Three flows into and out of the UK, from 2010 where the sources allow: legal migration, asylum, and illegal entry routes. Each chart marks the June 2016 referendum and the end of EU free movement on 1 January 2021.
Shown, not ranked. Migration has no agreed better or worse direction, and much of the data covers the UK only. This page presents figures and published arguments. It does not rank the UK and does not say whether any level is too high or too low.
Terms. Illegal entry routes is the Home Office's current term (it titles its chapter this way) for people detected arriving in the UK without permission, mainly by small boat. Earlier releases called this irregular migration; both names describe the same group, and the series here is the same. Net migration is the number of people moving to the UK for a year or more, minus the number leaving for a year or more.
Every count has a rate per 1,000 people beside it, using World Bank population estimates for the UK in the year the figures end, so sizes can be read against countries of different populations. Charts plot each twelve-month total at the middle of its twelve months.
Flow 1
Legal migration
Net long-term migration by nationality group (ONS), and entry clearance visas granted by route (Home Office).
Net long-term migration, by nationality group
All nationalities
British
EU+
Non-EU+
Dotted: provisional or revised in the latest release
Shaded gap: ONS method change, from survey to administrative data
1Marker 1: EU referendum (23 June 2016)
2Marker 2: End of EU free movement (1 January 2021)
Method change. ONS estimates up to the year ending March 2021 used the International Passenger Survey for British nationals and administrative data for EU nationals. From the year ending June 2021 all groups use administrative data, and ONS advises caution when comparing estimates before and after. From that period the EU group is EU+, which adds Norway, Iceland, Liechtenstein and Switzerland. Periods from the year ending Mar 25 onward are provisional. Estimates from the year ending March 2025 onward are provisional: ONS completes them when 12 months of travel data are available, and revises them. Periods marked R have been revised.
Net long-term migration to the UK for years ending December, by nationality group, with a rate per 1,000 UK residents
Year ending
All nationalities
British
EU+
Non-EU+
All, per 1,000 people
Dec 2012
+195,000
−72,000
+206,000
+61,000
+3.06
Dec 2013
+244,000
−81,000
+262,000
+63,000
+3.80
Dec 2014
+284,000
−80,000
+277,000
+87,000
+4.40
Dec 2015
+303,000
−73,000
+287,000
+88,000
+4.66
Dec 2016
+249,000
−94,000
+253,000
+90,000
+3.80
Dec 2017
+208,000
−81,000
+176,000
+114,000
+3.15
Dec 2018
+276,000
−72,000
+170,000
+178,000
+4.16
Dec 2019
+184,000
−82,000
+80,000
+186,000
+2.76
Dec 2020
+93,000
−78,000
+70,000
+101,000
+1.39
Dec 2021
+467,000
−69,000
+15,000
+522,000
+6.97
Dec 2022
+891,000
−81,000
−9,000
+981,000
+13.17
Dec 2023
+848,000
−104,000
−53,000
+1,005,000
+12.38
Dec 2024revised
+331,000
−117,000
−63,000
+511,000
+4.78
Dec 2025provisional
+171,000
−136,000
−42,000
+350,000
+2.46
People, rounded by ONS to the nearest thousand. Components may not add to the total. Before the year ending June 2021 the EU column is EU nationals only; after, EU+ . Net migration is arrivals minus departures, so a negative figure means more people left than arrived.
Sources
ONS, Long-term international migration, provisional. Net migration, immigration and emigration by nationality group. Table: Table 1 (headline cross-check against Table 2). Release: year ending december 2025, published 21 May 2026. Fetched 2026-11-06. Licence: Open Government Licence v3.0. Data file.
World Bank, Population, total (SP.POP.TOTL). Per-1,000-people rates. Table: SP.POP.TOTL. Release: last updated 2026-10-08. Fetched 2026-11-06. Licence: Creative Commons Attribution 4.0.
Visas granted, by route
Work
Study
Family
Dotted: provisional or revised in the latest release
1Marker 1: EU referendum (23 June 2016)
2Marker 2: End of EU free movement (1 January 2021)
Entry clearance visas granted for work, study and family by year, with a rate per 1,000 UK residents
Period
Work
per 1,000
Study
per 1,000
Family
per 1,000
2010
160,724
2.56
334,735
5.33
52,361
0.83
2011
149,310
2.36
323,276
5.11
45,041
0.71
2012
145,110
2.28
278,100
4.37
40,231
0.63
2013
154,760
2.41
296,226
4.62
33,133
0.52
2014
167,146
2.59
293,648
4.54
34,852
0.54
2015
165,900
2.55
272,919
4.19
37,711
0.58
2016
163,922
2.50
294,454
4.49
37,781
0.58
2017
164,943
2.50
330,101
5.00
40,123
0.61
2018
175,812
2.65
354,878
5.35
43,620
0.66
2019
192,559
2.89
404,410
6.07
55,643
0.84
2020
114,528
1.72
250,121
3.75
37,087
0.56
2021
239,193
3.57
435,110
6.50
42,182
0.63
2022
421,565
6.23
623,698
9.22
47,119
0.70
2023
613,627
8.96
604,253
8.82
80,083
1.17
2024
368,139
5.31
418,932
6.05
85,161
1.23
2025
260,430
3.75
427,340
6.15
65,971
0.95
Year ending June 2026provisional
234,841
3.38
385,974
5.56
57,590
0.83
Visas granted include dependants and exclude visitor, transit and other visas. Entry clearance visa data from the first quarter shown as provisional are provisional (Home Office note 1). The latest row is the twelve months to 2026 Q2; the Home Office marks data from 2026 Q1 onward as provisional.
Caveats
ONS moved from the International Passenger Survey to administrative data. Estimates from the year ending June 2021 are not directly comparable with earlier ones, and the chart breaks the line there.
ONS estimates for the latest periods are provisional and are revised as travel data arrive. Figures are rounded to the nearest thousand.
Visas granted count permission to enter, not arrivals, and include dependants. They are not the same measure as ONS net migration.
Inflows per 1,000 people for EU states from the OECD International Migration Database are not included on this page, so legal migration is shown for the UK only.
Sources
Home Office, Immigration system statistics: entry clearance visas. Visas granted by route group: work, study, family. Table: Vis_D02 (outcomes, issued); checked against Vis_01. Release: year ending june 2026, published 27 August 2026. Fetched 2026-11-06. Licence: Open Government Licence v3.0. Data file.
World Bank, Population, total (SP.POP.TOTL). Per-1,000-people rates. Table: SP.POP.TOTL. Release: last updated 2026-10-08. Fetched 2026-11-06. Licence: Creative Commons Attribution 4.0.
Flow 2
Asylum
Asylum applications and initial grant rate for the UK and the 27 EU states, from UNHCR. Context, not ranked.
Applications per 1,000 people
UK
Median of the EU27
1Marker 1: EU referendum (23 June 2016)
2Marker 2: End of EU free movement (1 January 2021)
Initial grant rate
UK
Median of the EU27
1Marker 1: EU referendum (23 June 2016)
2Marker 2: End of EU free movement (1 January 2021)
The EU27 line is the median of the 27 countries' own rates, not a pooled total. A year is plotted only where at least 24 countries have a value.
All 28 countries, 2025
Listed from most to fewest applications per 1,000 people. The order is a way to find a country, not a judgement: neither a higher nor a lower figure is better. 28 of 28 countries have data for 2025.
Asylum applications, rate per 1,000 people, first-instance decisions and initial grant rate for 28 countries in 2025. Context, not a ranking.
Order
Country
Applications
per 1,000 people
Decisions (granted or refused)
Initial grant rate
Grant-rate order
1
Greece
55,361
5.32
41,663
64%
2
2
Spain
144,387
2.93
143,577
52%
7
3
Luxembourg
1,655
2.41
1,142
55%
6
4
Ireland
13,144
2.40
19,176
18%
24
5
Belgium
27,551
2.31
27,614
32%
19
6
Italy
126,621
2.15
104,547
17%
25
7
Cyprus
2,813
2.05
4,492
15%
26
8
Slovenia
3,944
1.85
204
37%
16
9
France
111,343
1.62
154,213
41%
13
10
United KingdomUK
100,620
1.45
134,583
41%
14
11
Austria
12,493
1.36
18,622
63%
5
12
Germany
113,216
1.36
196,255
28%
21
13
Netherlands
23,949
1.32
13,963
51%
8
14
Latvia
1,462
0.79
485
30%
20
15
Estonia
994
0.73
805
93%
1
16
Malta
388
0.67
488
35%
17
17
Bulgaria
3,876
0.60
4,638
21%
23
18
Sweden
4,961
0.47
4,852
42%
12
19
Poland
13,196
0.36
7,543
48%
10
20
Finland
2,031
0.36
3,411
51%
9
21
Croatia
1,229
0.32
279
7%
27
22
Denmark
1,862
0.31
1,060
38%
15
23
Portugal
1,755
0.16
1,380
33%
18
24
Lithuania
324
0.11
342
48%
11
25
Czechia
911
0.08
750
26%
22
26
Romania
1,144
0.06
981
64%
3
27
Slovakia
142
0.03
111
64%
4
28
Hungary
101
0.01
28
– fewer than 100 decisions
–
Applications are new first-instance claims. Grant rate is recognised plus other protection, divided by granted plus refused, at first instance; a dash means fewer than 100 such decisions. Initial grant rate is the share of first-instance asylum decisions that recognised refugee status or gave other protection, among decisions that granted or refused. Closed cases are left out.
Caveats
Countries record asylum claims and decisions differently. UNHCR reports what each government supplies, so cross-country gaps partly reflect procedure and definitions.
Grant rates depend on who applies, which differs by country, and are not shown where a country made fewer than 100 first-instance decisions. A single year can move a small country's figure sharply.
This table is context, not a ranking. No order is better than another, and the order is only a way to find a country.
Sources
UNHCR Refugee Data Finder (asylum applications and decisions). Asylum applications and initial grant rate, 28 countries. Table: asylum-applications and asylum-decisions endpoints. Release: data to 2025, as served on 2026-11-06. Fetched 2026-11-06. Licence: Creative Commons Attribution 4.0 (UNHCR Refugee Population Statistics Database).
World Bank, Population, total (SP.POP.TOTL). Per-1,000-people rates. Table: SP.POP.TOTL. Release: last updated 2026-10-08. Fetched 2026-11-06. Licence: Creative Commons Attribution 4.0.
Flow 3
Illegal entry routes
Detected arrivals and returns, from Home Office immigration system statistics. Illegal entry routes is the Home Office's current term (it titles its chapter this way) for people detected arriving in the UK without permission, mainly by small boat. Earlier releases called this irregular migration; both names describe the same group, and the series here is the same.
Small boat arrivals and other detected entries
Small boat arrivals
Other detected entries
Dotted: provisional or revised in the latest release
1Marker 1: EU referendum (23 June 2016)
2Marker 2: End of EU free movement (1 January 2021)
The Home Office series starts in 2018, so the June 2016 marker has no data beside it.
Detected arrivals by illegal entry routes by year, with a rate per 1,000 UK residents
Period
Small boat arrivals
per 1,000
Other detected entries
per 1,000
All detected
per 1,000
2018
299
0.01
13,078
0.20
13,377
0.20
2019
1,843
0.03
14,688
0.22
16,531
0.25
2020
8,466
0.13
8,564
0.13
17,030
0.26
2021
28,526
0.43
8,259
0.12
36,785
0.55
2022
45,774
0.68
8,800
0.13
54,574
0.81
2023
29,437
0.43
6,926
0.10
36,363
0.53
2024
36,816
0.53
6,474
0.09
43,290
0.63
2025revised
41,472
0.60
5,036
0.07
46,508
0.67
Year ending June 2026revised
33,374
0.48
4,712
0.07
38,086
0.55
Quarterly Home Office data summed to calendar years; the last row is the latest four quarters. Other detected entries are inadequately documented air arrivals plus recorded detections in the UK and at UK ports. Small boat and air arrival data from January 2025 has been revised, and these figures come from a live operational database that can change (Home Office notes 11, 15 and 16).
Returns from the UK
Enforced returns
Voluntary returns
Refused entry, then departed
Dotted: provisional or revised in the latest release
1Marker 1: EU referendum (23 June 2016)
2Marker 2: End of EU free movement (1 January 2021)
Returns from the UK by type and year, with a rate per 1,000 UK residents
Period
Enforced
per 1,000
Voluntary
per 1,000
Refused entry, then departed
per 1,000
2010
13,928
0.22
31,762
0.51
18,446
0.29
2011
14,566
0.23
30,088
0.48
15,969
0.25
2012
15,134
0.24
31,002
0.49
14,045
0.22
2013
14,850
0.23
31,694
0.49
14,781
0.23
2014
13,541
0.21
27,815
0.43
16,559
0.26
2015
12,921
0.20
30,210
0.46
18,082
0.28
2016
11,903
0.18
28,474
0.43
17,916
0.27
2017
11,741
0.18
20,979
0.32
18,520
0.28
2018
9,236
0.14
15,702
0.24
19,655
0.30
2019
7,198
0.11
12,574
0.19
18,464
0.28
2020
3,383
0.05
5,009
0.08
10,136
0.15
2021
2,788
0.04
7,352
0.11
18,769
0.28
2022
3,806
0.06
11,014
0.16
23,531
0.35
2023
6,362
0.09
20,475
0.30
24,742
0.36
2024provisional
8,171
0.12
26,851
0.39
23,193
0.34
2025provisional
9,908
0.14
29,807
0.43
18,354
0.26
Year ending June 2026provisional
9,669
0.14
30,936
0.45
18,646
0.27
Returns cover everyone returned, not only people who arrived by illegal entry routes. The Home Office marks 2024 onward as provisional.
Caveats
Detections are of people, from Home Office management information that is not designed for statistical use. They do not show how many people entered without being detected.
The series starts in 2018. Recent quarters have been revised and can change again. Counts for other detected entries use different coverage from small boat arrivals and should not be added up to judge a total.
No per-country comparison is shown: EU border data count external-border crossings for the bloc as a whole, not entries into each member state.
Sources
Home Office, Immigration system statistics: illegal entry routes. Small boat arrivals and other detected entries, quarterly summed to years. Table: IER_D01 (detected arrivals by method of entry); checked against IER_01. Release: year ending june 2026, published 27 August 2026. Fetched 2026-11-06. Licence: Open Government Licence v3.0. Data file.
Home Office, Immigration system statistics: returns. Enforced returns, voluntary returns and refused-entry departures. Table: Ret_D02 (returns by destination and type); checked against Ret_01. Release: year ending june 2026, published 27 August 2026. Fetched 2026-11-06. Licence: Open Government Licence v3.0. Data file.
World Bank, Population, total (SP.POP.TOTL). Per-1,000-people rates. Table: SP.POP.TOTL. Release: last updated 2026-10-08. Fetched 2026-11-06. Licence: Creative Commons Attribution 4.0.
Open questions
Debated questions
Three claims people dispute. Each shows arguments from more than one side, with sources. None is settled here.
Did leaving the EU's Dublin returns system affect Channel crossings?
The Dublin III Regulation let a state ask another European state to take back some people who had claimed asylum there first. It stopped applying to the UK at the end of the transition period on 31 December 2020. Small boat arrivals in the Home Office series begin in 2018. See the small boat arrivals chart above for the timing.
Arguments that it may have contributed
After 31 December 2020 the UK could no longer use Dublin to request the return of some claimants to the European state responsible for them, or to join family members through it.
The Migration Observatory reports that migrants in northern France often cite the UK's exit from Dublin as a reason to seek asylum in the UK, particularly if they had already been refused protection in the EU.
Arguments that other factors matter more, or the effect is unclear
The Migration Observatory says it is unclear how much leaving Dublin shaped decisions, because few people were returned under it while the UK was a member (about 560 a year on average from 2008 to 2020).
The same briefing points to tighter enforcement on lorry and rail routes, larger smuggling groups, and global factors such as conflict, and says global factors appear more important than small policy differences.
Small boat arrivals rose from 2018, before the transition period ended. Fact checkers and parliamentary reports describe only a handful of crossings a year before 2018.
This page does not settle the question. The data above show timing and size; they do not show cause.
How far did non-EU migration after 2021 replace EU migration?
ONS estimates show net migration of EU+ nationals falling and turning negative after 2021, while net migration of non-EU+ nationals rose. Whether one replaced the other is a separate question from whether the two moved at the same time. See the net migration chart above.
Arguments that non-EU migration largely took its place
EU net migration has been negative since 2021 on official estimates, while non-EU nationals have made up most of net migration since.
Work routes for non-EU workers were widened as free movement ended, and the Migration Advisory Committee says employers should recruit more from overseas when the labour market is tight.
The Migration Observatory attributes much of the rise in non-EU visas to bespoke humanitarian schemes (Ukraine, British National (Overseas)) and to the return of international students, not to routes designed to replace free movement.
ONS publishes the reasons non-EU+ nationals gave for migrating (work, study, family, humanitarian, asylum), which show how much came by each route.
The Migration Advisory Committee's 2022 report says new routes can temporarily raise net migration and expected much of the recent rise to be temporary.
This page does not settle the question. The data above show timing and size; they do not show cause.
What did the end of free movement do to labour shortages?
Vacancies reached record highs in 2022 while free movement had ended, the pandemic had changed the workforce, and demand was recovering. See the visas granted chart above for the work route.
Arguments that it contributed to shortages
The Migration Advisory Committee treats the end of free movement as a labour supply shock, mainly for lower-skilled jobs in hospitality, logistics, manufacturing, agriculture and administration, and says it may have contributed to higher vacancies.
In 2021 the Committee recommended making care worker jobs eligible for the Health and Care visa, citing severe recruitment and retention difficulties in the sector.
The same Committee report notes that vacancy rises were similar across other developed countries, so ending free movement is unlikely to be the only cause, and names pent-up post-pandemic demand as a likely driver.
It also points to higher economic inactivity, concentrated among people aged 50 to 64, and to foreign workers who left during the pandemic.
The Committee cautions against reopening low-wage routes, which it says would in effect undo the end of free movement for those sectors, and says work routes should flex with the labour market.
This page does not settle the question. The data above show timing and size; they do not show cause.
Data checks on this release
Before any figure is published, the pipeline checks each spreadsheet's column headers, row counts, missing quarters, totals against the release's own published figures, and quarter-on-quarter jumps. If any check fails, the run stops and the previous figures stay. 24 checks passed on 6 November 2026.
List the checks
Column headers of Data_Vis_D02 match the expected layout (10 compared)
Data_Vis_D02 has 407,284 rows (floor 300,000) (407,284 compared)
Every quarter from 2005 Q1 to 2026 Q2 is present (86 compared)
No quarter-on-quarter move above 3x (largest 2.44x) (85 compared)
Quarterly totals match the grand total row of Vis_D02 (8 compared)
Annual and rolling-year figures match the release's summary table Vis_01 (65 compared)
Column headers of Data_IER_D01 match the expected layout (8 compared)
Data_IER_D01 has 15,762 rows (floor 10,000) (15,762 compared)
Every quarter from 2018 Q1 to 2026 Q2 is present (34 compared)
No quarter-on-quarter move above 3.5x (largest 2.67x) (33 compared)
Quarterly totals match the grand total row of IER_D01 (8 compared)
Annual and rolling-year figures match the release's summary table IER_01 (36 compared)
Column headers of Data_Ret_D02 match the expected layout (7 compared)
Data_Ret_D02 has 49,486 rows (floor 40,000) (49,486 compared)
Every quarter from 2004 Q1 to 2026 Q2 is present (90 compared)
No quarter-on-quarter move above 3x (largest 1.54x) (89 compared)
Annual and rolling-year figures match the release's summary table Ret_01 (69 compared)
ONS Table 1 column headers match the expected layout (6 compared)
ONS Table 1 has 55 periods for each of three flows (165 compared)
All nationalities equals British + EU+ + non-EU+, and net equals immigration minus emigration, within rounding (3,000) (385 compared)
Table 1 all-nationality estimates equal the headline estimates in Table 2 (24 compared)
No period-on-period move above 2.5x in immigration or emigration (the method change, YE Jun 21, is exempt) (432 compared)
UNHCR response totals equal the sum of their rows (applications and decisions) (2 compared)
UNHCR: 28 of 28 countries have applications for 2025 (minimum 25) (28 compared)
What this site can't tell you
Trends show correlation, not causation. Policy, conflict, the pandemic and the economy overlap the same years, and these charts do not separate their effects.
Sovereignty, control of borders, public services and views on how many people should come are not measured here. Migration policy is one of the non-economic factors people weigh.
Counts of detected people are not counts of everyone who arrived, and estimates for recent months are revised.