How immigration is measured

Updated 17 January, 2007

1. The United Nations definition of a migrant, standard in demography, is anyone who changes his or her country of usual residence for at least a year.

2. There are three main sources of international migration data international migration statistics produced yearly by the Office for National Statistics which measures immigration and emigration on a sample basis; settlement data from the Home office which measures long-term immigration of non-EU citizens only, but not emigration, and the census which provides a snapshot of the foreign-born population in the UK every 10 years.

3. The internationally accepted United Nations definition of a migrant is someone who changes his or her country of usual residence for a period of at least a year, so that the country of destination effectively becomes the country of usual residence.

4. This definition will include people who come to the UK (immigrants), or who leave the UK (emigrants), for more than a year for a variety of reasons. The main streams of migrants are: those who enter or leave for work and their dependants, family formation and reunion (i.e. marriage and family members joining existing migrants) and asylum seekers. Students are also included but will normally be partly counterbalanced by those leaving at the end of their course. Many now stay on after their studies. The definition will not include visitors and other people who come to the UK for less than a year such as seasonal agricultural workers.

5. In recent years the UK has had no system of embarkation controls. In 2005 about 32.5 million foreign nationals arrived in the UK 20.7 million of whom were nationals of the European Economic Area (the European Union plus Iceland, Norway, Switzerland and Liechtenstein) and 11.8 million other foreign nationals. Given the scale of these movements it is clearly difficult to measure migration with great accuracy using the existing low technology systems with their very incomplete coverage.

6. Three main measures are used:
The Office for National Statistics (ONS) international migration estimates, produced annually.
Settlement data produced from Home Office statistics each year.
The UK census which is carried out each decade.

Each of these is described in more detail below.

ONS international migration statistics
7. This is the measure most commonly used to estimate the numbers of migrants to and from the UK. The first estimates are produced in the last quarter of the subsequent year and more detailed information is produced in the following Spring.

8. The main source is the International Passenger Survey (IPS) - a voluntary sample survey of passengers travelling via the principal airports, sea routes and the Channel Tunnel. Not all ports of entry are covered. The survey asks questions about the intended length of stay of travellers to and from the UK on which the first estimates of international migration are based. The survey has a number of limitations:

a) It excludes land routes between Ireland and the UK. Ireland is part of a common travel area with the UK and there is no survey of people crossing the border between the Republic and Northern Ireland.
b) It excludes most asylum seekers and their dependants.
c) The survey asks about peoples intentions which may not accord with what they do in practice.
d) As a sample survey it only includes a very small sample of migrants so there is a great degree of uncertainty. All the information on inflows for the 2004 statistics were based upon 2801 interviews (standard error 3.8%) and for the outflow, on just 755 interviews (standard error 4.7%). That means that the grossed-up estimate of inflow has a 95% chance of being somewhere between 478,600 and 557, 600, and the outflow between 281,500 and 339,300. At the extreme, that means that net inflow could be between 139,000 and 276,000.

The estimates for Pakistan, quite an important source country, were based upon 231 interviews of immigrants and 6 interviews of emigrants. The estimate of a net 3,000 inflow from the Caribbean was based on the difference between 28 interviews in and 6 interviews out.

Apart from the uncertainty, all this means that the data can only be analysed and tabulated by very gross aggregated categories of (for example) country of origin and citizenship, purpose of visit, and age. Among other things it is not easy to separate those who enter for family re-unification, and those who enter in order to marry.
e) It is voluntary.  About 17% of those approached decline to participate
f) At present it does not cover the airports used by budget airlines, nor coach travellers.  This makes it particularly inaccurate in respect of East Europeans

9. ONS adjust their results to attempt to allow for the first 3 of the limitations listed above as described in paragraph 10 below. The fourth problem can only be resolved by a much larger sample size, or preferably, by a completely different system.

10. ONS takes great care to make 3 adjustments to the IPS estimate to produce their overall estimate of net migration.

1) First they allow for migration between the UK and the Republic of Ireland using estimates agreed with the Irish Central Statistical Office. (Net migration flows have been relatively insignificant in recent years).
2) Second, it is assumed that most asylum seekers will not have taken part in the IPS, or if they did, will not have stated their intention of staying in the UK. The ONS therefore make an allowance for asylum seekers from Home Office records of asylum claims. The total adjustment is essentially the total number who have requested asylum (including dependants) less the total number who have been removed or who have left voluntarily. Small allowances are made for some asylum seekers being included in the survey data on entry to the UK and some leaving voluntarily without being recorded in Home Office figures.
3) The third adjustment is for people who come to the UK or leave the UK intending to stay for less a year who then extend their stay (so called visitor switchers) and for people who come to the UK, or leave the UK, intending to stay for more than a year but who leave in less than a year (migrant switchers). The number of visitor and migrant switchers is estimated from the IPS data based on the number of people who say that their length of stay is 6-11 months and those who say that they may stay for more than a year but are not sure.

11. To summarise the above the final net international migration figure is calculated as: Net total international migration = net IPS migration flows + net Irish flows + net total visitor switchers net migrant switchers + net asylum seekers and dependants.

12. For 2005 the net total international migration was estimated to be 185,000 people calculated as:

Components Number
Net inflow from the IPS 168,000
Plus asylum seekers adjustment 11,000
Plus visitor switchers 38,000
Less net migration to Ireland* 10,000
Less migrant switchers 22,000
= Total net international migration 185,000

* In recent years there has been a net flow of people from the UK to Ireland.

13. The international passenger survey provides data on the internationally recognised definition of international migrant. Beyond that, it is a comprehensive measure of migration to and from the UK only insofar as it gives estimates of inflow and outflow. Its small sample size creates an inevitable problem of statistical uncertainly and prevents any detailed analysis of the characteristics of immigrants. Its voluntary nature is an additional problem. It is unlikely fully to include illegal immigrants. Those who enter the UK clandestinely will not be included in the data and those who enter with false documentation are unlikely to take part in the IPS which is a voluntary survey. It is also unlikely to include those who enter on short-term permissions and then stay on without getting an extension to their visa permission.

Settlement data
14. Another source is acceptance for settlement data compiled by the Home Office on people who have been granted indefinite leave to remain (settle) in the United Kingdom. The criteria used are based upon UK immigration law and are not the same as those used in the IPS. The data are published each year in the Control of Immigration Statistics issued by the Home Office.

15. Those who wish to stay in the UK indefinitely must seek permission to do so if they are not citizens of the European Economic Area (EEA). The Home Office figures comprise people who are granted settlement on arrival at ports, and people initially admitted to the country subject to a time limit and/or the fulfilment of conditions that are subsequently removed after application to the Home Office. For example, people entering on a work permit are given permission to stay for up to 5 years and may apply for settlement after 4 years in the UK. Very few people are accepted for settlement on arrival, most are accepted for settlement after a time limit or other conditions have been fulfilled.

16. Settlement data are therefore a good measure of long-term immigration to the UK. It has the advantage over the ONS International Migration measure of effectively excluding people who come here on an extended visit without intending to stay (eg. as working holidaymakers or students). It does, however, have limitations:


It is a measure of immigration only - there are no comparable data on how many British citizens have settled abroad.


It is a delayed measure of immigration it may be 4 years from the time of immigration before a person applies for settlement.


It only measures immigration from outside the European Economic Area. EEA nationals have the right to live in the UK and hence do not need to apply for settlement (although a few do).


It does not, by definition, include any illegal immigrants. Nor, in contrast to the ONS International Migration estimates, does it include failed asylum seekers just those who have been granted asylum or exceptional leave to remain (ELR) in the UK who have subsequently been granted settlement. (Asylum grants until recently gave an immediate right to settlement. Those being granted asylum currently and those granted exceptional leave are given limited leave to remain in the UK. At the end of this period their case is re-assessed and they may then be given settlement rights).

17. Settlement numbers therefore lag net immigration of non-EU citizens by four years or so. The following graph shows this relationship. The figures are the net immigration figures of non-EU citizens (a subset of the ONS international migration data) and settlement figures from the Home Office. (The settlement figures will typically be rather smaller because the international migration figures allow for asylum seekers whose claims have failed but who remain in the UK. The settlement figures only include those whose claims have succeeded).

18. At the census, which is carried out every ten years, one of the questions concerns birthplace. The census provides a very valuable snapshot of the numbers of migrants (defined as those who were foreign-born) in the UK and of the change that has taken place in the migrant population since the previous census. Comparing two censuses will also give a good indication of the number of people who have migrated to the UK over the decade if allowance is made for the ageing of the previous census’s migrant population.

19. As the census should include anyone who is normally resident it should cover broadly the same people who are covered by the ONS international migration data. In this respect the 2001 census results revealed a huge discrepancy when compared with the numbers predicted for 2001 from ONS birth and death registrations and international migration over the period 1991 to 2001, primarily in white males aged under 45. As the new methodology of the 2001 census was deemed to be infallible, the error was assumed to lie first in under-counting of emigration, thanks, it was thought, to a large number of migrant switchers. Some of the discrepancy was found to be over-correction of the 1991 census. Further investigation revealed substantial under-enumeration in the 2001 census in some urban centres. The census data has now been corrected as far as they can be, and the net migration estimates substantially reduced for all years since 1991 to enable them to fit the revised census data.

20. Migration Watch received reports from the public (including census enumerators) that the census under-enumeration in areas of high immigration was widespread. However, the census methodology allowed for households where no census return had been completed by analysing the results from similar households in the neighbourhood and adding to the census numbers accordingly. We believe, therefore, that the census should provide an accurate snapshot of immigration to the UK. The ONS international migration methodology, having been independently compared against the census output, should also be quite robust.