← Brexit

Brexit · Investment and productivity · United Kingdom · 2016–2026

What is observed, and what is modelled relative to remaining?

The official forecaster assumes UK productivity will settle 4% below where it would have been inside the EU; a recent academic study estimates GDP was 6% to 8% lower by 2025. Both compare the UK with a path it did not take, so neither is an observed loss — and observed growth since 2016 does not disprove them.

Evidence last reviewed 3 October 2026

The estimates, side by side

These estimates use different measures, periods and methods, so they are listed rather than averaged. Each is a comparison with an unobserved UK that stayed in the EU. They agree on the direction — a cost — and differ on size.

Modelled estimates relative to remaining in the EU (not observations)
WhoMeasureEstimateMethodStatus
Office for Budget ResponsibilityLong-run productivity; trade intensityProductivity 4% lower and exports and imports about 15% lower in the long run than if the UK had remained; about two-fifths of the productivity effect already by 2021, through weaker investment.An assumption in the official fiscal forecast, drawn from external studies; reviewed against the data in March 2022 and March 2024 and left unchanged.Official forecast assumption
Bloom, Bunn, Mizen, Smietanka, Thwaites and Abrahams (NBER)GDP, investment, employment and productivity, by the end of 2025GDP 6% to 8% lower; investment 12% to 13%; employment 3% to 4%; productivity 3% to 4% lower than without Brexit.Combines estimates from macro data with micro data from the Decision Maker Panel survey of UK firms. A working paper: circulated for discussion, not yet peer reviewed.Independent academic estimate (range)
Freeman, Manova, Prayer and Sampson (CEP, LSE)UK–EU trade relative to trade with the rest of the worldA sudden, persistent 25% fall in imports from the EU relative to other partners after January 2021; a smaller, temporary relative fall in exports; many small export relationships ended.Compares UK trade with the EU and with other countries before and after the trade deal, using detailed trade data.Independent academic estimate

How the OBR’s assumption works

Modelled scenario relative to remaining in the EU — a sequence in the OBR’s assumption, not observed data or measured effect sizes

  1. 1

    2016–2020: uncertainty

    The OBR judges that about two-fifths of its 4% productivity effect had occurred by the time the trade deal took effect, as uncertainty weighed on investment.

  2. 2

    2021: new trade barriers

    Leaving the single market and customs union adds customs, rules-of-origin and regulatory checks to UK–EU trade.

  3. 3

    Lower trade intensity

    Exports and imports assumed about 15% lower in the long run than if the UK had remained — judged broadly on track in March 2024.

  4. 4

    Long run: lower productivity

    Productivity assumed 4% lower than in the remain path. The gap is between two paths, not a fall from 2016 levels.

Source: Office for Budget Responsibility, Brexit analysis (page last updated 20 July 2026). Equal-sized steps show sequence only.

Three readings of the 4% are wrong. It does not mean GDP is 4% lower than in 2016 — GDP has grown since. It does not mean every region, sector or household is 4% worse off; it is an economy-wide average. And it does not mean the whole effect has already happened. The opposite reading is wrong too: growth since 2016 says nothing about how much growth there would have been inside the EU.

What investment actually did

What has happened to business investment since the referendum?

Business investment is up 14% on the referendum quarter — the modelled question is how much higher it would have been

UK business investment, chained volume measure, seasonally adjusted, indexed to 2016 Q2 = 100, 2000 Q1 to 2026 Q2.

United Kingdom · 2000 Q1 to 2026 Q2

1996–2026 displayed window · observations only. Shading marks years outside the displayed record.

Volume index, 2016 Q2 = 100No observations shown406080100120axis does not start at 02000201020202016 Q2 (referendum quarter)114
Investment barely grew for three years after the vote, fell sharply in the pandemic and has since recovered to 14% above its 2016 Q2 level. Growth since 2016 does not rule out a Brexit effect: Bloom and co-authors estimate investment is 12% to 13% lower than it would otherwise have been. That is a comparison with a path that cannot be observed, and it uses survey evidence from firms as well as macro data.

United Kingdom · 2000 Q1 to 2026 Q2 · Accredited official statistics; recent quarters are revised often and substantially

Sources, definitions and data — Business investment is up 14% on the referendum quarter — the modelled question is how much higher it would have been
Geography:
United Kingdom — Business investment (private and public corporations; excludes government and dwellings)
Measure:
Volume index, 2016 Q2 = 100
Period:
2000 Q1 to 2026 Q2
Status:
Accredited official statistics; recent quarters are revised often and substantially
Calculation:
Site calculation: each quarter divided by 2016 Q2 × 100.
Compare with:
2016 Q2, the quarter of the referendum
Not included / breaks:
This is the observed path only. Estimates that investment is lower than it would have been compare it with an unobserved path and are not shown as a line.

Source: ONS, Business investment, chained volume measure (NPEL), CDID NPEL, quarterly national accounts. Chart BREXIT-05.

View the data as a table
QuarterIndex (2016 Q2 = 100)
2000 Q162.5
2000 Q267.2
2000 Q361.0
2000 Q463.8
2001 Q161.1
2001 Q266.4
2001 Q365.5
2001 Q462.8
2002 Q160.8
2002 Q260.4
2002 Q364.0
2002 Q466.3
2003 Q166.7
2003 Q261.2
2003 Q361.3
2003 Q460.8
2004 Q162.6
2004 Q254.0
2004 Q359.3
2004 Q462.6
2005 Q164.2
2005 Q288.1
2005 Q371.1
2005 Q470.7
2006 Q172.7
2006 Q269.7
2006 Q372.5
2006 Q469.2
2007 Q173.6
2007 Q277.5
2007 Q376.1
2007 Q481.0
2008 Q174.7
2008 Q279.5
2008 Q376.3
2008 Q472.8
2009 Q169.2
2009 Q265.2
2009 Q363.7
2009 Q463.9
2010 Q167.2
2010 Q264.8
2010 Q367.2
2010 Q471.7
2011 Q167.1
2011 Q270.1
2011 Q372.1
2011 Q473.9
2012 Q176.6
2012 Q272.7
2012 Q373.7
2012 Q478.8
2013 Q176.4
2013 Q278.7
2013 Q381.7
2013 Q480.7
2014 Q181.8
2014 Q283.9
2014 Q384.9
2014 Q487.9
2015 Q192.0
2015 Q291.5
2015 Q390.4
2015 Q495.8
2016 Q196.0
2016 Q2100.0
2016 Q3104.6
2016 Q499.3
2017 Q1100.9
2017 Q299.3
2017 Q3100.4
2017 Q4102.4
2018 Q1102.8
2018 Q299.4
2018 Q397.1
2018 Q498.5
2019 Q1101.4
2019 Q2102.0
2019 Q3101.8
2019 Q4101.1
2020 Q199.7
2020 Q279.1
2020 Q392.0
2020 Q497.9
2021 Q193.5
2021 Q298.1
2021 Q397.9
2021 Q4101.3
2022 Q1103.9
2022 Q2105.0
2022 Q3104.1
2022 Q4102.5
2023 Q1105.7
2023 Q2105.0
2023 Q3102.6
2023 Q4103.5
2024 Q1104.0
2024 Q2106.0
2024 Q3106.6
2024 Q4105.0
2025 Q1110.0
2025 Q2108.4
2025 Q3111.1
2025 Q4108.7
2026 Q1112.0
2026 Q2114.0

What the data cannot settle

How much of the goods-trade weakness is Brexit? Not precisely. The pandemic, the energy-price shock and the war in Ukraine fall in the same years, and exports to non-EU countries fell too. Studies that compare EU with non-EU trade attribute a substantial share to Brexit; none isolates it perfectly.

Would investment have been stronger inside the EU? Both the OBR and the NBER study say yes, and much of the effect they estimate runs through investment. That is a judgement about an unobserved path, supported in the NBER case by firms’ own survey answers.

Do the new trade deals offset it? On the government’s own impact assessments, the Australia, New Zealand and CPTPP agreements each add less than 0.1% to GDP in the long run — far smaller than the estimated EU effect. The trade estimates.

Options that could reduce trade costs within the present relationship are compared in What can change within the present relationship? For the wider causes of slow growth, see Growing the economy.

Sources and methods (4)

Method notes

  • Modelled estimates are shown in a table and a schematic, never as a line on a chart of observed data.
  • Business investment is a chained volume measure from the quarterly national accounts; recent quarters are often revised.