Working Paper · The Prison and Punishment System
The state has been poor at running things for as long as anyone can remember. The usual explanation is that public bodies are simply bad at delivery. The real fault is structural. Government fuses two jobs that ought to be held apart: working out what to do, and getting it done. When one body owns both, and no single person owns the result, delivery is what gets lost.
· I ·
Take a hard task. Reform of the prison and punishment system in England and Wales runs for a decade, crosses prisons, probation, the courts and the estate, and depends on perhaps eight separate changes all landing. The standard response inside government is that a department can manage one or two large reforms in a Parliament and no more. That response is not a law of nature. It is a description of a delivery culture that has never separated the deriving of policy from the driving of it.
There is a way to test the claim. Hand the task to the discipline that exists precisely to deliver complex, concurrent operations under uncertainty, and see what it produces. A military headquarters given a defined objective does not ask how many of the constituent tasks it can afford to attempt. It runs an estimate. It fixes the end state, finds the one thing on which everything depends, sequences the rest against it, and drives delivery through a single commander who answers for the result. What follows is that estimate, applied to the prison problem, written so that a reader who has never seen one can follow the method as it works.
This paper is the delivery companion to the full policy document, A Ten-Year Strategy for the Prison and Punishment System.
· II ·
In Whitehall usage the strategic level has come to mean boards, frameworks and ownership spread so wide that no individual answers for anything. The operational level is different. It is where a defined task, however complex, is delivered by a commander who holds both the authority to act and the responsibility for the outcome.
To operationalise government is to move a problem from the first condition into the second. That is why this is written as an operational estimate and not a strategic one. The choice of method is the argument itself: a defined task, a named owner, a result somebody carries.
· III ·
An estimate begins with the commander's intent: the condition to be reached, stated plainly, so that everything below is measured against it. Here the intent is the mature system the reform is meant to leave behind by 2036. The figures are planning ranges, built from published baselines, and are meant to be varied rather than defended to the last thousand.
holds a stable population of the order of 87,000, near today's level, within a usable operational capacity of roughly 95,000 to 97,000 certified and staffed places; keeps a deliberate reserve of 8,000 to 10,000 places, so that it can repair itself rather than run permanently full; holds the high-security and specialist tier to around 8,000 to 9,000 places on shared-perimeter campuses; carries an officer workforce of broadly today's size but far greater depth and integrity; runs an expanded and redesigned probation service; and renews its estate through matched closure and replacement.
The intent is deliberately unheroic. It does not promise a smaller prison system. It promises one that can carry its load without the recurrent emergency releases that have marked recent years. Everything that follows concerns only how to reach that condition from where the system stands now.
· IV ·
Every estimate looks for the centre of gravity: the one thing from which the plan draws its coherence, such that if it fails, everything else is wasted motion. For this campaign the centre of gravity is not a building and not any single reform. It is the union of authority and responsibility in one accountable command.
The reason is the diagnosis the whole project rests on. In the present arrangement the courts fill prisons they do not pay for, prisons hold people they did not sentence, and the centre spends roughly a third of the money above the prison gate with no named owner. Each part optimises its own fragment and no one carries the whole. A reform delivered into that structure is absorbed by it. So the decisive requirement is a single commander who can move resource between the moving parts and who answers, by name, for the end state.
· V ·
An estimate distinguishes the tasks that can be willed to happen quickly from the ones that cannot. Here one factor sets the tempo for everything else, and it is the workforce. A probation officer takes fifteen to twenty-seven months to qualify. A prison officer takes six to seven months to recruit and then years to become fully experienced. Concrete cures at its own rate.
These are lead times, not attitudes. No amount of will compresses them.
The operational consequence is exact. The workforce lines must start on day one, running in parallel with everything else, because they gate every later effect. A plan that puts up buildings before it has the people to run them produces nominal capacity, which is no capacity at all. This is also the honest answer to the objection that the whole programme is too broad to attempt at once. The breadth is not the constraint. The lead times are.
· VI ·
The campaign runs seven lines at once. Concurrency is the normal condition, assumed rather than avoided. The estimate's job is to name the main effort and mark the genuine rate limits, so that when resource runs short it is stripped from supporting lines to protect the decisive one, rather than every line being shaved until all of them underdeliver. That last habit, trimming everything and stopping nothing, is the failure the whole approach is built to avoid.
Certified, staffed, usable capacity: recovery of repairable closed cells, rapidly deployable lower-security accommodation, and the officers to run all of it. This line carries the pacing constraint and the main weight of command attention.
The single accountable command established with authority across every line and a fixed-tenure chief executive held to the end state.
Professional formation, paid mentoring, specialist careers, targeted supplements, and counter-corruption treated as defence against an attack on the institution. Shares the pacing constraint.
Aggregate sentencing demand reconciled with capacity without touching judicial independence in the individual case, through a capacity remit on the existing sentencing body and firm control of recall and breach returns.
Intensive community punishment that substitutes for real custody rather than widening the net, enforced by swift and certain consequences where the monitored behaviour is objectively testable.
Probation expanded in step with intensive-order volumes, with professional judgement separated from routine compliance work. Shares the pacing constraint.
Obsolete establishments replaced through matched closure and modern build, funded from a ring-fenced renewal fund, with no closure preceding operational replacement.
· VII ·
This is where the answer to the "too many things at once" objection becomes visible. The seven lines run down the side; the five phases run across the top. Every line is live from the first phase. What the matrix shows is not sequence in the sense of one thing after another, but concurrency with dependencies marked, so the eye can find the single thread that gates the rest.
| Line of Operation | Phase 126–27 | Phase 227–29 | Phase 328–31 | Phase 430–33 | Phase 533–36 |
|---|---|---|---|---|---|
| 1 Capacity spine main effort | |||||
| 2 Accountable command | |||||
| 3 Officer profession | |||||
| 4 Demand management | |||||
| 5 Community punishment | |||||
| 6 Probation capacity | |||||
| 7 Estate renewal |
· VIII ·
Honesty requires naming the gap. The workforce investment across the ten years, the professional premium, the longer training, the paid mentoring, the targeted supplements, is real and recurrent, and it has not yet been costed to a firm figure. A dedicated workforce costing model is the necessary next piece of work.
What matters more is that the size of that figure is not the quantity on which the plan turns. The workforce is the centre of the whole design for a reason that has nothing to do with its price. Without the officers and probation staff established first, every other line fails regardless of what it costs: the buildings cannot open, the community sentences cannot be enforced, and the accountable command has nothing to command. The decisive question is not how much the workforce costs. It is whether it is built before the rest of the system is asked to rely on it. Spend more or spend less, that sequence is what determines whether the plan delivers or joins the long record of announcements that never became results.
· IX ·
The plan is broad. It is also deliverable, and the estimate shows why. Set the accountable command first. Start the workforce on day one and protect it as the pacing effort. Run the immediate capacity plan as the first phase of the main effort. Hold the more speculative lines, community substitution above all, behind evaluation gates. Let the genuine lead times, and only those, set the tempo.
Done that way, the campaign does not fail for being too broad to attempt. It fails only if the accountable command is diluted, or if a downstream line is pushed ahead of the workforce that must deliver it. Both are failures of sequence and command. Neither is a failure of feasibility. The state's long record of poor delivery is not evidence that the state cannot deliver. It is evidence of what happens when the deriving of policy and the driving of it are left fused in one confused hand, and no one is left holding the result.
This estimate is one worked example of a single idea: that Britain's institutions fail because accountability has been engineered out of them, and that the cure is a named person, trusted with intent and held to the result. That idea runs through both books in the Decline and Renewal series.