Sources of migration statistics

During each quarter, the Office for National Statistics brings together in one place the latest international migration statistics collected from survey and administrative sources. These include international migration flows in and out of the UK from the International Passenger Survey (IPS), employment of overseas nationals in the Labour Force Survey (LFS), as well as administrative data on national insurance registrations by overseas nationals, non-EU visas granted and asylum applications. These data are available in the Migration Statistics Quarterly Reports.

The statistics from the various sources differ from each other for good reasons, mainly due to their different coverage. The trends from the different sources, as well as their levels, will also differ over time, especially during periods of changing migration patterns. This is true, for example, currently for the International Passenger Survey (IPS) and the National Insurance Number (NINo) data for overseas nationals entering the UK and this note explains why such differences might be expected to occur.

National Insurance numbers are registered to EU nationals who come to the UK and seek employment or commence work. These EU nationals can fall into four groups:

a) Those who come to the UK who do not immediately make an application for a NINo, i.e. those that were not seeking work straightaway (for example, students) or have not been able to register for a NINo because of employment restrictions (for example, until recently, Bulgarian or Romanian citizens);

b) Those who come to the UK to work and stay for a period of 12 months or more (long-term migrants);

c) Those who come to the UK to take up short-term work (short term migrants) - for example, in the agricultural or hospitality sectors; and

d) Those who come to the UK but subsequently return home in a short time period, for example because they cannot find work or for other reasons (eg. they don’t like the work or are homesick).

There is no requirement when allocating a National Insurance number to establish how long the individual intends to remain in the UK. The NINo is required to ensure that an individual pays the correct amount of tax and National Insurance contributions regardless of their length of stay / employment in the UK.

The long term migration estimates produced by ONS are in line with United Nations definitions of a long term migrant and will only include the first (3a) and second (3b) of the categories, above, and the first category will be included during the year when the EU national entered the UK.

Differences between the IPS and National Insurance Number statistics are not new and will not necessarily reflect a measurement problem with either source. They have been large historically when there are changes to the rules on migration, such as removing restrictions on which citizens can enter into the UK, as these can cause a change in migration behaviour that can have quite different impacts on the two sources at different times. The removal of restrictions on work for EU2 (Bulgarian and Romanian) citizens on 31/12/2013 is an example of such a change. The differences between the sources mentioned in 3.a-d are likely to explain all or a large proportion of the gap currently being noted in the figures.

The IPS-based estimates are for long-term international immigration (LTIM) only i.e. people who intend to stay in the UK for 12 months or more for any reason (not just to work). This is consistent with the United Nations definition of a migrant. Estimates are based on around 800,000 interviews conducted each year with those people travelling to and from the UK. There is currently no administrative count of all people entering and leaving the United Kingdom that enables an estimate of long-term migration to be made. In order to produce these estimates of the change in the resident population which is due to migration, we interview people at the places of entry and exit into the UK to estimate levels of migration and those estimates are subject to both sampling and non-sampling errors (for example, if the person interviewed gives incorrect information). There is no need to interview at other places within the United Kingdom, such as London Victoria Coach station where some migrants may alight, because the same migrants have already entered the country at an airport, at the Channel Tunnel or by sea through a ferry port and are eligible for interview then.

Since the IPS is a sample survey, and not an exact count, confidence intervals around the estimates are also published which provide a measure of the reliability of the estimates and can be used to identify statistically significant changes. The IPS sample is carefully designed to provide the most accurate migration estimates possible from a survey. However, at times when migration patterns change significantly, there is a risk that the IPS sample design may need to be changed to fit these after the event when they are observed. The decennial Census provides a check on these numbers. Several changes were made, retrospectively, for example in the mid-2000s as a consequence of changing patterns of migration from Central and Eastern Europe. ONS monitors travel patterns, for example flights to and from the EU2 countries (Romania and Bulgaria), to ensure that new routes are picked up quickly and the sample adjusted where necessary.

The Migration Statistics Quarterly Report includes statistics on National Insurance registrations for all overseas nationals who come to the UK seeking to work and who register for a NINo. The figures are based on the recorded registration data from the HMRC National Insurance Recording and Pay As You Earn System, i.e. after the NINo application process has been completed, and so are not a direct measure of when someone migrated to the UK. You would expect a “spike” in National Insurance Number registrations following the removal of restrictions on work for EU2 (Bulgarian and Romanian) citizens and indeed this is evident in the statistics, some of which will reflect people coming to the UK to work and apply for a NINo after the employment restrictions have been lifted. However, some individuals were residing in the UK prior to the employment restrictions being lifted on 31/12/2013 and would only have been able to register for a NINo after that date. Those individuals resident in the UK prior to 31/12/2013 will have been counted in previous years in the IPS.

Unlike the main international migration estimates produced from the IPS, National Insurance registrations will cover all of those who register seeking to work, including those who stay for under 12 months in the UK. This will include many short term migrants (such as those who might come to work over the summer, for example in the agricultural or hospitality sectors). For instance, in the year ending June 2013 (the latest date for which we have short-term migration estimates), EU long-term immigration was 183,000 and short-term immigration (for work, study and business) was 239,000. The number of National Insurance Number registrations over that period to mid-2013 was 398,000.

ONS is undertaking on-going reconciliation work on the four main sources of international migration data (International Passenger Survey, Labour Force Survey, National Insurance numbers and visas), including the difference between the IPS estimates and NINos. It is currently planning to provide a background note on this work in May as part of the Migration Statistics Quarterly Report, incorporating the latest available short term migration data published in that month, to help explain further why the two datasets are showing different trends. The exercise will compare long-term and short-term migrants and examine IPS data for those who arrive for less than one month.

When available, DWP and HMRC data on NINo activity (those who have applied for a national insurance number and are still active in the UK) will be incorporated into this work to provide additional information for the users of our statistics and a more complete picture.