Brexit was expected to impact the UK economy through two main channels: trade and immigration. There have been numerous attempts to estimate the impact of Brexit on trade (e.g. Freeman et al. 2025), but relatively few that look at immigration and the resulting impact on the UK labour market. Our new analysis (Portes and Springford 2026) uses synthetic differences-in-differences to construct new quantitative estimates.

Before and immediately after the 2016 referendum, a large literature anticipated that Brexit would reduce immigration overall. Ending free movement was expected to cut migration from the EU, partially offset by a modest increase in non-EU inflows under a more flexible visa regime. The Home Office impact assessment, for example, projected that the new points-based system would reduce work-related migration by around 40,000 per year (Home Office 2021). EU-origin migration did indeed decline after the referendum (Di Iasia and Wahba 2023), even though free movement remained in place until the end of the Brexit transition period in January 2021, when the new system, which broadly equalised rules for EU and non-EU origin migrants (except for Irish citizens), was introduced.

Migration then rose sharply (Portes 2024), with net migration peaking at over 900,000 in 2023, about four times its average level in the 2010s, before falling back sharply in 2025. However, interpreting this surge is complicated by the fact that migration to many other advanced economies also rose sharply after the end of the pandemic in early 2021, and the emergence of widespread labour shortages. Razin (2026), comparing the UK with Germany, shows that UK migration patterns diverged significantly from those in comparable EU economies following the referendum and the end of free movement, with much of the change driven by rising non-EU inflows.

Measuring Brexit’s impact on migrant employment

Our data are not immigration flows, which have been repeatedly revised in the UK and are not always collected on a comparable basis across countries. Nor do we rely on the UK Labour Force Survey, which has had well-publicised issues with its sampling of foreign-born residents. Instead, we use data on non-national employees from HMRC for the UK, and data on foreign-born employees from the EU Labour Force Survey data compiled by Eurostat. This is not perfect but, in our view, represents the most credible and reliable relevant data currently available. Focusing on migrants in the workforce also has the advantage that it is less likely than overall immigration to be affected by country-specific shocks to refugee and asylum flows.

We use synthetic difference-in-differences (SDID) (Archangelsky et al. 2021). SDID is similar to the synthetic control method (SCM) that has been used to estimate Brexit’s impact on GDP (Springford 2022). An algorithm finds a weighted basket of countries from a ‘donor pool’ (in our case, the EU15 and EEA) which, when combined together, minimise the difference in ‘pre-treatment’ trends, before the Brexit referendum in June 2016, just as with SCM. That basket of countries forms the control which can be used to measure divergence in trends after a policy intervention.

SDID extends this procedure by also estimating ‘time weights’, which prioritise quarters in the pre-treatment period when trends in the countries that make up the control are closest to trends in the post-treatment period. This is helpful in our case, because our employment data, which start in July 2014, give us only a short period of time before the 2016 referendum. Furthermore, transitional controls on free movement for Bulgarians and Romanians ended in 2014, which raised the rate of intra-EU migration. The SDID method allows us to form a control from the trends shortly before the referendum.

We estimate impacts separately for EU-origin and non-EU-origin workers. For EU workers, we treat the referendum in June 2016 as the start of the treatment period, reflecting both the immediate policy uncertainty and the depreciation of sterling that followed the vote. For non-EU workers, the relevant policy break occurs in January 2021 when the new immigration system was implemented. In practice, our results do not change materially if we take June 2016 as the start of the treatment period for both groups.

For EU-origin migrants, our results are shown in Figure 1. The countries with the largest weights in the control are Germany (11%), Austria (10%), and Finland, Ireland and Belgium (all about 9%). The figure shows that by 2024, the number of EU-origin workers in post-Brexit UK was roughly back to its level at the time of the referendum, while in the basket of comparable countries, the counterfactual UK, the number had risen by about 30%. This implies that Brexit had reduced the number of EU-origin employees by about 785,000 in 2024.

Figure 1 EU workers

If we repeat this process, but take the start date for our impact estimation to be January 2021, the results are very similar. It is worth noting that essentially all of the estimated impact materialises by about 2023. During 2024, both the actual UK and the counterfactual are broadly flat, meaning that our estimate suggests that the impact of Brexit on the number of EU-origin employees had largely played itself out by then.

For non-EU employees, the picture is quite different. Here the counterfactual shows a sharp rise after 2021, rising to over 150% of its 2016 level by the end of 2024; but the actual UK showed a much larger rise to about 225%. This implies an impact, looking at 2024 as whole, of about 992,000 employees. Basing the calculation in 2016 instead of 2021 slightly reduces the results. The largest country weights for this control are Switzerland (9%), France, Denmark, Italy, and the Netherlands (all about 8%).

Figure 2 Non-EU workers

As well as going in the opposite direction to that for EU workers, the profile of the impact is quite different. Instead of flattening off, it continues to grow, roughly in a linear fashion. The new system appears during this period to have led to a roughly constant increase in work-related migration flows, of about an extra 70,000 employees per quarter. Obviously, if it were to continue this would lead to a steadily growing impact on the number of non-EU origin employees in the UK workforce. However, we already know from both visa and immigration data that non-EU immigration to the UK for work slowed very sharply in 2025, because the visa regime was tightened and demand–supply mismatches in the British labour market eased A continued, linear rise in non-EU workers is highly unlikely to be the case – 2024 may reflect the peak for the impact of Brexit on non-EU origin employees.

Our estimate of the net effect of Brexit, then, is an increase in non-UK origin employees of approximately 207,000 in 2024, averaged over the calendar year; this is equivalent to about 0.62% of the overall workforce. However, this relatively small impact is the net effect of two large and mostly offsetting changes; a reduction in the number of EU-origin employees of about 775,000 (2.33% of the total workforce) and an increase in non-EU origin employees of about 992,000 (2.95% of the total workforce).

As visa issuance has been dropping rapidly, net migration to the UK is likely to fall further over 2026, with one forecast (Bowes 2025) suggesting it might be less than 100,000 in 2026. The dilemma facing all European governments – between maintaining employment as societies age and political pressure to reduce immigration – has not been resolved by Brexit, and our estimates show that the decision to leave the EU made only a modest difference to the number of foreign-born workers in Britain.

References

Arkhangelsky, D, S Athey, D Hirshberg, G Imbens, and S Wager (2021), “Synthetic Difference-in-Differences”, American Economic Review 111(12): 4088–4118.

Bowes, J (2025), “The coming collapse in immigration”, UK in a Changing Europe, October.

Di Iasio, V and J Wahba (2023), “Expecting Brexit and UK migration: Should I go?”, European Economic Review 157, 104484.

Freeman, R, M Garofalo, E Longoni, K Manova, R Mari, T Prayer and T Sampson (2025), “Deep Integration and Trade: UK Firms in the Wake of Brexit”, CEPR Discussion Paper No. 19869.

Home Office (2020), “Impact Assessment for Immigration and Social Security Co-ordination (EU Withdrawal) Bill 2020”, May.

Portes, J (2024) “Unintended consequences: the changing composition of immigration to the United Kingdom after Brexit”, National Institute Economic Review 268: 63–78. 

Portes, J and J Springford (2026), “The impact of Brexit on immigration to the UK”, Centre for European Reform, March.

Razin, A (2026), “Brexit migration shifts: Benchmarking the UK to Germany”, VoxEU.org, January.

Springford, J (2022), “The cost of Brexit to June 2022”, Centre for European Reform.