16 July, 2018
1. A comparison of ONS population estimates with cumulative net migration totals indicates that net migration from the European Union may have been undercounted by as much as 40,000 per year since 2004 while non-EU net migration may have been overcounted by as much 46,000 per year. This suggests that Brexit may represent a greater opportunity to deliver a substantial reduction in net migration than previously thought.
2. The Annual Population Survey (APS) measures the UK population by country of birth (and does not include people who live in ‘communal establishments’) while Long-term International Migration (LTIM) figures record movement by citizenship and are mainly based on the International Passenger Survey (IPS). The IPS has been running continuously since 1961 and collects information, 700,000 and 800,000 interviews a year, about passengers entering and leaving the UK.
3. The ONS advise against comparing datasets which measure different things. However, comparing the change in the size of the population of EU and non-EU migrants with cumulative net migration statistics produces interesting results.
4. Between 2005 and 2017, the population of EU born in the UK (as measured by the APS) increased by 2.21 million (see Annex for full table).
5. Cumulative EU net migration meanwhile (as measured by LTIM/IPS) totalled 1.7 million between 2004 and 2016. This included the 315,000 undercount of EU net migrants for the period 2004-2008 which was identified by the 2011 Census (of which 250,000 was attributed to EU8 and 65,000 to migrants from the Republic of Ireland).
6. Thus the change in the population of EU born in the UK is 508,000 higher than cumulative net migration of EU citizens. This, averaged out over the 13 years is 39,000 per year. So either the LTIM statistics have undercounted EU net migration or the APS is overestimating EU born residents in the UK.
7. The population of non-EU born in the UK (as measured by the APS) has increased by 1.94 million between 2005 and 2017 yet net migration of non-EU citizens has totalled 2.54 million between 2004 and 2016, 600,000 higher than the change in the population of non-EU born residents in the UK. This is an average of 46,000 a year.
8. Comparing the APS and the LTIM for EU and non-EU migration shows that the total non-UK born population has increased by 4.15 million while net migration of non-UK citizens totalled 4.24 million. This may confirm the view of the ONS that overall net migration is a broadly sound estimate (assuming British immigration and emigration are being properly recorded).
9. It appears possible that an overcount of non-EU migration (possibly being driven by an undercount of student departures) may have been masked by an undercount of EU migration.
10. Professor Jonathan Portes of King’s College London recently stated the view that an underestimate of EU net migration is likely, noting that ‘the volume of National Insurance numbers being issued to EU citizens suggested their numbers were still being undercounted [and the] data seems to bear this out’. [1]
11. The IPS was improved in 2009 to better capture EU migration at the borders. However, the adjustment methodology is now a decade old and the ONS is looking at it again. It is possible that the adjustment methodology is undercounting the number of EU migrants who enter the country as visitors (i.e. for less than 12 months) but who end up staying for longer than a year. These people would be absent from the migration figures but present in the population statistics.
12. Post-Brexit policy should aim to reduce immigration into lower-skilled work which is less beneficial to the economy while adding to population growth and to pressure on public services. This analysis suggests that Brexit may provide greater scope for the government to deliver a substantial reduction in overall net migration - a goal supported by nearly two-thirds of the public.
Table 1. Change in UK population by country of birth, 2005 to 2017; IPS/LTIM net migration, 2004 to 2016 (thousands).
Year | Non-EU APS | Change in non-EU APS | Non-EU Net migration | EU APS | Change in EU APS | EU net migration | Non-UK APS | Change in non-UK APS | Non-UK net migration |
---|---|---|---|---|---|---|---|---|---|
2004 | 3741 | 266 | 1492 | 87 | 5233 | 352 | |||
2005 | 3975 | 234 | 198 | 1577 | 85 | 96 | 5552 | 319 | 294 |
2006 | 4228 | 253 | 218 | 1769 | 192 | 104 | 5997 | 445 | 322 |
2007 | 4374 | 145 | 204 | 1969 | 200 | 127 | 6342 | 345 | 330 |
2008 | 4568 | 195 | 187 | 2115 | 146 | 63 | 6683 | 341 | 250 |
2009 | 4727 | 159 | 184 | 2183 | 68 | 58 | 6910 | 227 | 242 |
2010 | 4856 | 129 | 217 | 2283 | 100 | 77 | 7139 | 229 | 294 |
2011 | 4992 | 136 | 204 | 2517 | 234 | 82 | 7509 | 370 | 286 |
2012 | 5071 | 79 | 157 | 2609 | 92 | 82 | 7679 | 170 | 239 |
2013 | 5106 | 35 | 142 | 2674 | 65 | 123 | 7780 | 101 | 266 |
2014 | 5252 | 146 | 194 | 3025 | 351 | 174 | 8277 | 497 | 368 |
2015 | 5387 | 135 | 189 | 3183 | 158 | 184 | 8569 | 292 | 372 |
2016 | 5616 | 229 | 175 | 3537 | 354 | 133 | 9152 | 583 | 308 |
2017 | 5677 | 61 | 3705 | 168 | 9382 | 230 | |||
Census revision (net migration undercount) 04-08 | 315 | 315 | |||||||
Total | 1936 | 2535 | 2213 | 1705 | 4149 | 4238 | |||
Difference between net migration and change in population | 599 | -508 |