, time-limited and protected from harm if an experiment fails.
The Cost of Failure
The cost of failure is astronomical, but it is currently invisible and growing. It shows up in a housing shortage that prices a generation out of home ownership, in productivity that has barely grown since 2008, and in some of the highest industrial energy prices in the developed world.
The deeper cost is to democracy itself. When elected governments cannot deliver, citizens lose faith not just in their leaders but in the system. That opens the door to those who promise to get things done by sweeping democracy aside.
The good news is that the problem is solvable. Britain’s bottlenecks are organisational, not mysterious: a discretionary planning system, a hollowed-out state, constant churn and incentives that reward saying no.
Plato’s answer to the ship of state was to take the vote away from the passengers. The better answer is to hire skilled navigators, let the passengers choose the destination, and show everyone the charts along the way.