This article was published in the Times on 28 October 2025.
The Office for Budget Responsibility (OBR) is reported to have dealt a hammer blow to Rachel Reeves’s numbers by downgrading its productivity forecasts for the economy. I remember only too well the sense of helplessness chancellors feel over OBR forecasts in the run-up to budgets. Your entire budget hangs on the verdict of steely and rather unresponsive bureaucrats.
But when it comes to productivity, this view is mistaken. Chancellors do have some control over our national productivity — in one crucial respect. According to the Office for National Statistics, around a fifth of our economic output comes from the public sector. Nearly half of that is the NHS. The Treasury is famous — or rather infamous — for micromanaging public spending far more than mostfinance ministries. Why not turn that to our advantage to transform productivity in the public sector so that next time the OBR decides to upgrade our trend growth rather than downgrade it?
Public sector reform is hard. I still have the scars on my back from the 2016 junior doctors’ strike. I won the dispute, but it lasted nearly as long as the miners’ strike and severely damaged my relationship with the medical profession. That made further productivity improvements much harder. David Cameron and Theresa May backed me to the hilt, but other reforming ministers have found themselves reshuffled to oblivion.
Resistance from unions is not the only issue. Sometimes the blob fights back. Rishi Sunak and I planned to reduce civil service numbers to 2019 levels to fund our increase in defence spending. Within days of the new government’s arrival last year, plans the mandarins detested were quietly shelved. Civil service numbers have since risen.
But the biggest issue is the lack of strategic cost reduction that is central to how the private sector works. When the Home Office overspends on the asylum budget, the Treasury demands it “consume its own smoke”. Cutting police officers is politically explosive, so a new national police computer gets shelved instead — the very project that would have made policing far more efficient for taxpayers.
It is hardly surprising that the Institute for Fiscal Studies believes that, on one measure, productivity declined by an average 0.3 per cent between 1997 and 2019. Over the same period, the private sector managed 0.8 per cent growth, hardly stellar but at least a gradual improvement.
A government that was serious about public sector productivity would need to galvanise the whole of the state to be more efficient.
That means measuring productivity in every police service, hospital and school. It means measuring both cash savings and things like improvements in school standards, which matter but don’t save money. Most of all, it means making clear to public sector workers that any pay rise above inflation will need to be funded by efficiency gains — so no more £9.4 billion bungs to the unions as happened last July. If train drivers are to be paid more, then there needs to be an agreement that new train services can be driverless. When Thameslink introduced self-driving trains on the core part of its network in 2018 (albeit with drivers still in the cab), the number of trains it could run per hour nearly doubled.
For the chancellor, there is a double bonus: better productivity forecasts from the OBR but also a big boost to public finances.
Treasury officials told me that increasing annual public sector productivity growth by 0.7 per cent a year (still well below private sector levels) would stabilise debt as a proportion of GDP, rather than allowing it to grow inexorably. For the public, that is the difference between better schools and hospitals — or painful cuts.
There are two bits of good news that should encourage the chancellor to be bold. Firstly, the Office for National Statistics, despite some well-publicised challenges on labour market data, has developed probably the most sophisticated tools of any country for measuring public sector productivity. I commissioned it to overhaul our approach, and that work is now completed. What gets measured gets done, so it would be perfectly possible to embark on a huge national efficiency programme.
Indeed, the NHS has already started this journey. In my last budget, it signed up to a 2 per cent annual productivity target in return for a big investment in its antiquated IT systems. That programme has been kept on by the new government. In its first year, it managed 2.7 per cent. Now, most GP services have digital telephony. Most hospitals are in the process of installing IT systems that, once teething problems are overcome, will reduce the admin burden of doctors and nurses. There are lots more challenges ahead, including dealing with the incomprehensible strikes being pursued by the BMA. But if our biggest and most expensive public service can deliver efficiency improvements on that scale, so too can everyone else.
The second reason for optimism is that AI is about to unlock a further explosion in efficiency. Analysis from Microsoft and Accenture says it could boost output across the economy by 2 per cent a year over a decade. The gains in the most inefficient parts of our economy, often — sadly — those run by the state, are likely to be even greater. Turning round our creaking public sector is not just essential — it is also possible.