1. Introduction

1.1. Purpose

This paper provides a description of the methodology used to estimate Short-Term International Migration (STIM) for local authorities in England and Wales.

Local authority STIM estimates are calculated by taking the national estimate of short-term international immigration for 3 to 12 months for the purposes of work or study (based on the United Nations (UN) definition of short-term migration) and distributing it using administrative sources. Counts of short-term international migrants arriving for work (excluding business trips) and study (for formal higher or further education) are taken from the International Passenger Survey (IPS). For information on how the national-level estimates are calculated, please see the Methodology to Estimate Short-Term International Migration for England and Wales.

A quality and methodology information report is also published in which STIM estimates for local authorities are assessed against the six dimensions of statistical quality defined by Eurostat.

1.2. Background

A user consultation as part of the Migration Statistics Improvement Programme (MSIP) identified the need to produce robust, on-going STIM estimates for local authorities. Local authority level STIM estimates were first published as part of a one-off research report in 2009, based on a joint distribution and modelling method. Further estimates were published as experimental statistics in February 2012 on the basis of a distribution-only methodology that produces statistics in line with the improved long-term immigration distribution methodology for producing population estimates for England and Wales.

Since 2013, STIM estimates for local authorities and based on this improved methodology have been designated as National Statistics, following assessment by the UK Statistics Authority and it is these estimates that this methodology report covers.

This method uses main reason for migration information from the IPS to identify short-term flows of workers and students for England and Wales, which are then distributed directly to local authorities using administrative data sources.

The methodology for STIM estimates for local authorities:

  • is based on the UN definition of short-term migration, that is, moves made for 3 to 12 months for work or study reasons

  • refers to inflows only

  • provides estimates for mid-years (that is 1 July to 30 June)

  • uses administrative data from several different sources, including data on visas and student registrations, see Table 1

1.3. Definition of a short-term international migrant

STIM estimates for local authorities in England and Wales are based on the UN definition of a short-term international migrant, that is:

"A person who moves to a country other than that of his or her usual residence for a period of at least 3 months but less than a year except in cases where the movement to that country is for purposes of recreation, holiday, visits to friends and relatives, business, medical treatment or religious pilgrimage".

In practice for local authority STIM estimates, this refers to moves made to the UK for 3 to 12 months for the purposes of employment or study by non-UK residents.

1.4. Issues with measuring migration

The IPS, weighted by airline and other carrier information, is the primary source used to produce STIM estimates. While this is the best data currently available, it is not specifically designed to capture information solely on international migration. Therefore, these data do not specifically cater for the need to produce STIM estimates.

In addition, because it is not possible to produce an accurate figure for the number of people who are in the country illegally we do not produce estimates on the size of the illegal immigrant population. For more information on estimating the size of the illegal population, please refer to the following Home Office reports:

A later report, Economic impact on the London and UK economy of an earned regularisation of irregular migrants to the UK, written by the London School of Economics (LSE), also estimated that in 2007 the number of “irregular” migrants in the UK was 533,000.

Back to table of contents

2. Current methodology

2.1. Overview

We developed the current methodology to distribute England and Wales STIM estimates to local authority level during MSIP. The current methodology covers:

  • inflows only

  • moves made for 3 to 12 months for work and study reasons only (i.e. the UN definition)

  • mid-years from mid-2008 onwards

The methodology uses published short-term immigration totals derived from IPS data for workers and students staying for 3 to 12 months1. The worker and student totals are then distributed by mapping the IPS definitions as closely as possible to the relevant data within the administrative sources provided by other Government departments. It should be noted that this method does not:

  • measure outflows

  • measure stocks

  • include those people migrating for 1 to 3 months

  • include people migrating for reasons other than study or work

Figure 1 gives a high level overview of how short-term immigration estimates are broken down for distribution in the current methodology and which administrative sources are used in the distributions. The following sections give more information on definitions used and the detail of the methodology.

2.2. Data sources

Table 1 provides a summary of each data source used in the current methodology (as outlined in Figure 1). For more information on the coverage of each source and how it is used in the methodology, please refer to section 3.

2.3. Differences between the current and previous methodologies

Local authority STIM estimates published in 2009 used a combination of methods including modelling data from the IPS and distributing migrants using NINo data. In line with the long-term migrant methodology for producing population estimates for local authorities in England and Wales, the current STIM methodology is based solely on a distribution method. However, the short-term and long-term methodologies are not identical. The reasons for this include new data sources becoming available to us and difficulties in identifying short-term migrants in some of the data sources used in the long-term methodology. Additionally, to align with the UN definition of short-term migration, there are only two streams in our STIM estimates for local authorities (students and workers) compared to four streams (students, workers, returning migrants and others) in the long-term methodology.

The previous local authority level STIM estimates related to moves made for 1 to 12 months for all reasons for migration. The current methodology is based on the UN definition (i.e. moves made for 3 to 12 months for work and study) because:

  • the Code of Practice for Official Statistics states that National Statistics Institutes, such as ourselves, should promote comparability both within the UK and internationally by adopting common definitions

  • STIM estimates for workers and students are the most important for resource planning, as they are more likely to remain in a single local authority during their stay; the majority of other short-term migrants come to visit friends or family, or for extended holidays and are more likely to move between several local authorities

  • the data available are of better quality for students and workers staying for 3 to 12 months, as those coming for other reasons are unlikely to appear in an administrative data source (with the exception of those who require a visa to enter the UK) - those coming for shorter periods to work may not get a job and hence not register for a NINo

In changing the methodology to produce local authority level STIM estimates, we consulted with our own internal Methodology Quality Assurance and Research Review groups, as well as an Academic Reference Panel. The decision was communicated to local authorities via Local Insight Reference Panels in May 2011 and in MSIP quarterly updates.

The consultation on new country groupings ran from 21 January 2014 to 18 March 2014. As a result these new groupings have been applied to the new STIM tables.

A consultation on international migration outputs ran from 11 November to 23 December 2016 which looked in part at the presentation and timing of the STIM output. This will feed into shaping the format of future STIM publications.

Notes for: Current methodology

  1. These published IPS estimates are based on interviews undertaken at the end of the short-term migration event, so lengths of stay are known. Therefore, no error arises from changes of intention (e.g. intending to stay 2 months and actually staying 4 months). In this case, the end of the short-term migration event is when the individual leaves the UK after a 3 to 12 month stay in England or Wales.
Back to table of contents

3. Overview of methods – workers

As outlined in section 1, STIM estimates for local authorities are divided into short-term migrants who come to England and Wales for work and those who come for study. Both of these groups are distributed using a combination of administrative sources. This section describes the methodology for workers, while section 4 outlines the methodology for students.

The UN definition is used as a basis for selecting short-term migrants in the IPS to gain total estimates for distribution. In the administrative data sources used to calculate the distribution, it is not possible to use this exact definition. However, short-term workers and students staying for 3 to 12 months have been identified as accurately as possible with the data available.

3.1. Data used to distribute short-term immigration for work estimates

International Passenger Survey (IPS)

The IPS is a sample survey of passengers arriving at, and departing from, UK air and sea ports and the Channel Tunnel. The England and Wales level estimate of short-term international immigration for 3 to 12 months for the main reason of work, which is distributed to local authorities, is derived from IPS data.

It should be noted that, unlike for estimates of Long-Term International Migration (LTIM), short-term migrants are interviewed at the end of their stay (when leaving England and Wales). Therefore, STIM estimates are based on actual events, as opposed to migration intentions. This means no adjustment has had to be made for those whose actual behaviour differs from their reported intentions.

The IPS estimate for “work” used in the local authority level STIM estimates refers to as defined in the UN definition. This includes short-term migrants going to a definite new job, including au pair work, and those seeking a job, but excludes those visiting on business for their existing employer. Estimates for the latter are shown in the “Other Employment” category in England and Wales level STIM reference tables, but are not available for local authorities.

For more information about the IPS, please refer to International Passenger Survey: Quality Information in Relation to Migration Flows.

Migrant Worker Scan (MWS)

The MWS is used to distribute European Economic Area (EEA) short-term migrant workers to local authority level and is a 100% extract of migrants registered on the National Insurance and PAYE Service (NPS)1. This dataset includes registration details, self reported arrival dates and some demographic details along with address data at registration. The MWS includes both short-term and long-term migrants but since there is no activity information included with the data, and migrants do not have to de-register from the MWS when they leave the country, there is no way to distinguish between these two types of migrants. However, in the methodology, only those migrants who had 6 months or less between their arrival in the country and registration for a NINo have been included, to try to identify a larger proportion of the data that are short-term migrants.

Grants of extensions of leave to remain for work

Home Office data on extensions of leave to remain applied for by migrants planning to stay in the UK for 3 to 12 months are used to distribute short-term immigration estimates for non-EEA workers to local authority level.

Over 100 nationalities require a visa for entry to the UK for any purpose or for any length of stay. EEA migrants and Swiss nationals do not require a visa to come to the UK, but other nationalities, known as “non-visa nationals” (such as Australia and the USA), need to be granted a visa or an extension to come to the UK for over 6 months or for work. However, it is assumed that students who stay 3 to 6 months distribute in a similar way to students coming in for 6 to 12 months and therefore this should not affect the distribution.

The main advantage of extensions data is that they contain the current UK residential addresses of the applicants. Some issues with the data for the purpose of the STIM methodology are:

  • these extensions may not be the initial grant of leave to remain and may relate to individuals who have been in the UK for at least a short time already; these individuals may not have required a visa when initially entering the UK

  • an individual may have multiple consecutive extensions due to further grants of extensions (or, in limited circumstances, via switching to a different Tier)

  • all applications for extensions, not just first applications, are included in the data provided; this is due to the complicated and time consuming nature of linking together different applications made by the same person and it is assumed that this does not affect the distribution

Certificates of Sponsorship (CoS)

Those applying for extensions of leave to remain under PBS Tiers 2 and 5 (which cover most work routes for non-EEA nationals) must be sponsored by institutions for work, by being provided with a CoS. CoS data only cover workers and split into applications for visas and extensions. The data used in the local authority distribution methodology uses a combination of visa out of country applications and extensions of leave to remain for 3 to 12 months and are ‘used’2 CoS.

The CoS contains two addresses: one for the employer (place of work) and one for the headquarters of the sponsoring organisation. Each sponsor is assigned a certain number of certificates (by Tier) which they can distribute to potential migrants. The issues with this system for the purposes of the STIM methodology are:

  • applicants can have multiple certificates assigned to them either concurrently or consecutively, however, only in relatively rare cases would more than one be ‘used’ simultaneously; since this methodology uses “used” CoS data, this is unlikely to have a substantial impact

  • address information is based on employer address, this is not an issue for the majority of the country as most people live and work in the same local authority, however, grants of extensions data to include home address are used to distribute migrant workers to local authorities in London where migrants do not tend to live and work in the same local authority

3.2. Methodology used to distribute short-term immigration for work estimates

The method for distributing the England and Wales level STIM for work estimates is summarised in Figure 1. The England and Wales IPS short-term immigration for work estimate is split into four different nationality groupings: the EU83, EU24, Rest of EEA5 and Non-EEA.

Non-EEA workers

We use applications for extensions of leave to remain from non-EEA nationals to distribute the non-EEA part of the short-term immigration for work estimate from the IPS. The length of the extension is used to separate out short-term migrant workers from the extensions data.

The distribution for non-EEA short-term immigration is calculated separately for local authorities in London and outside of London. The London/non-London split is taken from the average of the extensions and CoS distributions. It is then applied to the IPS data to provide a ‘count’ of short-term migrants in London and outside of London. The work address has been used for the CoS data, and the CoS data have been subset to include only those who applied ‘out-of-country’ for a CoS. This is to eliminate double counting between the CoS and extensions datasets.

For non-EEA short-term workers outside of London local authorities, the average of the distributions from the CoS and extensions datasets has been used to distribute to local authority level. This is because using the average of the two sources smoothes out any outliers in the CoS data (due to its basis on work address) and any outliers in the extensions data (due to its basis on applications from individuals already in the country).

London is treated differently because it has a lower percentage of people living and working in the same local authority compared to the rest of the country. Using the CoS data, based on work address, in London, even in combination with the extensions data, does not give a plausible distribution. The distribution taken from the extensions data alone is therefore used to distribute non-EEA short-term migrant workers within London.

EEA workers

For short-term immigration for workers from the EEA, the MWS has been used to distribute to local authority level. Although the MWS contains both short- and long-term migrants – and it is not possible to distinguish between them – this is the only data source that could be used in the methodology to distribute this group of workers. Research during the development of this method compared the distribution of the MWS EU8 subgroup with a distribution from the Worker Registration Scheme (WRS)6. This comparison showed that the two distributions were similar enough to conclude that the MWS is a valid source for distributing short-term workers.

To try to capture as high a percentage of short-term migrant workers as possible and exclude as many long-term migrants as possible, only those in the MWS who apply for a NINo less than 6 months after arrival have been used in these distributions. Length of stay estimates in table STIM.02 show that the majority of short-term migrant workers identified in the IPS stay for 6 months or less.

Notes for: Overview of methods – workers

  1. The NPS collects, validates and stores information on NI contributions. In 2008 the database was expanded to record and validate payment of tax through the PAYE system. The NPS holds detailed information on periods of employment, self employment, sickness, tax credits, unemployment and periods of non-liability (e.g. caring for children) back to 1975. The NPS data holds records for anyone who has ever had a NINo in the UK. This includes data on people who may no longer be resident in the UK and people who have died. The database currently holds over 75 million records, which continues to grow as new people reach age 16 and enter the National Insurance / PAYE system. The NPS also includes data on the registration and identification of migrants from abroad who need a NINo to work or claim benefits.

  2. A CoS is classed as ‘used’ if it is linked with an application for a visa or an extension of leave to remain. It is not known if the application was successful or, in the case of applications for visas, if the individual subsequently migrated to the UK.

  3. The EU8 are the eight central and eastern European countries which joined the European Union (EU) in May 2004. These are: Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia and Slovenia.

  4. The EU2 refers to Bulgaria and Romania, which joined the EU in January 2007.

  5. The rest of the EEA includes all other EEA member countries. Of these only Iceland, Liechtenstein and Norway are not also members of the EU. Switzerland is also included in this group because, although not a member of the EEA, it is part of the Schengen zone and its citizens have the same right of travel into the UK as EEA citizens and thus behave in a similar manner.

  6. The WRS was a scheme with which EU8 nationals needed to register if they wished to take up employment in the UK. It ran between 2004 and 2011. Self-employed workers did not need to register with the WRS.

Back to table of contents

4. Overview of methods – students

4.1. Data used to distribute short-term immigration for study estimates

International Passenger Survey (IPS)

The England and Wales level estimate of short-term international immigration for 3 to 12 months for the main reason of study, which is distributed to local authorities, is derived from IPS data. A short-term migrant student in the IPS is defined as having been in the country for 3 to 12 months on departure and having arrived for a ‘formal course’. In addition, IPS data are used to split the short-term immigration for study estimate into the following three groups:

  • Higher Education (HE)

  • Further Education (FE) students of EEA1 origin

  • FE students of non-EEA origin

For more information about the IPS, please refer to section 3.1.

Higher Education Statistics Agency (HESA) Student Record

The student record is a record level dataset giving demographic details of all students at HESA reporting institutions. Reporting institutions are all HE providers that are funded via the Higher Education Funding Councils for England or Wales. This dataset includes details such as length of course, year of study (i.e. 1st year, 2nd year), term-time postcode and country of previous residence. The dataset has been linked with both the Student Record from the previous year and the MWS to identify and remove those in the country prior to starting study and so is considered to accurately identify short-term migrant students. In the HESA Student Record data, short-term migrant students are identified as those:

  • domiciled abroad

  • aged 17 to 59 inclusive

  • in their first year of study

  • whose previous institution is not in the UK or is unknown

  • whose length of course was or is expected to be between 3 months and a year

  • who are not in the Student Record the previous year

  • who, if successfully linked to the MWS, have a lag between arrival and start of study of less than three months

Data are split by academic year (1 August to 31 July). This is a slightly different reference period than that of the IPS data. This is not thought to significantly affect the distribution, as only 1.4% of short-term migrant HE students in the Student Record arrived in July in 2009/10.

Individualised Learner Record and Lifelong Learning Wales Record

Short-term immigration estimates for EEA FE students are distributed using aggregate data from the Learner Records provided by BIS and WG. These record government funded learners at FE level, including those studying English as a second language, at their term-time postcode. It also details length of course and previous domicile (place of residence).

Short-term migrant students in the BIS and WG datasets are those who are domiciled abroad, aged 17 to 59 and whose course is expected to last 3 to 12 months which only include those who receive funding from the UK government and those studying at further education level. As with the HESA Student Record data, the data are split by academic year. The data are received as aggregates by length of study, local authority and previous domicile. Therefore, it is not possible to investigate the proportion of students that start their study in July.

For the 2014 estimate onwards the previous year’s distribution was used as the domicile variable is no longer being collected (BIS) or not on the same basis (WG). Further research is to be undertaken in order to decide whether this method should continue or be replaced.

Confirmations of Acceptance to Study (CAS)

Short-term immigration estimates for non-EEA FE students living in local authorities outside London are distributed using CAS data from the Home Office.

It should be noted that historically CAS data covered those applying for both grants of extensions and visas2. This was a key difference between the CoS data on workers, where these two groups could be separated, and the CAS data on students, where they could not. Where a combination of in and out of country applications were previously used, only out of country visa applications are now used for the distribution of non-London visitors. Other issues relating to CoS data, such as its basis on institution address rather than residential address, also apply to CAS data.

In the CAS data, short-term migrant students are those whose certificates state that their course is expected to last 3 to 12 months. The CAS data include the name of the institution at which the student is studying. All students attending institutions that are in the HESA student record are removed in order to reduce double counting. CAS data are split into quarters and so exactly match ONS’s mid-year definition.

Grants of extensions of leave to remain for study

Short-term immigration estimates for non-EEA FE students living in local authorities in London are distributed using data from the Home Office on grants of extensions of leave to remain applied for from within the UK by migrants planning to stay for 3 to 12 months. These data relate to extensions and hence those covered are not new inflows. The extensions data are based on residential address and have the same characteristics for both work and study (for more information, please see section 3.1, grants of extensions of leave to remain for work).

In the extensions data, short-term migrant students are those granted an extension of leave to remain for the purpose of study and for a period of 3 to 12 months. Representatives from the Home Office have stated that student visas tend to be issued for a longer period than the course length in order to give migrants time to leave the UK after they finish studying. Therefore, it is possible that some short-term migrant students could have grants of extensions for over one year and consequently not be included in the distribution. However, given that there is no evidence that these different effects are biased towards particular areas (therefore skewing the distribution); it is assumed that this will not affect the distribution.

4.2. Methodology used to distribute short-term immigration for study estimates

The method for distributing England and Wales level short-term immigration for study estimate is summarised in Figure 1.

IPS data are used to derive the proportion of the short-term immigration for study estimate to be distributed by each data source. The HESA Student Record is used to produce the HE distribution. BIS and WG Learner Record data are used to produce the FE distribution for EEA domiciled students. For local authorities in all regions except London, out of country CAS data are used to distribute non-EEA FE short-term migrant students. A different methodology is used for local authorities within London, because using the sponsoring institution’s address, as is used in the CAS data, does not produce feasible estimates for London. The proportion of short-term migrant students in the London region is taken from the CAS data. It is then distributed down to local authority level using extensions data based on residential address.

Notes for: Overview of methods – students

  1. In this case, ‘EEA’ refers to all EEA member countries, as well as Switzerland. This is in line with the definition used in the worker distribution.

  2. Visas are applied for from outside the country. Those entering the UK on visas include the new arrivals which ONS is aiming to distribute, as well as individuals who have previously entered the UK and are returning on a new visa. It is assumed that where a previous visa is in existence, the period between exit on one visa and entry on the new visa is large enough that the second entrance is considered to be a second migration event. However as these individuals have not entered the UK at the time of application their subsequent residential address is unknown.

Back to table of contents

5. Quality and reliability

The IPS is a sample survey. Therefore, the England and Wales level estimates being distributed to local authorities are subject to some uncertainty – both sampling and non-sampling error. For more information, please refer to the Methodology to Estimate Short-Term International Migration for England and Wales.

Local authority level estimates are provided unrounded and to the nearest unit in publication table STIM.07. However, this is to enable and encourage further calculations and analysis; the estimates must not be assumed to be as exact as the level of detail implied by unit level data. The number of records in each administrative data source used in the distribution ranges from between 5,000 and 150,000. Given that these sample sizes are large enough to be confident in the distributions, ONS are confident that the method is sustainable over time.

Back to table of contents

6. Appendix A

Worked example – calculating local authority short-term immigration estimatesThis example shows the calculations for short-term immigration estimates for four local authorities, A, B, C and D for one mid-year

IPS Data and Administrative Data Proportions

Calculating short-term immigration for work

Short-term immigration estimates for work for nationals of EU8, EU2 and EEA countries (see Table A1) are distributed using the MWS (see Table A3 below).

Short-term immigration estimates for work for nationals of non-EEA countries are distributed using a combination of CoS and extensions data. Only the extensions data is used to distribute these estimates for London.

Therefore, the total estimates for short-term immigration for work are as shown in Table A5.

Calculating short-term immigration for study

Short-term immigration estimates of those studying in Higher Education are distributed using data from the HESA Student Record. Short-term immigration estimates for EEA students studying in Further Education are distributed using BIS and WG Learner Record data. This is shown in Table A6.

Short-term immigration estimates for non-EEA students studying in Further Education are distributed using a combination of the CAS and extensions data. Only the extensions data is used to distribute these estimates for London.

Therefore, the total estimates for short-term immigration for study are as shown in Table A8.

Calculating short-term immigration estimates

The total short-term immigration estimate for each local authority is shown in Table A9, calculated by summing the work and study totals together.

Back to table of contents