In 1978 Joel Barnett, Chief Secretary to the Treasury in James Callaghan's government, approved a simple rule for adjusting Scottish spending ahead of a devolution referendum that ultimately failed. It was a stopgap, designed to spare ministers an annual negotiation. Barnett himself later called its survival "a terrible mistake" and told a Lords committee it had never been intended to last more than a year or two. Nearly five decades on, it still determines the block grants of the Scottish Government, the Welsh Government and the Northern Ireland Executive — the single largest financial fact of the devolution settlement, and one that has never been put into statute.
The mechanics are narrower than the arguments around them suggest. The formula does not decide how much Scotland, Wales or Northern Ireland receive in total. It decides how their block grants change. When the Treasury alters the budget of an English (or England-and-Wales) department at a spending review or fiscal event, each devolved administration receives a population-share slice of that change, scaled by a "comparability percentage" reflecting how far the programme is devolved. Health is close to 100% comparable, so a £1 billion increase for NHS England generates roughly £97 million extra for Scotland at current population ratios. Defence and foreign affairs are 0% comparable, so changes there generate nothing. The devolved administrations then spend their blocks however they choose: a Barnett consequential triggered by English health spending can lawfully be spent on Welsh roads.
What the formula never does is ask what anyone needs. That omission cuts in every direction, which is why every part of the UK can produce a sincere grievance. The baseline each annual change is added to reflects spending patterns inherited from the 1970s, when Scotland's per-head allocation was already well above England's — a premium the Goschen proportion had entrenched since 1888. Because Barnett only adjusts at the margin, that advantage rolled forward untouched. Treasury figures have consistently shown identifiable spending per head in Scotland and Northern Ireland running 15–25% above England's, and English regions with acute deprivation see none of it.
Wales holds the opposite grievance. The 2009–10 Holtham Commission calculated that a needs-based assessment — using the same indicators the Treasury applies to English local government and NHS allocations — would give Wales around 115% of English per-head spending, while Barnett was drifting Wales down towards 112% and falling. The culprit is the so-called Barnett squeeze: because consequentials are allocated per head, every uplift nudges devolved per-head spending arithmetically closer to the English level. The higher a nation's starting premium and the faster nominal spending grows, the sharper the convergence. Rather than replace the formula, the 2016 Welsh fiscal framework bolted on a patch: a 105% needs multiplier on Welsh consequentials, with a floor stopping relative funding falling below that level. The formula itself was preserved, and its output simply gets corrected after the fact.
A convention, not a law
None of this rests on legislation. The formula lives in the Treasury's Statement of Funding Policy, a document the Treasury writes and can rewrite, and successive governments of both parties have kept it precisely because the alternative — an explicit, contestable needs assessment across four nations — is politically radioactive. The 2009 Calman Commission, the Holtham Commission and a dedicated House of Lords select committee all concluded the formula lacked any objective justification; the Lords committee said bluntly in 2009 that it should be replaced. Every government since has declined, and the 2014 Smith Commission, negotiated in the days after the Scottish independence referendum, entrenched it further by promising its continuation as part of the settlement.
Devolved tax powers have complicated rather than displaced it. Since the Scotland Acts of 2012 and 2016, Holyrood sets Scottish income tax rates and bands, and the block grant is reduced by a "block grant adjustment" reflecting forgone UK revenue — a mechanism whose indexation method, the awkwardly named "indexed per capita" approach, was itself fought over for years because different formulas shift hundreds of millions of pounds depending on relative population growth. Barnett now sits atop this machinery as the default everyone criticises and nobody replaces. Its defenders make one honest argument: it is automatic, cheap to run and starves no one of a predictable settlement. Its critics note that a funding system for a quarter of UK public spending rests on a forty-eight-year-old back-of-envelope rule kept alive by the fear of arguing about anything better.
