Migration statistics: what the data tell us

Transcription

1 Migration statistics: what the data tell us getstats in Parliament event organised by Royal Statistical Society House of Commons Library All Party Parliamentary Group on Statistics

2 Migration statistics: what the data tell us Jakub Bijak, Southampton University Sin Yi Cheung, Cardiff University David Coleman, Oxford University Ian Cope, ONS Jonathan Portes, NIESR Hetan Shah, Royal Statistical Society (chair) 2

3 Ian Cope, ONS 3

4 International Migration Statistics Ian Cope, Director Population & Demography December 2014

5 Population and migrations statistics overview

6 Calculating LTIM estimates

7 Long-Term International Migration 583, , ,000 Source: Long-Term International Migration (LTIM), ONS

8 Net Migration EU/Non-EU/British Net Migration (thousands) Net migration 260, Non-EU Citizens 168, , EU Citizens British Citizens q1 q ,000 Calendar Year YE = Year Ending q1 = YE March q2 = YE June Source: Long-Term International Migration (LTIM), ONS

9 EU migrants living in UK (2011) 800, , , , , , , ,000 0 Source: 2011 Census UK passports held, or country of birth

10 Source: 2011 Census

11 Economic Activity Source: 2011 Census

12 Strengths and Limitations of sources IPS LFS/APS Visas NINos Patient Register Census Frequent collection More information than Semaphore Frequent collection Counts not estimates Published with 2 month lag to reference period Counts not estimates Published with 2 month lag to reference period Can analyse people registering with previous address abroad Rich data source including ethnicity, language, labour market, etc Limited sample size robust nationally but less robust for smaller sub groups Smaller sample than census Greater respondent burden Mainly non-eu citizens Includes short term migrants Includes short term migrants Migration for work only Lags before registering Does not record nationality Not registering or deregistering at all, or with a lag Once every 10 years

13 Impact of migration on population ONS calculate long-term international migration based on the UN definition: A person who moves to a country other than that of his or her usual residence for a period of at least a year (12 months), so that the country of destination effectively becomes his or her new country of usual residence ONS calculate 54% of population growth between mid-2001 and mid-2013 due to migration. Migrants tend to be young adults with higher fertility rates 60% of projected increase in population mid-2012 to mid is attributable to future migration either directly attributable to future migration (43%),or indirectly attributable to the effect of fertility and mortality on these future migrants (17%).

14 International Migration Statistics For more information:

15 Sin Yi Cheung, Cardiff University 15

16 Getstats in Parliament: Beyond Net Migration Statistics 4 December 2014, Houses of Parliament Sin Yi Cheung School of Social Sciences Cardiff University cheungsy@cardiff.ac.uk

17 What do migration statistics tell us? Labour market integra/on Evidence of ethnic penalty in the BriFsh labour market (Heath and Cheung 2007), parfcularly in the private sector (Heath and Cheung, 2006) using data from: Quarterly Labour Force Survey (ethnicity, labour force parfcipafon, unemployment, occupafon) General Household Survey (Fll 2001) parents country of birth, own COB, year of arrival Individual Sample of Annoymised Records SARs m cases

18 Labour Market Integration Clear evidence of language skills contribufng to labour market success for the second generafon s ethno- religious minorifes (Cheung 2014) Ethnic Minority BriCsh ElecCon Study (EMBES) 2010 Refugees IntegraFon - Social Network, language and employment (Cheung and Phillimore 2013) Survey of New Refugees (SNR) : data on social networks: co- ethnic, religious, relafves, friends, formal organisafons; also on health, housing, and language training

19 Signi<icant improvement for the second generation (EMBES 2010)

20 Sample size of Survey of New Refugees

21 Refugee s social network pro<ile at baseline survey

22 Recent advance in migration data Census 2011 New quesfons on migrafon: temporary residence, English language ability, passport held SARs 10% individual Samples of Annoymised Records with full occupafon and COB coding Labour Force Survey WhyUK since 2010

23 Policy questions cannot be answered without robust data Migrants own views on integrafon meaning and priorifes Longitudinal survey not just of refugees but all migrant groups Consistent quesfons in repeated cross- secfonal and longitudinal surveys Without Parents country of birth, impossible to idenffy (white/other white) second or any third generafon, in order to study intergenerafonal change/mobility.

24 Policy questions cannot be answered without robust data First language at home quesfon needed every year Language skills at arrival and language acquisifon over Fme Pre- migrafon socio- economic characterisfcs: educafon, employment (only in SNR) Lessons from other immigrafon desfnafons cross- nafonal comparison crucial to see how we fare

25 David Coleman, Oxford University 1

26 Migration statistics: what the data tell us RSS / House of Commons Library Briefing. Or what the data cannot tell us? David Coleman david.coleman@spi.ox.ac.uk

27 Structure of UK migration data Home Office Control of Immigration data inflow only. Entries classified according to Immigration Rules (2500 page handbook) PBS, visas, asylum etc. Nothing on stock: e.g. of resident migrants under each category by nationality and by visa type with which they entered the country (e.g. number of Indian migrants with ILR who came in as students). Many potential migrant flows not fully captured - over 200,000 family visit visas issued each year. How many actually went home? Data not easily related to: International Passenger Survey (plus tweaks: asylum, switchers etc) Unique direct measure of inflow and outflow. But small voluntary sample only (80% response rate, interviews 2620 in, 1824 out in 2011., +/- 35,000 confidence interval). Inadequate for detailed analysis. No data on many individual countries of interest (e.g. Syria, Brazil). Requires frequent, often substantial revision ( for 2006). New annual National Insurance Number allocations Stock data from decennial census, Annual Population Survey etc

28 What the data cannot tell us Reliable, complete numbers of persons entering and leaving the United Kingdom and their characteristics. Timely exact information on number of immigrants resident in the UK and their whereabouts. Consequent uncertainty on national population estimates Whether persons admitted on time-limits have actually left the UK and when. e.g. how many in UK have indefinite leave to remain? How many as students, etc? Abundant data and analysis, not always easy to find. Don t know what we don t know None of the data sources used, while offering the best data currently available, are specifically designed to capture information solely on long-term international migration. (ONS 2014)

29 Difficult data an example Question: Which is the correct number for net international migration in 2007? ;? ? ?

30 Difficult data the answers Answer: All are correct (insofar as any are) is the IPS number, the only one that can be used for most statistical analyses (by age, purpose of journey, citizenship etc) was the LTIM number (IPS plus asylum etc) before the 2011 census result revision, the difference between inflow and outflow is the revised net number which the ONS and HO now use following adjustment of the undercount of net migration revealed by the 2011 census However, the census revision cannot determine the gross flows in and out. The difference between the LTIM gross inflow and outflow gives 233, not is correct to make sense of the gross flows. Simples! (similar variations exist for other years).

31 What is to be done? Heroic efforts by ONS to improve / patch up data - Migration Statistics Improvement Programme since 2002 Therefore repeated revisions of figures. More data and detail has been provided. E-borders failure. Semaphore cannot provide better in / out data. But however ingeniously the IPS is patched up, it s still the IPS.

32 Some small ways forward 1 Wider synthesis of data relating to migration under a common hub: IPS, HO (crime as well as migration), Population Surveys, Census, NiNos, HIPE and NHS, etc. Intermediate menu-driven step in ONS website between headline data and the vast store of information

33 Some ways forward 2 Join the rest of Europe in the 21 st Century by abandoning the 18 th century census concept Develop a Population Register for residents and immigrants (ID card not needed, but desirable). Already much of the way there with administrative data. Political will and understanding lacking At least a Feasibility study urgently needed

34 A blast from the past. I heard from several distinguished persons that there was a general complain to the imperfection of elementary population documents in this country It is indeed a subject of wonder to every intelligent stranger, that in a country so intelligent as England, with so many illustrious persons occupied in statistical enquiries, and where the state of the population is the constant subject of public interest, that the very basis on which all good legislation must be grounded had never been prepared; foreigners can hardly believe that such a state of things could exist in a country so wealthy, wise and great. (Adolphe Quetelet, 1835).

35 Jakub Bijak, Southampton University 11

36 ESRC Centre for Population Change Jakub Bijak University of Southampton Uncertain migration statistics: How much can we know? RSS/GetStats in Parliament event on Migration Statistics Houses of Parliament, London, 4 December 2014 With thanks to: George N Disney, Jonathan J Forster, Sarah Lubman, James Raymer, Peter W F Smith and Arkadiusz Wiśniowski

37 The Challenge Migration statistics: A paradox of plenty? Issues: definitions, coverage, accuracy Migration flows according to a benchmark definition Source 1 Source 2 Source 3

38 Option one: Multiple sources Model-based reconciliation of evidence Source: Disney (2014) courtesy of the author

39 Option two: Multiple countries Model for a whole EU migration system: Integrated Model of European Migration Example: UK Raymer et al. (2014) JASA 108,

40 Option three: Multiple models IMEM + LFS-based model for short- and long-term Polish migration to the UK Source: Wiśniowski (2014), courtesy of the author

41 Key Messages Cornucopia of data Reconciliation via statistical models Uncertainty becomes a key feature Future: micro-level data linkages, other sources (e.g. Semaphore) Thank you! j.bijak@soton.ac.uk

42 Jonathan Portes, NIESR No slides, spoke from notes 18

43 House of Commons Library House of Commons Library Standard Notes available: Migration Statistics Asylum Statistics 19

44 Migration statistics: what the data tell us getstats in Parliament event organised by Royal Statistical Society House of Commons Library All Party Parliamentary Group on Statistics