The local economic cost of Brexit.
A synthetic control study of how Brexit reshaped output and income across every UK region, city region, and local authority.
What this project does
The Brexit Cost Explorer compares observed local outcomes against a synthetic-control counterfactual for each place. The core question is not whether a place grew or shrank in absolute terms, but whether it diverged from a plausible path it might have followed without Brexit.
The estimates are built from a placebo-weighted ensemble of 31 donor-pool combinations — drawn from European NUTS2 regions, US states, G20 countries, OECD members, and EU member states — scored on out-of-sample predictive performance. Two outcome series are tracked for every place: real Gross Value Added (GVA, a production measure) and nominal Gross Disposable Household Income (GDHI, a residence-based welfare measure). The table below summarises post-2016 and post-2020 GVA losses for the UK, its constituent countries, and English ITL1 regions.
Estimated Brexit losses by region
GVA losses as a share of observed output, in GBP billions, and per capita — for the United Kingdom, constituent countries, and ITL1 English regions. The post-2016 window captures the full period from the referendum vote; the post-2020 window isolates the period after legal exit. All figures are as of 2023, the latest year of ONS sub-national annual GVA data.
| Region | Post-2016 GVA (real) | Post-2020 GVA (real) | ||||
|---|---|---|---|---|---|---|
| Loss (£bn) | Gap (%) | Per capita | Loss (£bn) | Gap (%) | Per capita | |
| United Kingdom (2023) | −£188.8bn | −7.7% | −£2,835 | −£84.4bn | −3.6% | −£1,257 |
| Constituent Countries | ||||||
| Scotland | −£15.8bn | −8.6% | −£2,908 | −£7.6bn | −4.3% | −£1,386 |
| Wales | −£3.4bn | −4.4% | −£1,101 | −£7.3bn | −8.9% | −£2,338 |
| Northern Ireland | +£0.3bn | +0.5% | +£147 | −£0.9bn | −1.8% | −£490 |
| English Regions | ||||||
| North East | −£5.3bn | −7.8% | −£2,024 | −£2.1bn | −3.2% | −£797 |
| North West | +£1.1bn | +0.5% | +£160 | +£4.7bn | +2.2% | +£681 |
| Yorkshire & The Humber | −£8.5bn | −5.3% | −£1,752 | −£5.1bn | −3.3% | −£1,053 |
| East Midlands | −£6.8bn | −5.0% | −£1,667 | −£4.0bn | −3.0% | −£984 |
| West Midlands | −£11.8bn | −6.8% | −£1,999 | −£12.6bn | −7.3% | −£2,110 |
| East of England | −£8.4bn | −4.2% | −£1,337 | −£9.2bn | −4.7% | −£1,456 |
| London | −£70.0bn | −11.6% | −£7,927 | −£34.6bn | −6.1% | −£3,916 |
| South East | −£46.0bn | −12.4% | −£4,997 | −£19.8bn | −5.7% | −£2,131 |
| South West | −£12.1bn | −6.6% | −£2,385 | −£4.5bn | −2.5% | −£869 |
Source: Alabrese, Edenhofer, Fetzer & Wang (2026), Table 1. Estimates use the placebo-weighted ensemble across 31 donor-pool combinations. GVA is real chained-volume (2019 prices). Loss in £bn is computed as actual × (1 − exp(−gap)). Per-capita loss in £. All figures are as of 2023 (latest ONS annual release). Northern Ireland's positive post-2016 GVA outcome reflects its effective continuation in the EU customs union under the Trade and Cooperation Agreement. Losses are concentrated in the most trade-integrated regions: London, the South East, and Scotland.
How to read the estimates
These numbers are estimated gaps relative to a modelled counterfactual, not directly observed losses on a public balance sheet. They should be read as evidence about divergence after 2016, with extra caution around nominal household-income measures. The site separates project-level method from place-level evidence so users can move from the national framing to a local report without losing context.
Photo credits (left to right, top to bottom): Stevekeiretsu / CC BY-SA 4.0 · © User:Colin / Wikimedia Commons / CC BY-SA 4.0 · Bob Harvey / CC BY-SA 2.0 · Jeremy Segrott / CC BY 2.0 · Gwydion M. Williams / CC BY 2.0 · ChiralJon / CC BY 2.0. All images via Wikimedia Commons.