Brexit and the Labour Market Contents

Chapter 2: Migration statistics

“Evidence-based policy-making needs data that is fit for purpose”10

6.The starting points for any new immigration policy must be: to measure and understand the movement of people into and out of the UK; the distribution of migrants and the consequent pressures on local services and communities; the extent to which businesses rely on migrant workers; and the future needs of the British economy.

7.In the year ending December 2016, it is estimated that 588,000 people migrated to the UK. Of these 250,000 were EU nationals, 264,000 non-EU nationals, and 74,000 British nationals.11 When those leaving the UK are taken into account, net migration to the UK was approximately 248,000 for this period. Figure 1 shows how net migration has changed over the last 25 years.

Figure 1: Long-term trends: net migration, immigration and emigration 1991 to 2016 (thousands) (*figures for 2016 are provisional)

combination chart showing net migration, immigration and emigration from 1991 to 2016

Source: ONS, Long-Term International Migration into and out of the UK by citizenship, 1964 to 2015, December 2016: [accessed March 2017]; ONS, Provisional long-term International Migration estimates, May 2017: [accessed July 2017]

8.In this chapter we consider the current state of the immigration statistics, the difficulties which were raised by witnesses, and possible ways to improve data collection.

Measuring migration

9.There is no one simple measure of migration. A picture of the movement of people into and out of a country and changes in its population is built up using a number of different measurements. Some quantify the flow of migrants (the number of people entering and leaving the country). Other data are based on the number of migrants living in the UK at a particular point in time. All estimates are subject to considerable margins of error. There are currently no wholly reliable migration statistics.

10.There is no one definition of migrant; at least three are used in the official statistics:12

Table 1: Sources of immigration data

Measurement (year commenced)



Definition of migrant

Method of collection and size

Significant exclusions

International Passenger Survey (1961)

Quarterly (since 1991)


Change of residence

Sample face-to-face survey at points of entry. 289,000 interviews (2015) of which 4,000-5,000 were migrants.

Those entering via the UK/Ireland land border or on overnight flights.

Labour Force Survey (1973)/ Annual Population Survey (2004)



Foreign born and foreign nationals

Face-to-face or telephone sample survey. Labour Force Survey: 40,000 households (300,000 individuals) per quarter. Population Survey: 320,000 (includes LFS interviewees).15

Those in communal accommodation; short-term migrants.

National Insurance number allocations (data available from 2001)



Foreign nationals

DWP records.

In the year to Dec 2016 825,000 (all numbers issued).

Those who do not require a NI number.

Census (1801)



Foreign born and for England Wales foreign passports

Household survey. 56.1 million (90 per cent response rate).

Migrants have a lower response rate than the general population.

Entry Clearance Data (data available from 2001)



Requires entry clearance

Home Office records.

In 2016 350,000 visas were issued.

EEA migrants.

The International Passenger Survey

12.The International Passenger Survey is used to calculate the number of immigrants and emigrants (and therefore the net migration figure).16 The data for the survey is collected by ONS officials who interview passengers at 19 major and regional airports, eight ports and the Channel Tunnel rail link.17 Less than one per cent of interviewees (between 4,000 and 5,000 each year) are identified as immigrants or emigrants. These migrants are then asked questions about their reasons for travelling to or from the UK, proposed length of stay, and characteristics.

13.There are a number of specific concerns about the survey and the use of the data:

(a)As a result of the small sample size, the figures become less reliable when broken down and detailed estimates—such as the number of students or migrants from a specific area—are considered.18 The selection of interviewees in this sample was also questioned: Professor Portes pointed out “when you come into the country as an immigrant … you would think you had better things to do than stop and chat with an ONS official for 20 minutes”.19

(b)Some groups are underrepresented in the sample. The ONS acknowledged that data on those leaving the UK is poor and that there is a particular issue with counting the number of students leaving at the end of their degree. 20

(c)To attempt to account for these difficulties in how the survey is conducted, it has a large margin of error. The latest net figures (showing net migration of 248,000) are subject to a margin of error of 41,000 either way.21 The ONS says there is a one in 20 chance that the true value of net migration is outside this margin of error.22

14.Despite these issues, some witnesses pointed out that the data “tally reasonably well” with other sources.23 It was noted that the 2011 census revealed that net migration for the period 2001 to 2011 had been underestimated by 346,000 and the migration figures were revised from 2.18 million to 2.53 million for the decade to 2011.24 This was considered to be a “relatively small” discrepancy.25

The Labour Force Survey

15.The Labour Force Survey was described as “the best source of data to explore the impact of immigration.”26 This survey (together with the Annual Population Survey) provides information on employment and unemployment including about employees’ country of birth and nationality. However, it does not capture those living in communal households and may undercount very short-term migrants. A relatively low response rate (50 per cent) may mean the survey fails to capture some groups too.27 As with the Passenger Survey, the small sample size affects the reliability of sub-sets of the data.28

16.Some sectors reliant on seasonal or short-term migration—such as hospitality and farming—considered the survey underestimated their migrant workforce. The British Hospitality Association’s survey of their members revealed that 24 per cent of the hospitality workforce were from the EU. The Labour Force Survey estimate for this sector was only 12.3 per cent.29

National Insurance numbers

17.The allocation of National Insurance numbers to overseas nationals can, the ONS acknowledged, “provide some useful indication” about long-term migration.30 These figures show that for the year ending December 2016, 825,000 National Insurance numbers were allocated to overseas adults of which 626,000 were given to EU citizens.31

18.The National Insurance numbers should not be use in isolation. Many people register for a National Insurance number and leave after a short period. Professor Portes pointed out that the Government does not publish “on any regular or comprehensible basis the number of national insurance numbers that are in use and how many have been in use for what period over the last few years.” This uncertainty limits the value of the data available.32

19.Increasing reliance has been placed upon the migration statistics to formulate and judge government policy. Many of the available measures are wholly inadequate. In particular the long-standing and widely identified problems with the International Passenger Survey mean that it cannot bear the burden placed upon it and cannot be relied upon to provide accurate estimates of net migration.

Gaps in the available information

20.The above methodological weaknesses mean that there are gaps in the information available to policy makers.

21.Alan Manning, Chairman of the Migration Advisory Committee, told us that the presentation of the statistics currently does not give “an overarching view of arrivals and departures and how that is put into different boxes” and makes it hard to “have an overall view of what is going on.”33

22.Specific gaps identified in evidence included:

Box 1: Measuring student migration

The ONS stated that one of the “significant challenges” it faces is to understand what international students do when they complete their studies.39 The available data do not allow for a precise figure of how many students stay at the end of their degree to be calculated.40 Analysis by the Institute for Public Policy Research (IPPR) of the information to be gleaned from the various sources is displayed in the table below.

IPPR analysis: approximate number of non-EU students arriving in 2011 remaining in the UK after five years according to data sources41

International Passenger Survey42

Home Office Visa Data

Annual Population Survey

Higher Education Statistics Agency Destination Survey





23.The difficulties with the immigration statistics will become more acute as the Government seeks to develop and assess the impact of new immigration policies. Without greatly improved statistics the Government will be formulating policy in the dark. The Government must work urgently to improve the reliability of the data. The Government must also continue to monitor the stock and distribution of migrants to ensure adequate public service provision.

Improving immigration data

24.The problems with the immigration data are widely accepted and the ONS is focused on improving the reliability of the statistics.43 Three methods of improvement emerged from the evidence we heard.

Better use of administrative data

25.Measures contained in the Digital Economy Act44 will allow the ONS to access more administrative data. The Government consider that this will “provide more accurate, frequent and timely statistics, instead of relying on surveys.”45 Jon Simmons, Head of Migration and Border Analysis at the Home Office, told us that:

“If HMRC’s and DWP’s data, for example, could be linked with information that the Office for National Statistics might hold on nationality, it would suddenly become a really rich data source that could tell us things that we were not able to know previously.” 46

26.Professor Portes argued that data held by HMRC and the Department for Work and Pensions on migrants’ economic activity—such as paying tax or claiming benefits—could be used “much better”. He pointed out that “there is a lot of data there, although we do not tend to use it very much”. Dr Lucie Cerna of the OECD Migration Division stated that other countries such as Australia and Canada successfully used administrative data in this way.47

27.The Government should prioritise plans for the comprehensive sharing of data across departments. We recommend that the Government develops a systematic understanding of the movement of immigrants within the UK economy. This will require analysis of multiple sources of information, and should include:

(a)matching PAYE and National Insurance number registrations;

(b)matching self-assessment records for self-employed migrants and sole traders with issued National Insurance numbers; and

(c)using data on benefit claims and tax credits to ascertain whether those with unused National Insurance numbers remain in the UK claiming benefits.

Entry and exit checks

28.Whilst the available administrative data could provide a rich source of information, its use does not address the fundamental problem of the reliability of the International Passenger Survey. Without this the Government does not know how many migrants enter or exit the UK each year. One solution to this is to count those entering and leaving the UK.

29.The Government now collects limited information on all those leaving the UK through exit checks.48 The Minister for Immigration told us that these checks will “help us get a better picture” of movements to and from the UK.49

30.The data currently collected is from the Advanced Passenger Information provided by inbound and outbound air passengers. Since April 2015 checks have also been undertaken on ferry and rail passengers.50 The information required includes the passenger’s full name, nationality, date of birth, gender, and travel document number, type and country of issue.

31.The Home Office considered that when linked with other information — such as visa type— these checks could be used to monitor “the flows into and out of the UK for those people subject to immigration control”. However, it would not be possible for this information to replace the International Passenger Survey as:

“it is not possible to determine the very small fraction of long-term migrants among the very large number of travel movement per year”… It is also considerably slower to determine long-term movements through actual travel patterns than by asking individuals their intentions at the point of arrival … because it is necessary to wait for a year after arrival.”51

32.The re-introduction of checks on those leaving the UK provides an opportunity to count those entering and leaving the UK in a systematic manner. The Government should explore how the available data can be combined with other information and used to provide a long-term check to the information provided by the International Passenger Survey.

10 Written evidence from the Institute of Directors (LMT0012)

12 Some other statistics (such as Home Office Visa Statistics and Higher Education Statistical Authority data) use other definitions, such as those requiring entry clearance or those domiciled outside the UK.

13 United Nations Statistics Division, Recommendations on Statistics of International Migration, Revision 1, 1998: [accessed 12 July 2017]

14 These figures cover all regions of the UK, including Scotland and Northern Ireland. In the case of the census there are some differences in the questions asked of households in each part of the UK.

15 ONS, Annual population survey, Quality and Methodology Information, September 2012: [accessed 12 July 2017]

16 The rest of the international migration estimate is comprised of an adjustment for asylum seekers using Home Office data, an adjustment for ‘switchers’ who decide to stay longer or leave earlier than originally intended and data on migration from Northern Ireland and the Republic of Ireland.

17 Written evidence from the ONS (LMT0036). The majority of migrants are interviewed at Heathrow. ONS, International Passenger Survey: Quality Information in relation to migration flows, 2016: [accessed 12 July 2017]

18 Oral evidence taken before the EU Homes Affairs Sub-Committee, 21 December 2016, (Session 2016–17), Q 51 (Paul Vickers). 4,000 migrants provide the basis for around 90 per cent of the total international migration estimate.

19 Q 5 (Prof Jonathan Portes)

20 ONS, International migration data and analysis: Improving the evidence, February 2017: [accessed 12 July 2017]. In this report the ONS uses a traffic light system to classify the reliability of the available data on migration. Emigration is ‘red’ indicating data where the ONS has “little or no information or where more work is required to assess [its] reliability.” This issue is explored in more depth below.

22 Ibid. This equates to a 95 per cent confidence interval.

23 Q 5 (Prof Jonathan Portes) and Q 64 (Jon Simmons)

25 Q 64 (Jon Simmons); Professor Rowthorn disagreed that this was a small amount. Q 6 (Prof Robert Rowthorn)

26 Written evidence from the London School of Economics (LMT0020)

27 Such as irregular migrants, those working long/irregular hours, those in large households.

28 Written evidence from the ONS (LMT0036)

29 Written evidence from the British Hospitality Association (LMT0008); see also written evidence from the NFU (LMT0026).

30 Written evidence from the ONS (LMT0036)

31 Department for Work and Pensions, National Insurance number allocations to adult overseas nationals to December 2016, Summary Tables, Table 1, 23 February 2017: [accessed 12 July 2017]

32 In January 2017, the Office of Statistics Regulation temporarily revoked the ‘National Statistics’ status these figures whilst work to improve the supporting guidance and overall public value was undertaken. Office for Statistics Regulation. Assessment of compliance with the Code of Practice for Official Statistics Assessment Report 331 Statistics on National Insurance Number Allocations to Adult Overseas Nationals, January 2017: [accessed 12 July 2017]

33 Q 5 (Prof Alan Manning)

34 Q 13 (Prof Alan Manning)

35 Q 5 (Prof Jonathan Portes)

36 Ibid.

37 Written evidence from the National Farmers Union (LMT0026)

38 Q 7 (Prof Jonathan Portes)

39 ONS, International migration data and analysis: Improving the evidence, February 2017: [accessed 12 July 2017]. In this report the ONS uses a traffic light system to classify the reliability of the available data on migration. EEA and non-EEA student emigration is ‘red’ indicating data where the ONS has “little or no information or where more work is required to assess [its] reliability.”

40 Q 3 (Prof Alan Manning); see also Q 6 (Prof Robert Rowthorn)

41 IPPR, Destination education: Reforming migration policy on international students to grow the UK’s vital education exports, Table 3.3: September 2016: [accessed 12 July 2017]. The IPPR analysis is based on a cohort of students arriving in 2011. The IPPR caution that “Comparisons between the figures in the tables should be made with caution, given the different methodologies used and the limitations of each of the data sources.” IPPR, Destination education

42 The figures for the IPS are higher due to the challenges faced by the ONS in counting the number of student departures. The ONS acknowledge that “the IPS figures of international students immigrating to the UK are consistently higher than the IPS figures of former international students emigrating”. The ONS are seeking to collaborate with Government departments to improve the data: ONS, International migration data and analysis: Improving the evidence, February 2017: [accessed 12 July 2017]

43 Written evidence from the ONS (LMT0036); see also ONS, International migration data and analysis: Improving the evidence, February 2017: [accessed 12 July 2017]. In the short-term the ONS told us they wish to improve the data in high priority policy areas, such as international student departures. An update on the work to improve student migration statistics in due on ‘mid-2017’.

46 Q 71 (Jon Simmons)

47 Q 33 (Dr Lucie Cerna)

48 Home Office, Exit Checks fact sheet, March 2015: [accessed 12 July 2017]

49 Q 62 (Robert Goodwill MP)

50 Home Office, Exit Checks fact sheet, March 2015: [accessed 12 July 2017]

51 Home Office, A report on the statistics being collected under the exit checks programme, August 2016: [accessed 12 July 2017]

© Parliamentary copyright 2017