Arthur King | Writer & Commentator

In a U.S. Army war game in Germany, Ukrainian drone pilots playing the enemy caused about 70% of an American tank brigade’s losses in the first three days. Three days later, that figure was under 30%. defence-blog.com/ukrainian-d…

The official Battle of Britain memorial is at risk of closure, leaving the pilots who fought in the conflict in danger of being “forgotten”, veterans have said. 🔗: telegraph.co.uk/news/2026/09…

@JuliaHB1 Politicians line to bash each other and rarely generate detailed operations solutions to immediate problems. Here’s mine: Britain’s prison crisis looks like a shortage of cells. It isn’t. It is a shortage of certified, staffed, usable capacity — and that changes the solution. By April 2026 there were about 85,700 prisoners, against certified normal accommodation of roughly 70,000. The pressure valve is early release: 3,100 prisoners were released in one tranche in autumn 2024 simply to keep the system moving. The conventional answer is to build more prisons. We should — but it cannot solve the immediate crisis. The programme announced in 2021 eventually promised 20,000 additional places. Only 6,518 had been delivered by September 2024 and completion now runs towards 2031. Between 2010 and 2024 the net increase across the whole estate was just 1,005 places. So start with what makes a prison place usable: a serviceable cell, legal certification and enough warranted officers to operate it. The immediate plan: 1. Recover what we have. Some 4,151 cells have been indefinitely closed through dilapidation since 2010. Rank them by cost and speed of recovery and put those capable of rapid restoration ahead of slower new builds. 2. Create a capacity cascade. Put rapidly constructed modular accommodation onto existing Category C and D sites. Move suitable lower-risk prisoners into it. That releases conventional accommodation further up the estate, which can be upgraded or repurposed to relieve pressure on Category B and local prisons. This does not mean putting dangerous prisoners into lower-security accommodation. It means the opposite: Build quickly at C and D, move suitable prisoners down, and preserve the strongest accommodation for the most dangerous offenders. Creating one place lower down can release a more secure place above it. 3. Fix the real bottleneck: officers. In March 2026 there were 21,984 FTE Band 3–5 prison officers — 732 fewer than a year earlier, a fall of 3.2%. Yet retention is improving: the HMPPS leaving rate fell to 9.1%. That points towards recruitment and intake, not simply retention, as the immediate weakness. A cell without an officer isn’t capacity. Recruit aggressively while moving appropriate support functions — perimeter security, gates, escorts and transport — away from warranted officers where possible. 4. Attack the inflow. The remand population was 17,023 at the end of 2024 — around one prisoner in five and the highest level in half a century. With a Crown Court backlog above 73,000, thousands of prison places are occupied by people waiting for their cases to be decided. Build capacity without addressing that flow and we simply fill it again. 5. Keep building permanent prisons. New prisons remain the best long-term solution. They cannot be the main answer to a crisis now when the current programme runs towards 2031. Finally, put one person in charge. Not one official for construction, another for maintenance, another for certification and another for recruitment. One named person responsible for certified, staffed, usable capacity, with the authority and resources to deliver it. Publish four numbers monthly: usable places, officer strength, early releases and remand population. Hold that person accountable. We have spent years treating this as a construction problem. It isn’t. Recover the existing estate. Build rapidly at C and D to shuffle capacity upwards and keep the worst offenders in the secure accommodation designed for them. Rebuild officer numbers. Reduce remand. Keep building permanent prisons for the future. Stop measuring how many cells we promise to build. Measure how many prisoners we can actually hold safely and lawfully — and make someone responsible for delivering that number.

UK doesn’t need more tax. Here’s £50+bn a year within five years, without ⬆️taxes or ⬇️core entitlements. 🇬🇧 Government spent £1,360bn last year: 44.3% of national income. On the OECD’s comparable measure, public spending here is 46.0% of GDP, against a European average of 49.2%. Below Germany, Sweden and Italy. The size of the state is not the immediate problem. The case for higher taxes deserves a hearing. Britain borrowed £132bn last year while paying £130bn in debt interest. Defence is committed to ⬆️ to 3.5% of GDP by 2035. An ageing population will require ⬆️health and social care. But before asking taxpayers for more, take the uncontested money first. For now, leave; State pension £155bn. Universal Credit £84bn. Disability benefits £66bn. Social care £56bn. NHS clinical services. School budgets. Defence. Debt interest. Together, around three-quarters of government spending either reflects a legal entitlement or discharges an unavoidable public obligation. Those debates may be worth having later. That leaves roughly £250bn where government has genuine discretion. Within that more than £50bn can be found using the bottom of published estimates. Transport £10bn, from a £49bn budget where the railway alone receives £26.5bn. Discretionary grants £8bn, from a £34bn pool that entitles nobody to anything. Overhead savings across arm’s-length bodies £4bn. Civil service reform £6bn, around 100,000 posts over time. NHS management, administration, procurement and support functions £8bn. Business subsidies £5bn. Block grants £3bn. Digital transformation £3bn. Local government administration £2.5bn. Consultants £1bn. None of this cuts a pension, closes a school, reduces frontline NHS care or weakens the Armed Forces. Then there is the money that never appears in the spending debate. Britain has 341 tax reliefs costing £218bn in 2024–25. Only 25 have been formally evaluated in the past decade. The annual evaluation budget is around £600,000. That’s roughly £1 of scrutiny for every £360,000 of tax relief. No department owns them. No statute requires them to be reviewed whereas spending programmes must be defended. Tax reliefs simply continue because nobody decides to stop them. When government did evaluate one—Entrepreneurs’ Relief—it concluded the scheme was expensive, ineffective and poorly targeted. Reform reduced its annual cost from £2.3bn to around £600m. Exclude pensions and the family home, review the remainder properly, and £8bn a year is a conservative estimate. That produces recurring savings of roughly £58bn. This would take around five years, beginning at perhaps £12bn annually, with openly budgeted transition costs. These are gross savings, and reducing civil service numbers means real people lose jobs. Those consequences should be acknowledged honestly. So why hasn’t it happened? Because Britain’s institutions reward preserving programmes, not proving they still deserve to exist. No minister is criticised for leaving a programme alone. Many are criticised for abolishing one. Canada changed that in 1994. Every programme had to pass six tests: public interest, proper role for government, correct level of government, alternative providers, efficiency and affordability. The deficit disappeared within three years. Canada still publishes annual reviews of its tax expenditures. New Zealand completed the picture in 1988, making departmental leaders personally accountable for agreed results through fixed-term performance contracts. Both sides of the ledger. Every programme. Every tax relief. A named owner. A statutory review date. A sunset clause unless somebody justifies it. Last year, 🇬🇧 departmental administration spending rose 11.7% to £13.8bn. Nobody justified that increase. Nobody will justify it next year. That is why around £58bn remains readily available. No programme and no tax relief has to prove to a named person that it still deserves to continue. They just do.