Skip to content

A Comparison of International Estimates of Long-Term Migration This product is designated as National Statistics

Released: 27 November 2014 Download PDF

Summary

This paper analyses UK estimates of long-term international migration alongside estimates from six countries - Germany, Poland, Spain, Italy, Australia and Canada – from which there are relatively large annual flows of migrants to and from the UK1, and for which international migration estimates are readily available.

The aim of the paper is to set the scene for further research into how we could potentially use estimates of international migration from other countries to validate UK migration estimates.  It provides a brief overview of initial comparative work.  In particular, the paper aims to: 

  • set into an international context UK estimates of immigration, emigration and net migration;

  • set out the difficulties of using estimates of international migration from statistical producers in other countries (‘mirror statistics’) to validate UK migration estimates.

Notes

1. See Table 3.20 

Key findings

  • In 2011 and 2012, Germany had higher net migration than the UK, with figures of 279,000 and 369,000 respectively, reflecting a rapid increase in immigration since 2008. Prior to 2012, of the EU countries analysed in this report, Italy generally had the highest net migration peaking at 476,000 in 2007. The exception to this was 2005 and 2006 when the UK recorded the highest net migration, peaking at 320,000 in the year ending June 2005.

  • Between the year ending June 2007 and the year ending June 2012, Australia, Canada and the UK demonstrated fluctuating levels of net migration, with little similarity in trends shown. Net migration to the UK has been relatively volatile, peaking at 320,000 in the year ending June 2005, following EU enlargement, and reaching a low of 205,000 in the year ending June 2009, around the time of the 2008 economic downturn. Net migration figures for Australia rose from 105,000 in the year ending June 2004 to 300,000 in the year ending June 2009. This was followed by a decrease to 196,000 in the year ending June 2010, and since then net migration remained steady.  Canadian mid-year estimates for net migration have remained steady between the year ending June 2008 (253,000) and the year ending June 2012 (67,000).
  • UK estimates often differ from the estimates of other countries measuring the same flow of migrants. Direct comparisons are very difficult due to 3 main factors; (i) differing methods of data collection, some of which may be more rigorously collected than others or may be missing all together; (ii) a diverse range of definitions for long-term migration regarding the country of origin, the duration of stay and the intention of the migrant and (iii) differences in the timing of collection and the publication of estimates.

1. UK migration estimates in an international context

Figure 1 compares net migration estimates for the UK, with those of Germany, Poland, Spain and Italy. These are EU countries which have appeared frequently since 2002 in Table 3.20: the top 15 countries of last or next residence for International Passenger Survey (IPS) estimates of long-term international migration. ONS applies the UN definition of a long-term 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 a year (see  ’Comparability’ in section 2). Net migration is defined as the difference between immigration (the number of migrants entering a country) and emigration (the number of migrants departing a country).  A positive estimate indicates that more migrants arrived in a country than departed in a given time period.

Figure 1: Calendar year net migration estimates for the UK and selected EU countries, 2002-2012

Figure 1: Calendar year net migration estimates for the UK and selected EU countries, 2002-2012

Notes:

  1. Sources: Central Statistical Office Poland, Instituto Nacional de Estadistica, Destatis, I.Stat, Office for National Statistics, Long-Term International Migration (LTIM)

Download chart

Figure 1 shows that there are differences between the net migration figures for the UK and other European countries. Between 2002-2012, calendar year net migration estimates for the UK have remained between 172,000 and 273,0001. During this time, UK immigration figures decreased from their peak of 596,000 in 2006 to 498,000 in 2012. Likewise, UK emigration figures decreased from the peak of 427,000 in 2008 to 321,000 in 2012.

Poland also showed relatively steady estimates of net migration since 20092, although in contrast to the UK, it experienced negative net migration in 2008 and 2012, with more people emigrating than immigrating.

Net migration to Germany was variable between 2002 to 2012, with a steady decline from 2002 to 2008, followed by a sharp increase to 2012. In 2011 and 2012, net migration to Germany was higher than net migration to the UK, driven by a sharp increase in immigration (although some of these will be tt-term migrants, see Section 2).

Net migration in Italy varied between 171,000 and 476,000 from 2002 to 2007. After 2007 there was a decline in net migration, reaching 245,000 in 2012, a similar level to 2006. Net migration reflects the same pattern seen for immigration to Italy.  Following an initially rapid increase in the number of migrants entering the country, at the time of  the 2008 economic downturn there was a decrease in immigrants which has continued ever since. Conversely, since 2008 there has been a gradual increase in the numbers of emigrants leaving, concluding with a relatively large increase from 82,000 in 2011 to 106,000 in 2012.

Finally, Spain has demonstrated a rapid decrease in net migration from 2008 to 2012, with 2010 being the year that the number of people leaving the country exceeded those arriving for the first time. Over the period where data is available, there has been a decrease in immigration to Spain and an increase in emigration from Spain.

Figure 2 compares the net migration figures for two non-EU countries, Australia and Canada3, with the UK. These are non-EU countries shown frequently since 1975 in the top 15 list of countries of last or next residence from the IPS, and which also provide readily accessible data online. India has featured in the top 3 countries of last residence every year since 2003, and China has been in the top 10 countries of last residence every year since 1999. However official migration data from these countries is not readily available.

Figure 2: Mid-year net migration estimates for the UK and selected non-EU countries, 2002-2012

Figure 2: Mid-year net migration estimates for the UK and selected non-EU countries, 2002-2012

Notes:

  1. Sources: Australian Government Department of Immigration and Border Protection[4], Statistics Canada, Office for National Statistics

Download chart

Estimated net migration for the UK has fluctuated over the last decade. Net migration peaked at 320,000 in the year ending June 2005, following enlargement of the EU. Subsequently, net migration decreased reaching 205,000 in the year ending June 2009, at the time of the recent economic downturn. Net migration subsequently increased again to 263,000 in the year ending June 2011, followed by a further decrease in the year ending June 2012 to 167,000.

Net migration figures for Australia rose from 105,000 in the year ending June 2004 to 300,000 in the year ending June 2009. This was succeeded by a decrease in the year ending June 2010 to 196,000, before remaining steady until the year ending June 2012. This was mainly due to fluctuations in immigration estimates during this time period, with the lowest mid-year estimate being 297,000 in the year ending June 2004 and the highest mid-year estimate being 520,000 in the year ending June 2009. Conversely, emigration estimates gradually increased from 193,000 in the year ending June 2004 to 269,000 in the year ending June 2012. A contributory factor to the changes in both the immigration and emigration figures could be the introduction in 2007 of the’12/16 rule’ by the Australian Bureau of Statistics (ABS), which defined anybody present in Australia for 12 months or more, over a 16 month reference period, as an arriving Net Overseas Migrant and conversely, anybody absent from Australia for the same period of time as  a departing Net Overseas Migrant.

In contrast, Canadian mid-year estimates for net migration have remained fairly consistent between the year ending June 2008 (253,000) and the year ending June 2012 (267,000). Unlike Australia and the UK, Canada calculate net migration by adding together estimates for the numbers of immigrants, ‘returning emigrants’ (Canadian residents or immigrants having previously emigrated from Canada and subsequently returned to Canada to re-establish a permanent residence) and net flows of temporary residents (for example students, temporary workers and refugees). Deducted from this are the number of emigrants (who sever residential ties) and net temporary emigrants (Statistics Canada, 2012).

Definitional differences between the three countries will account for some of the observed differences in net migration. The net migration figures should also be viewed in the context of the relative population size of each country. For the year ending June 2012, the UK had by far the highest population of the three countries with 63.7 million usual residents, Canada was second with 34.8 million usual residents and Australia had the smallest population with 22.7 million usual residents. While the UK also had the highest level of immigration of the three countries, Australia had higher immigration than Canada even though it has a smaller overall population. Other factors which might affect immigration would include geographical location, government policies relating to migration and the overall political and economic climate .

Notes for 1. UK migration estimates in an international context

  1. This is based on revised calendar year estimates in light of the 2011 Census. The highest net migration estimate between 2002 and 2012 was 320,000 in year ending June 2005. Conversely, the lowest net migration estimate was 154,000 in the year ending September 2012.

  2. There are no net migration estimates available for Poland or Spain prior to 2008.

  3. There are no net migration figures for Canada prior to 2006/2007.

  4. In 2005–06 the method for calculating NOM changed in 2006 with the introduction of the 12/16 rule 2005–06, up to 2009–10 this table is based on final NOM data calculated by the Australian Bureau of Statistics (ABS), data for later years is based on DIBP forecasts of the components of NOM and preliminary NOM estimates from the ABS.

2. Use of ‘mirror statistics’ to compare and improve estimates of international migration

‘Mirror statistics’, in the context of migration, involves comparing two measures of the same flow of migrants produced by different countries, for example, comparing UK figures for emigration from the UK to Germany with German figures for immigration from the UK to Germany.  This analysis highlights differences in the way migration is measured, and may potentially be used to validate estimates of migration and possibly to identify ways to improve estimates of migration.

In 1976 the UN first made recommendations regarding the use of international migration data, however in 1998, it discussed revising these, realising the need to address the issues concerning lack of comparability of the migration data of different countries (UN, 1998).

The Organisation for Economic Co-operation and Development (OECD) also recognised the difficulties associated with harmonising data, and in a report (Lemaitre et al., 2006) calculated the differences between harmonised datasets  and the original published figures. Some of its recommendations for improvements included breaking down aggregate data (for example by nationality and sex) to improve understanding, extending statistics further into the past, and monitoring more closely phenomena such as changes to free movement regimes and ‘regularisations’.

ONS has previously published research investigating the potential use of immigration estimates of other countries which share large migration flows with the UK to improve emigration estimates (Dini, Horsfield and Vickers, 2007).

Eurostat later commissioned the three year MIMOSA project (Raymer & Abel, 2008) in which applied estimation methods of flows between 31 EU countries were developed, breaking these estimates down by  variables such as sex and age. Emigration numbers of sending countries were made consistent with immigration numbers of receiving countries. The aim was to account for the differences in definition, quality and sources of available migration data and to replace missing data with these estimates.

Comparability

There are differences in how countries define, measure and report on migration. These present inconsistencies when comparing other countries’ data with UK data.

According to the UN, a long-term migrant is defined as
“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.”

The UN definition of a short-term migrant is
“a person who moves to a country other than that of his or her usual residence for a period of at least three months but less than a year (12 months) 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.”

The UK and Italy apply the UN definition of a long-term migrant, while Germany and Spain do not specify a time period, and as such will include both short- and long- term migrants in their estimates (United Nations, 2011).

According to the UN, Australia, Canada, and Poland use ‘establishment of permanent residence’ as the definition of long-term immigration. Regarding emigration, Australia and Poland have ‘establishment of permanent residence abroad’ as the defining factor. While the UN (2010) states that “no migrant outflows are available for Canada”, StatCan refers to emigrants as “Canadian citizens or immigrants who have left Canada to establish a permanent residence in another country”. See the Annex for a summarised table of definitions applied.

There is also variation in the methodology for recording migrants:

  • UK migration data is based on intentions. These may change, for example a person may intend to stay for 12 months or more but leave within 12 months, becoming a migrant switcher. This is adjusted for in ONS migration estimates.

  • There may be a lag between arrival in a country and an application for permanent residence. By contrast, UK migration is measured at the point of arrival in the country. 

  • The UK produces migration datasets based on ‘Country of Birth’, ‘Nationality’ or ‘Country of previous or next residence’. However this is not the case for all countries. For example, while most of the countries in this report record the country of last and next residence of migrants, the USA reports migration based on country of birth. For instance a migrant born in Germany and living in the UK but then migrating to USA would be recorded as being from Germany not the UK. This would not be comparable with UK estimates of migration by country of last/next residence.

Accuracy

There are known accuracy problems in the measurement of migration.

In the UK, long-term international migration (LTIM) estimates are based on data from the International Passenger Survey (IPS), with adjustments made for asylum seekers, non-asylum enforced removals, migrant and visitor switchers and flows to and from Northern Ireland.  As such UK estimates of migration are based largely on a sample survey and so there are sampling and non-sampling errors around the estimates.

The other European countries in this study measure migration based on registration and de-registration on population registers. While covering more of a population than a sample survey, the disadvantage of this is that it may include both short- and long-term migrants (Germany is an example of this, see below).  Emigration estimates may be lower as individuals may be discouraged from de-registering if this would result in them sacrificing certain rights, such as the right to vote, access to social services or healthcare. This means that they would still be legally resident in their country of origin, while in fact being part of the resident population in a different country.

 Likewise there may be disincentives for an immigrant to register, such as taxation authorities, local police and security forces having access to the details of the immigrant.  Some countries may operate their registers more stringently than others.  For example, in Germany registration with a local authority is obligatory.  It is difficult to differentiate between short- and long-term migrants as anyone who registers and is present longer than a week is considered an immigrant. Information on immigration of foreign nationals can partially be extracted from the Central Register of Foreigners, but this would not include returning German nationals as they use separate registers. Indeed, there are many different registers to record migrants in Germany. As anyone who registers after a week can be included in the migration data, the figures will inevitably be higher than other countries where a longer time criterion is set.

Accessibility

Estimates of migration to and from the UK from a specific country of origin are not always readily available. Furthermore, some of the time series presented in this report are missing data from earlier years. Many non-EU countries such as India and China do not supply readily accessible data online.  This can in part be overcome by looking at the data collection of international bodies. European countries can be checked against Eurostat data and potentially supplemented by it. Other international bodies such as the OECD and the World Bank offer further information. The UN Department of Social and Economic Affairs Population Division monitor and publicise the different migration monitoring methods of the individual countries, but, as previously stated, recommend harmonisation of these to improve comparability (UN, 1998).

Timeliness

ONS publishes provisional estimates of long-term international migration five months after the reference period. If emigration data from other countries were used in the calculation of immigration estimates to the UK then this would result in a longer time lag between the reference period and publication. This is due to other countries measuring emigration by identifying a lack of activity in their administrative sources for at least one year.

Even a simple comparison of annual immigration, emigration and net migration estimates between countries can be difficult as each country will submit their data at different times of the year and collect it for different reference periods. For example, the Australian Government’s  Department of Immigration and Border Protection (DIBP) publishes their annual migration figures in September for their fiscal year (1 July to 30 June). Citizenship & Immigration Canada organise their data by calendar year, whereas Statistics Canada, like Australia, use their fiscal year (1 July to 30 June) as a basis for their figures, but publish in October. The UK publishes final calendar year and mid-year estimates in November, in addition to quarterly provisional estimates with a 5 month lag to the reference period.

The Annex summarises the differences in definitions and methodologies used to measure international migration across the countries included in this report.

3. Case Study: Using mirror statistics to compare migrant flows into and out of the UK

Figure 3 compares the ONS figures for immigration and emigration flows between the UK and Australia with the Australian Government’s DIBP figures for the same flows.

Figure 3: Mid-year emigration estimates for the UK and selected non-EU countries, 2002-2012

Figure 3: Mid-year emigration estimates for the UK and selected non-EU countries, 2002-2012

Notes:

  1. Sources: Australian Government Department of Immigration and Border Protection, Statistics Canada, Office for National Statistics

Download chart

The UK immigration figures are measuring the same flow as the Australian emigration figures (both in red), while the UK emigration figures are measuring the same flow as the Australian immigration figures (both in blue). Although the series do show similar trends, the UK estimates are higher than those of Australia (with the notable exception of the 2012 data for the migrant flow into Australia).

A notable issue of comparability is that the UK figures for flows to and from Australia are based on calendar years whereas the equivalent Australian figures are based on mid-years, which is the Australian fiscal year. UK mid-year data for flows specific to Australia are unavailable. For example, the year marked as 2006-2007 will be from 1 July 2006 to 31 June 2007 for the Australian figure, but for the UK will cover 1 January 2007 to 31 December 2007.

Figure 4 is a comparison of the ONS figures for the number of migrants departing from the UK and arriving in Canada with Citizenship and Immigration Canada’s (CIC) figures for the same flow. Emigration data to specific countries from Canada is not available.

Figure 4: Canadian and UK immigration flow estimates between Canada and UK, 2002-2012

Figure 4: Canadian and UK immigration flow estimates between Canada and UK, 2002-2012

Notes:

  1. Sources: Citizenship and Immigration Canada (CIC), Office for National Statistics (ONS), International Passenger Survey (IPS).

Download chart

According to CIC, the number of immigrants into Canada from the UK increased until 2009. From 2010 onwards the number of immigrants from the UK began to decrease. By contrast the UK figures show a completely different trend. Year on year figures fluctuate which may reflect sampling variability. Canada’s definition of a long-term migrant includes newly arrived migrants and those changing status once temporarily resident, based on residence permits.

Figure 5 is a comparison of the ONS data for migration flows between the UK and Spain with that of Instituto Nacional de Estadistica (INE).

Figure 5: UK and Spanish migration flow estimates between UK and Spain, 2002-2012

Figure 5: UK and Spanish migration flow estimates between UK and Spain, 2002-2012

Notes:

  1. Sources: Instituto Nacional de Estadistica, Office for National Statistics (ONS), International Passenger Survey (IPS).

Download chart

The ONS data for migrant flows from the UK (in dashed blue) were fairly steady from 2002 to 2006, before a decline until 2009, after which the estimates have remained stable. Similarly to the emigration figures, the number of migrants arriving in the UK from Spain was also relatively stable between 2002 and 2008, however there was a peak of 22,000 immigrants in 2009. Similarly, there was  a statistically significant increase in migrants from Spain, from 17,000 in 2011 to 27,000 in 2012. It is clear that emigration figures to Spain were far higher than immigration figures from Spain for the period 2002 to 2007, however after this time the numbers became closer to each other and since 2010, flows for migrants arriving into the UK from Spain have been higher than those departing for Spain. 

Spanish and UK figures for the migrant flow from the UK into Spain (in blue) from 2008 to 2012 clearly mirror each other, but the Spanish figures are somewhat higher than the UK figures. This may be because the Spanish immigration figures would include short- and long-term immigrants. In contrast, the figures for the number of migrants travelling from Spain to the UK (the red lines) differ completely each year, with no similarity in trend, although the levels were similar in 2008 and 2011.

Figure 6 is a comparison of the ONS data for migration flows between the UK and Poland with that of the Central Statistical Office of Poland.

Figure 6: UK and Polish migration flow estimates between UK and Poland, 2002-2012

Figure 6: UK and Polish migration flow estimates between UK and Poland, 2002-2012

Notes:

  1. Sources: Central Statistical Office of Poland, Office for National Statistics (ONS), International Passenger Survey (IPS).

Download chart

The UK data in Figure 6 shows a rapid increase in immigration from Poland shortly after Poland’s accession to the EU in 2004. It is also notable that emigration from the UK to Poland almost trebled in 2008, during the most recent economic downturn. Despite having the same definition of a long-term migrant, the different methods of data collection result in the Polish estimates being much lower than their UK equivalent (see Annex).

Figure 7 shows a comparison of the ONS data for migration flows between the UK and Italy with those produced by the Italian National Institute of Statistics (IStat).

Figure 7: UK and Italian migration flow estimates between UK and Italy, 2002-2012

Figure 7: UK and Italian migration flow estimates between UK and Italy, 2002-2012

Notes:

  1. Sources: I.Stat, Office for National Statistics (ONS) - International Passenger Survey (IPS).

Download chart

Conclusion and Next steps

This paper offers a brief overview of the difficulties inherent in comparing international migration estimates produced by different countries. It shows how UK data differs greatly from the data of other countries measuring the same flow of migrants. Direct comparisons are very difficult due to 3 main factors; (i) differing methods of data collection, some of which may be more rigorously collected than others or may be missing all together (ii) a diverse range of definitions for long-term migration regarding the country of origin, the duration of stay and the intention of the migrant (iii) differences in the timing of collection and publication of data.

Having set the scene, ONS plans to undertake further research to assess the extent to which estimates of international migration produced by statistical producers in other countries could validate UK migration estimates. As part of this ongoing work, ONS will collaborate closely with these producers in order to overcome some of the issues outlined in this preliminary research. 

Annex

Australia

Department of Immigration and Border Protection (DIBP) data define long-term immigrants as those granted permanent visa status in a given year (except immigrants from New Zealand, who have the right to live and work in Australia). Figures do not include those on student, skilled graduate, or any other temporary visas. Long-term emigration statistics are collected from passenger cards as residents leave the country. These data only include those intending to leave permanently.

Australian Bureau of Statistics estimates of the usually resident population include residents who have been in the country for twelve of the sixteen months prior to when the estimate is made. They also include students and other temporary residents.

Canada

Statistics Canada and Citizenship and Immigration Canada (CIC) define long-term immigrants as those granted the right to live in Canada permanently by immigration authorities in the given year, including those who have recently arrived in the country and those who have changed status while temporarily resident.

Net migration is calculated by adding together estimates for the numbers of immigrants,  ‘returning emigrants’ (Canadian residents or immigrants having previously emigrated from Canada and subsequently returned to Canada to re-establish a permanent residence) and net flows of temporary residents (for example students, temporary workers and refugees). Deducted from this are the number of emigrants (who sever residential ties) and net temporary emigrants.

Germany

Destatis long-term migration figures are based on local population registers and the central register of foreigners. Long-term immigrants are defined as those resident for more than twelve months. Figures do not include military and diplomatic personnel and those not registered.

Estimates quoted in this paper include both long- and short- term migrants.

Italy

I.Stat measure international migration using a population register. A pre-requisite of registration is a residence permit, with an ongoing requirement for regular demonstrations of continuous residence by the migrant in the municipality in which they reside.

Poland

The Central Statistical Office of Poland estimate long-term migration based on registered migration on their population register validated by data from other countries to estimate and correct for the underestimation of migration as recorded on the population register.

Spain

The Instituto Nacional de Estadistica (INE) include both long- and short- term migrants in their estimates which are based on registration and de-registration from their population register.  In Spain laws were passed in 2005 to enforce re-registration every 2 years to avoid removal from the registers (Sandell, 2006). It is worth noting that Spain officially include illegal migrants in their counts.

United Kingdom

The Office for National Statistics (ONS) base Long-Term International Migration (LTIM) estimates on data from the International Passenger Survey (IPS), supplemented by adjustments for asylum seekers and their dependents, non-asylum enforced removals, flows into and out of Northern Ireland and visitor and migrant switchers. Estimates do not include military and diplomatic personnel.

 

Glossary

Harmonised datasets

Sets of data from different sources that are standardised based on common classifications, definitions and standards. The intention is to improve comparability.

International Passenger Survey (IPS)

The International Passenger Survey (IPS) is a survey of a random sample of passengers entering and leaving the UK by air, sea or the Channel Tunnel. Between 700,000 and 800,000 people are interviewed on the IPS each year. Of those interviewed, approximately 4,000-5,000 people each year are identified as long-term international migrants.

Regularisation

An immigrant gaining ‘legality of residence’. Regularisation campaigns encourage immigrants to sign up to the population register and are sometimes considered an ‘amnesty’. (International Centre for Migration Policy Development 2009)

References

Dini, E., Horsfield, G. and Vickers, L. (2007) Using data from overseas to improve estimates of emigration, Population Trends (ONS).

International Centre for Migration Policy Development (2009) Regularisations in Europe.

Lemaitre, G. Liebig, T. Thoreau, C. (2006) Harmonised statistics on immigrant inflows-preliminary results, sources and methods.

Raymer, J. and Abel, G. (2008) The MIMOSA model for estimating international migration flows in the European Union, UNECE/Eurostat work session on migration statistics Working papers, (8).

Sandell, R. (2006) Spain’s Immigration Experience: Lessons to be Learned from Looking at the Statistics.

Office for National Statistics (2012) Table 3.20: IPS Top Countries of Last or Next Residence by Citizenship, 1975-2012.

Statistics Canada (2012) Chapter 6. Emigrants, net temporary emigrants and returning emigrants. In: Population and Family Estimation Methods at Statistics Canada.

United Nations Department of Economic and Social Affairs Statistics Division (1998) Recommendations on statistics of international migration.

United Nations Department of Economic and Social Affairs Population Division (2011) International migration flows to and from selected countries: the 2010 revision.

 

Background notes

  1. Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: media.relations@ons.gsi.gov.uk

    These National Statistics are produced to high professional standards and released according to the arrangements approved by the UK Statistics Authority.

Content from the Office for National Statistics.
© Crown Copyright applies unless otherwise stated.