Following the Money North: The Fiscal Geography of England’s Devolution Experiment
Download PDF
Executive Summary
In July 2026 the Prime Minister announced that English mayors would receive a share of income tax and business rates for the first time. The Fiscal Devolution Roadmap published with the October Budget set the timetable: mayoral strategic authorities will receive a share of business rates, plus 20 per cent of any growth, from April 2027, and a share of income tax from April 2028, replacing ring-fenced grants. The stated aim is to give places a direct stake in their own growth. The amounts, and the rules for adjusting grants, have not yet been set. This paper uses the data that will drive those rules to show who would win and who would lose.
- Tax bases are extremely unequal. In 2024-25 income tax receipts were 8,710 pounds per head in London, 5,795 in the South East and only 2,583 in the North East. London alone raised 29 per cent of England’s income tax.
- Income tax growth follows existing wealth. Between 2010-11 and 2024-25, income tax per head grew 11.5 log points faster than the English average in London and 11.1 points slower in the North East. Initial GDP per head explains 95 per cent of the differences in growth across the nine regions.
- Assignment would have redistributed toward London. With a grant adjustment indexed to English income tax per head, a 10 per cent income tax share assigned from 2010-11 would by 2024-25 have given London an extra 95 pounds per head a year and cost the North East 30, the East Midlands 29 and Yorkshire 27. From 2018-19 the pattern is more mixed, but London still gains most.
- Regions would carry new risks. Over three-year periods, the gain or loss from a 10 per cent share swings between minus 13 and plus 16 pounds per head in the North East and between minus 20 and plus 36 in London. Central grants currently absorb this risk.
- The business rates growth share is small and also favours stronger regions: 20 per cent of growth since 2017-18 would be worth 12.6 pounds per head a year in London and minus 1 in the North East.
- England has not been converging. The dispersion of GDP per head across English regions has risen steadily since 1998, and poorer local areas have not grown faster than richer ones.
The conclusion is not that fiscal devolution is wrong, but that its design matters more than its principle. Retention of tax growth rewards places that are already growing. If the aim is to help lagging regions, the income tax arrangement needs strong equalisation, protection against regional shocks, and a reference path for grants that does not simply lock in past divergence.
1. The Promise of Devolution
England is one of the most centralised countries in the developed world. Local and regional government raises a small share of its revenue through its own taxes; most comes as grants from Whitehall, often ring-fenced for specific purposes. Successive governments have tried to change this, through city deals, combined authorities, metro mayors and business rates retention pilots. The 2026 reforms go further. For the first time, English mayors will receive a share of income tax, the largest tax in the country.
Table 1. The Fiscal Devolution Roadmap: main elements
| Element | Timing | Detail |
|---|---|---|
| Business rates share for mayors | April 2027 | A share of business rates in place of Investment Funds, Mayoral Capacity Funding and Local Growth Funds, indexed to inflation, plus 20 per cent of any growth |
| Income tax share for mayors | April 2028 | A share of income tax receipts in place of elements of grant funding; amounts to be set at the Spending Review |
| Full business rates retention | By 2030 | Local and regional government to retain 100 per cent of business rates |
| Overnight visitor levy | From 2028 | Power for strategic authorities to levy a charge on overnight stays |
Source: HM Treasury, Fiscal Devolution Roadmap, October 2026; Budget 2026.
The design of such arrangements matters enormously. When a tax is assigned to a region, the grant it replaces is reduced by an amount that is meant to equal the tax at the start. What happens next depends on how that grant reduction evolves. If it grows in line with the national tax base per head, the region gains whenever its own tax base per head grows faster than the national average, and loses when it grows more slowly. This is the logic of the block grant adjustments that accompanied the devolution of income tax to Scotland. The Roadmap leaves these rules to be set. This paper asks what they would have produced in the past.
2. The Fiscal Map of England
The data used are the Office for National Statistics country and regional public sector finances, which allocate every tax and every item of public spending to the nine English regions, Wales, Scotland and Northern Ireland, from 1999-2000 to 2024-25. They are combined with regional GDP and population from the ONS regional accounts. The English regions are larger than the areas covered by mayors, but they provide the longest consistent series, and most mayoral areas sit inside a single region.

Figure 1 shows how unequal the income tax base is. Receipts per head in London, 8,710 pounds, are 3.4 times those of the North East and almost twice the English average of 4,623. The three regions of the Greater South East, London, the South East and the East of England, are the only ones above the average. Every northern and midland region raises between 2,600 and 3,200 pounds per head.

The other side of the map is spending. Only London and the South East raise more in revenue than is spent on them. London’s net contribution was about 5,400 pounds per head in 2024-25; the North East received a net transfer of about 7,200 pounds per head. These flows are the implicit insurance and redistribution of the current, centralised system. Every reform that ties regional budgets to local taxes alters them.
Table 2. The fiscal map, 2024-25
| Region | Income tax per head (pounds) | Net balance per head (pounds) |
|---|---|---|
| London | 8,710 | +5,374 |
| South East | 5,795 | +1,628 |
| East of England | 4,854 | -685 |
| South West | 3,659 | -2,500 |
| East Midlands | 3,183 | -3,585 |
| North West | 3,100 | -5,171 |
| West Midlands | 3,010 | -5,052 |
| Yorkshire and The Humber | 2,925 | -4,285 |
| North East | 2,583 | -7,158 |
| England | 4,623 |
Source: ONS; author’s calculations.
3. Growth Follows Wealth
The promise of tax retention is that regions which grow their economies will be rewarded. The question is which regions have in fact grown their tax bases. Figure 3 shows the growth of income tax per head in each region relative to England between 2010-11 and 2024-25.

Only London grew faster than England, by 11.5 log points. Every other region grew more slowly, the North East by 11.1 points less, the East Midlands and Yorkshire by 8.8 points less. The ranking is almost exactly the ranking of GDP per head at the start of the period. A regression of relative income tax growth on the log of GDP per head in 2010 across the nine regions gives a slope of 27.5 (standard error 2.5) and an R-squared of 0.95: the richer a region was in 2010, the faster its income tax base grew afterwards.
This is not surprising. Income tax is progressive, and high earners are concentrated in London and its commuter belt. When wages at the top grow faster than wages in the middle, as they did in the 2010s, and when thresholds are frozen, as they have been since 2021, income tax receipts grow fastest where top earners live. A system that rewards regions for the growth of their income tax base will therefore reward London for the growth of top incomes, not for local economic policy.
4. Simulating Income Tax Assignment
To quantify what the 2028 reform could mean, the paper simulates a simple assignment rule. At a base year, each region receives a share of its income tax receipts, and its grant is reduced by the same amount. In later years the grant reduction grows with English income tax per head and with the region’s population. The region gains if its income tax per head grows faster than England’s and loses if it grows more slowly. Results are expressed for an assigned share of 10 per cent of income tax; they scale proportionally for other shares.

Table 3. Gain or loss in 2024-25 from assigning 10 per cent of income tax (pounds per head)
| Region | From 2010-11 | From 2014-15 | From 2018-19 |
|---|---|---|---|
| London | +94.9 | +76.0 | +20.8 |
| South East | -11.3 | -20.8 | -11.6 |
| East of England | -10.1 | -13.9 | -4.7 |
| South West | -15.2 | -8.2 | +8.1 |
| East Midlands | -29.4 | -17.7 | -2.7 |
| West Midlands | -21.9 | -11.3 | -10.5 |
| North West | -23.2 | -10.4 | +1.4 |
| Yorkshire and The Humber | -26.9 | -10.0 | +5.4 |
| North East | -30.2 | -10.6 | +4.3 |
Note: grant adjustment indexed to English income tax per head and to regional population. In millions of pounds, assignment from 2010-11 would have given London 863 million and cost the North West 179 million and Yorkshire 152 million in 2024-25.
Had income tax been assigned in 2010-11, every region except London would have lost money by 2024-25: between 10 and 30 pounds per head for each 10 per cent of income tax assigned, with the largest losses in the North East, the East Midlands and Yorkshire. London would have gained 95 pounds per head, or 863 million pounds a year. Starting in 2014-15 gives the same pattern. Starting in 2018-19, the picture is more mixed: the North East, Yorkshire and the South West would have made small gains, because their income tax bases recovered well from the pandemic, while the South East and West Midlands would have lost.
Two conclusions follow. First, over long periods income tax assignment with standard indexation transfers resources toward London, the region that needs them least. Second, the outcome depends heavily on the start date, which means that the choice of base year in 2028 will itself redistribute substantial sums.
5. The Risk Transferred to Regions
Under the current system, if a region’s income tax base grows slowly because of a local recession or the closure of a large employer, its grants are unaffected. Under assignment, the region bears the shortfall. The cost of this lost insurance can be measured by the variability of the gains and losses over shorter periods.

Table 4. Volatility of the regional income tax base
| Region | Std. dev. of relative annual growth (pp) | Worst three-year outcome (pounds per head) | Best three-year outcome (pounds per head) |
|---|---|---|---|
| North East | 2.7 | -13.0 | +16.4 |
| London | 2.3 | -19.7 | +36.0 |
| West Midlands | 2.0 | -14.6 | +8.6 |
| Yorkshire and The Humber | 2.0 | -11.2 | +11.0 |
| East Midlands | 2.0 | -12.0 | +12.6 |
| North West | 1.8 | -10.2 | +6.8 |
| South West | 1.4 | -9.7 | +12.5 |
| South East | 1.1 | -14.8 | +7.8 |
| East of England | 1.1 | -10.7 | +5.1 |
Note: relative growth of income tax per head against England, 2000-01 to 2024-25. Three-year outcomes per 10 per cent share.
The North East has the most volatile income tax base relative to England, with a standard deviation of 2.7 percentage points a year. For each 10 per cent of income tax assigned, its three-year outcome has ranged from a loss of 13 pounds per head to a gain of 16. For London the range is wider in pounds, from minus 20 to plus 36, but London’s budget is also larger. For a mayoral authority whose discretionary budget is small, swings of this size are significant. Unless the system includes a mechanism to share regional shocks, poorer regions would be exposed to risks they cannot diversify.
6. Business Rates: Small Stakes, Same Direction
The business rates element of the Roadmap is better defined: from April 2027 mayors will receive a fixed share in place of existing funds, uprated with inflation, plus 20 per cent of any growth in business rates in their area. The value of the growth share can be estimated from the growth of business rates since the last revaluation cycle.

The sums are modest. Twenty per cent of nominal business rates growth since 2017-18 would be worth about 12.6 pounds per head a year in London, 11.3 in the East Midlands, where logistics has expanded, and 10.6 in the East of England, but only 1.7 in Yorkshire and minus 1 in the North East, where the business rates base has shrunk. Business rates growth is driven by revaluations and by property values, which again favour the South and the logistics corridor of the Midlands. The growth incentive is real but small, and it does not favour the regions the reform is meant to help.
7. Convergence or Divergence?

The case for tax retention rests on the hope that local control will help lagging places catch up. The record since 1998 shows the opposite trend under the existing system. The dispersion of GDP per head across the nine English regions rose from 0.227 in log terms in 1998 to 0.260 in 2024; across the smaller ITL3 areas it rose from 0.390 to 0.440. A regression of real GDP growth per head on initial GDP per head across 132 English ITL3 areas finds no convergence: the coefficient is slightly positive (0.09, standard error 0.24) for 1998 to 2024, and positive in each sub-period.
Table 5. Sigma and beta convergence in England
| Measure | 1998 | 2008 | 2018 | 2024 |
|---|---|---|---|---|
| Std. dev. of log GDP per head, 9 regions | 0.227 | 0.239 | 0.255 | 0.260 |
| Std. dev. of log GDP per head, ITL3 areas | 0.390 | 0.405 | 0.437 | 0.440 |
| Beta coefficient, ITL3 real growth on initial level | 1998 to 2024: +0.09 (se 0.24) | 1998 to 2008: +0.19 (0.24) | 2008 to 2019: +0.22 (0.39) | 2019 to 2024: +0.19 (0.19) |
Note: a negative beta coefficient would indicate convergence.
In an economy where regions are diverging, a system that lets each region keep more of its own tax growth will amplify divergence unless it is offset by stronger equalisation. The results in Sections 3 and 4 show the mechanism: the regions whose tax bases grow fastest are those that are already richest.
8. Conditions for Success
Fiscal devolution can still serve its purpose if four conditions are met.
- Index grants to the right benchmark. Indexing the grant adjustment to English income tax per head, the standard approach, rewards regions with fast-growing top incomes. An alternative is to index to the region’s own past trend, so that mayors are rewarded only for improvements relative to their own trajectory.
- Build in equalisation. A share of income tax growth above the English average could be pooled and redistributed to regions growing below it, as in many federal systems.
- Share regional shocks. A corridor, within which regions bear gains and losses, with the centre absorbing shocks beyond it, would preserve incentives while limiting the risk transferred to regions with volatile tax bases.
- Keep shares small at first. The simulations scale with the assigned share. Starting with a small share, and a clear review after three years, limits the redistribution that the choice of base year would otherwise lock in.
9. Conclusions
England’s fiscal devolution experiment begins on an unequal map. London raises 3.4 times as much income tax per head as the North East, and its tax base has grown faster for fifteen years. Simulations of the assignment rules most likely to be used show that, over the long run, income tax retention would have moved money toward London and away from every other region, and would have exposed poorer regions to fiscal risks that central grants now absorb. Regional output in England has been diverging since 1998. If the 2028 income tax arrangement is designed without strong equalisation, it is more likely to entrench that divergence than to reverse it.
Methodological Appendix
A1. Data
Revenue and expenditure by region: ONS Country and regional public sector finances, financial year ending 2025 edition, tables of total public sector current receipts and total managed expenditure, 1999-2000 to 2024-25, in millions of pounds. Business rates are those paid by market-sector bodies plus private non-profit institutions. Population and GDP: ONS Regional gross domestic product, all ITL regions, 1998 to 2024. Financial year t-1 to t is matched with mid-year population of calendar year t-1.
A2. Assignment simulation
For region r with assignment starting in base year b, the gain in year t per share s is s x [IT(r,t) – IT(r,b) x (ITpc(E,t) / ITpc(E,b)) x (POP(r,t) / POP(r,b))], where IT is income tax, ITpc(E) income tax per head in England and POP population. This is a per-capita indexed deduction similar to the comparable model used for Scotland. The three-year windows apply the same formula with b running from 2000-01 to 2021-22 and t = b + 3.
A3. Convergence
Sigma convergence: cross-sectional standard deviation of log GDP per head at current market prices. Beta convergence: OLS regression of the annual average growth of real GDP per head (chained volume measures) on the log of initial real GDP per head across English ITL3 areas, with heteroskedasticity-robust standard errors.
A4. Limitations
The analysis uses the nine English regions, not the boundaries of mayoral strategic authorities, which are smaller and do not yet cover all of England. Regional allocation of income tax in the public sector finances is based on the residence of taxpayers and is subject to estimation. The simulations apply a stylised rule; the actual rules will be set at the Spending Review. The analysis does not model behavioural responses, such as the effect of local incentives on growth, which are the intended benefit of the reform.
References
HM Treasury (2026), The Fiscal Devolution Roadmap: Reforming the Funding Model for Regional and Local Government in England, October 2026.
Office for National Statistics (2026), Country and regional public sector finances: financial year ending 2025.
Office for National Statistics (2026), Regional gross domestic product: all ITL regions, 1998 to 2024.
Driss EFFINA, PhD
Independent Researcher, Global Strategy Files
The opening pages of our four country reports: United States, United Kingdom, Japan and China. A free PDF, sent to your inbox.
You will also receive occasional analysis from Global Strategy Files. Unsubscribe in one click, at any time.