3 Mapping migration in the European Union
In this chapter, we take a closer look at the amount, composition and dynamics of immigration into and within the EU, considering the source countries and other characteristics, such as education level, age and gender. We start with total immigration into, and emigration from, EU member states. We then look specifically at intra-EU mobility, including so-called ‘posted workers’, and at immigration from outside the EU, including the recent wave of asylum seekers. We also present data on the share of foreign citizens and foreign-born people in EU countries and their origins.
3.1 Annual immigration and emigration trends
Figure 13 shows trends in the movement of people from 2009-15 for the aggregate of 27 EU countries and three groups of EU countries. We separate home-country citizens and foreign citizens – the foreign-citizens category includes people from other EU countries and non-EU countries (including refugees).
Annual emigration of home citizens (to both EU and non-EU destination countries) has been relatively stable at about 0.28 percent of population on average across the countries considered in Figure 13, with a minor increase recently. This minor increase is entirely due to increased emigration from southern EU countries from a very low level of 0.09 percent in 2009 to 0.23 percent in 2015. High unemployment in southern EU countries was the main driver of the increase, yet it is notable that even after the increase, the annual emigration flow remains below the EU average. The highest levels of emigration rates are from the central and eastern EU members, with a relatively stable emigration rate of about 0.53 percent of population per year, more than the double than rate in north-west and southern EU countries.
Return migration – the immigration of home citizens – is also highest in central and eastern European countries, though returns do not compensate for emigration. On average, central and eastern European countries had the largest net emigration rates of 0.23 percent of population in 2015, followed by the four southern EU countries at 0.13 percent and the 10 north-west EU countries at 0.08 percent. Net emigration of home-country citizens was recorded by 25 of the 28 EU countries in 2015 (Figure 14). The largest net outflows of home citizens were from Lithuania (-0.64 percent of population), Latvia (-0.58 percent) and Croatia (-0.52 percent), while the only three countries where citizens returned on a net basis were Cyprus (0.25 percent), Malta (0.12 percent) and Denmark (0.10 percent). Figure 14 also shows that emigration of home citizens from Italy and Spain remained rather low compared to other EU countries, even if there was some increase in the preceding years.
A clear message from Figure 13 is that immigration of foreign citizens increased significantly from 2009-15, primarily because of refugee inflows to north-west EU countries. In 2015, the largest inflow of foreign-citizens as a share of population was to Luxembourg (4.02 percent of population), followed by Malta (2.61 percent), Austria (1.83 percent) and Germany (1.79 percent). In terms of number of people, Germany received the bulk of asylum seekers, as we discuss later in this chapter.
Figure 13: Trends in immigration and emigration: home and foreign citizens, 2009-15 (percent of population per year)
Source: Bruegel based on Eurostat ‘Emigration by age group, sex and citizenship [migr_emi1ctz]’, ‘Immigration by age group, sex and citizenship [migr_imm1ctz]’ and ‘Population on 1 January by age and sex [demo_pjan]’ datasets. Note: EU27: current EU members except Bulgaria. North-West 11: Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Luxembourg, Netherlands, Sweden, United Kingdom. CEE 10: Croatia, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia and Slovenia. South 4: Greece, Italy, Portugal and Spain. Belgian data is not available for 2009: we assumed that the growth rate (in terms of number of people) from 2009 to 2010 was the same as in the Netherlands. For Latvia, total flows are available for the full period of 2009-15, but the home citizen/foreign citizen breakdown only from 2011-15. For 2009-10, was assumed that the annual growth rate (in terms of number of people) of foreign citizens was the same as in Lithuania.
Some of immigrants go home: the emigration of foreign citizens from the EU has been relatively stable at about 0.25 percent of EU population annually. Net immigration of non-EU citizens has increased from about 0.25 percent of EU population in 2009 to about 0.5 percent in 2015, while EU citizens are leaving the EU at a rate of about 0.1 percent of population on average.
Figure 14: Emigration and immigration of home-country citizens in 2015 (percent of population)
Source: Bruegel based on Eurostat ‘Emigration by age group, sex and citizenship [migr_emi1ctz]’, ‘Immigration by age group, sex and citizenship [migr_imm1ctz]’ and ‘Population on 1 January by age and sex [demo_pjan]’ datasets.
3.2 Intra-EU mobility
The central and eastern EU member states have the most mobile populations. Citizens of these countries primarily moved to the west, including Italy and Spain.
East-west mobility
The eastern enlargements of the European Union in 2004 (10 countries), in 2007 (Bulgaria and Romania) and in 2013 (Croatia) increased the potential for east-west mobility within the EU, given that labour mobility is one of the fundamental freedoms of the EU.
However, the accession treaties with new EU member states allowed the older countries to impose transition periods of up to seven years, during which immigration restrictions on the citizens of newer member states could be maintained. The older member countries also had the option to introduce such controls during the seven-year transition period, even if they had abolished the restrictions earlier, provided that there was a serious disturbance of their labour markets.
In 2004, twelve of the fifteen older EU members used this option and adopted temporary immigration controls, but the UK, Ireland and Sweden opened their labour markets directly from 1 May 2004 for nationals of the eight central European countries (EU8) that joined the EU on 1 May 2004. When Bulgaria and Romania joined, ten countries opened their labour markets already from 2007, while the other 15 countries imposed restrictions on free movement (usually requiring a work permit). These temporary restrictions were gradually removed during the seven-year transition periods. It is notable that among all older EU members, there was only one instance of new restrictions being introduced after a complete abolition of restrictions: Spain introduced new controls on immigrants from Romania in August 2011, after having removed all controls in 2009.
The right of movement from central and eastern European countries to western Europe was rapidly taken up. For the 2004 entrants, the peak was in 2006. However, the peak of migration from Romania and Bulgaria was also in 2006 – one year before they entered the EU (Figure 15).
In terms of the impact on destination countries, the United Kingdom and Ireland received disproportionally large inflows after 2004. The non-introduction of temporary controls by these two countries likely diverted immigration to them.
The big surge in migration from Bulgaria and Romania in 2006 is largely attributable to migration from Romania to Italy and Spain. The big surge suggests there was a pent-up desire to move to western Europe before these countries entered the EU. However, it is also noteworthy that the net inflow from Bulgaria and Romania in 2003-05 – before these countries entered the EU – was practically the same (even slightly higher) than migration in 2008-12, after the enlargement-related surge abated. Evidently, many people from Bulgaria and Romania were able to find ways to move to western Europe well before these countries entered the EU, while the subdued 2008-12 flows could be explained by the temporary mobility restrictions that were imposed by receiving countries, and the increase in unemployment in the two main destinations countries, Spain and Italy.
Figure 15: EU8 and Bulgarian/Romanian (EU2) citizens living in subgroups of older EU members, 2000-15 (in thousands)
Source: Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset, code: migr_pop1ctz, and UK Office for National Statistics ‘Population of the UK by country of birth and nationality’ dataset. Note: the top panel shows 1 January data of a given year as 31 December data of the previous year, so that the difference in year-end data shows the annual net flows, as reported in the bottom panel. EU8 countries: Czech Republic, Estonia, Hungary, Poland, Lithuania, Latvia, Slovenia, Slovakia. EU2 countries: Romania and Bulgaria. The destination countries for EU8 includes 12 older EU member states, not including France, Greece and Spain, while the destination countries for EU2 includes 13 older EU member states, not including France and Greece. Therefore, EU8 and EU2 data are not comparable because Spain is not included among the destination countries for EU8 but included for EU2. There are a few missing values in the Eurostat database, which we approximated in order to have constant country-composition aggregates. Missing UK data was approximated using data from the UK Office for National Statistics data. Irish data is missing for 2000-06: we assumed that the growth rate of immigrants from EU8 and EU2 was the same as in the UK. The 2004-07 missing data for Belgium was approximated by using the growth rate for the Netherlands. Similarly, missing data points for Luxembourg were approximated using Dutch growth rates.
Country-specific mobility numbers show that Germany and the UK are main destination countries for migrants from the central and eastern EU members (Table 1), but there are notable exceptions. Romanians preferred to move to Italy and Spain: almost three times as many Romanians moved to Italy and Spain combined than to Germany and the UK combined. Estonians primarily move to Finland because of geographical closeness and cultural similarity. Austria is the second most important destination country for neighbouring Croatia and Slovenia, and the third most important destination country for Hungary.
Finally, the minor role of France is noticeable. The number of central and eastern European citizens in France is just about one-tenth of the German figure. Distance is unlikely to be a main reason for this, because Spain – further than France – is a more popular destination country for migrants from Bulgaria, Lithuania, Poland and Romania (countries for which Spanish residency data is available).
The crisis had an influence on east-west migration patterns, because of diminished labour market opportunities in some key destination countries, such as Italy and Spain. This was reflected in a slowdown in mobility from the east to the west (Figure 15). However, economic and labour market developments in certain central and eastern European countries also had a major influence on migration patterns. There was a massive exodus from hard-hit Latvia and Lithuania, where GDP and employment fell by 10-20 percent (Darvas, 2013). The population declined by more than 10 percent in these countries from 2008-12.
Table 1: The main EU destination countries for central and eastern European migrants, stocks of foreign citizens as of January 2016 (thousand people)
Source: Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset, dataset code: migr_pop1ctz. Note: blue: main destination country, red: second main destination country.
In Latvia and Lithuania, most of the population decline in 2008-12 was a result of the emigration of young cohorts and the resident population of younger people declined by about 20-25 percent, causing major losses to these countries. Large-scale emigration from Latvia and Lithuania continued in 2013-16.
Since about 2013, emigration from central and eastern European countries has started to increase again (Figure 15), most likely because of better labour market opportunities in the main destination countries. Croatia, which entered the European Union in 2013, has also seen increased emigration.
Even during the crisis, there was no mass return migration, as noted by Zaiceva and Zimmerman (2012). Instead, the rate of return migration to the central and eastern European countries has proved to be relatively stable (Figure 13). However, there are country-specific differences (Table 2). For example, by 2016 return migration to Bulgaria, Hungary and the three Baltic countries had increased, while to the Czech Republic, Poland and Romania it declined.
Table 2: Return migration flows (thousands)
|
Bulgaria
|
-
|
4.9
|
10.7
|
|
Czech R.
|
21.7
|
8.1
|
4.5
|
|
Estonia
|
1.7
|
2
|
8
|
|
Croatia
|
-
|
4.7
|
6.5
|
|
Hungary
|
2.3
|
5.5
|
32.6
|
|
Latvia
|
0.5
|
1.5
|
5
|
|
Lithuania
|
4.8
|
14
|
18.4
|
|
Poland
|
142.3
|
102
|
84.8
|
|
Romania
|
124.9
|
138.4
|
115.5
|
|
Slovakia
|
1.2
|
1.1
|
3.2
|
|
Slovenia
|
2.9
|
3.3
|
2.8
|
Source: Eurostat ‘Immigration by age group, sex and citizenship’ dataset, code: migr_imm1ctz.
Even though return migration flows are not large, it is useful to look at the profile of returnees to central and eastern European countries. They are mostly 25-44 year olds – the most mobile age bracket (Figure 16). For the Czech Republic, almost two-thirds of returnees are in the 24-45 age bracket; for Slovenia, the same age group makes up one third of returnees. A typical returnee to a central and eastern European country is below the age of 45, single, male and employed in a lower-skilled job abroad despite having attained a tertiary degree, which suggests that individuals who are returning are overqualified and could signal a brain-waste (Zaiceva and Zimmermann, 2012).
Figure 16: Age profile of return migrants, 2015
Source: Eurostat ‘Immigration by age group, sex and citizenship’ dataset, code: migr_imm1ctz. Note: data for Poland, Slovakia, Slovenia and Romania is for 2013.
The EU 2010 Labour Force Survey documents the economic activity of returnees one year before the survey. The results suggest that the majority of returnees (above 60 percent in most central and eastern European countries, with the exception of Slovenia where students constitute the biggest group) were employed abroad prior to returning. Romania, Bulgaria and Latvia reported the largest shares of returnees who were unemployed while abroad.
Moreover, with the gradual return of post-enlargement migrants, which might accelerate as they start to retire, cross-border pension transfers and social remittances might become major issues (Duszczyk and Matuszczyk, 2016).
Figure 17: Southern EU citizens living in eight north-west EU countries, thousands, 2000-15
Source: Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset, dataset code: migr_pop1ctz. Note: southern Europe: Greece, Italy Portugal and Spain. North-west EU-8: other older EU countries not including France, Luxembourg and Ireland. End-2015 data is available also for these three countries: 1.04 million southern EU citizens were living in France, Luxembourg and Ireland altogether, so the omission of these three countries from the chart is major.
South-north mobility
Migration from southern (Greece, Italy, Portugal and Spain) to northern and western EU countries is also of interest. From 2000 to 2012, there was a decline in the number of southern European citizens living in northern EU countries (Figure 17). That is, even though labour markets in southern Europe were hit hard in 2008-12, there was still net return migration to southern Europe, which is puzzling. However, since 2013, an increasingly large number of southern European citizens have decided to move to northern EU countries.
Table 3: The main EU destination countries for migrants from southern Europe, stocks of migrants as of January 2016 (thousand people)
|
Greece
|
314.6
|
8.1
|
56.9
|
n.a.
|
16.7
|
2.6
|
14.1
|
5.5
|
7.0
|
8.4
|
2.4
|
0.2
|
0.4
|
|
0.9
|
|
Italy
|
557.4
|
192.8
|
195.1
|
191.6
|
156.8
|
20.3
|
29.5
|
25.1
|
-
|
8.6
|
8.1
|
6.1
|
4.8
|
5.2
|
2.4
|
|
Portugal
|
124.0
|
553.9
|
222.3
|
101.8
|
44.2
|
93.1
|
19.4
|
3.2
|
5.8
|
2.3
|
2.8
|
-
|
1.8
|
0.3
|
0.6
|
|
Spain
|
148.1
|
156.7
|
133.8
|
-
|
61.7
|
5.5
|
26.8
|
6.6
|
22.6
|
8.8
|
6.0
|
10.0
|
4.2
|
1.0
|
2.4
|
|
All countries
|
1144.1
|
911.5
|
608.0
|
293.4*
|
279.3
|
121.5
|
89.7
|
40.4
|
35.4
|
28.1
|
19.3
|
16.4
|
11.1
|
6.5
|
6.4
|
Source: Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset, dataset code: migr_pop1ctz. Note: blue = main destination country, red = second main destination country. * The total for Spain as a destination country does not include Greek citizens.
Some studies have reported a shift from east-west to south-north mobility (Fries-Tersch and Mabilia, 2015, and Barslung and Busse, 2014), as emigration from crisis-hit countries increased. However, while Figure 17 indicates a rapid increase in south-north mobility (even though the omission of France as a destination country is highly significant), the east-west flows (the sum of EU8 and EU2 net flows, Figure 15, right panel) continue to be much larger in terms of number of people, and even larger as a share of source-country population, because the combined population of Greece, Italy, Spain and Portugal is greater than the combined population of central and eastern European EU members.
Mobility within the north-west of the EU
We also look at intra-region mobility within north-west countries of the EU, such as Germans living in France and French people living in Germany. The top panel of Figure 18 shows that the level of north-west EU nationals living in other north-west EU countries increased after 2000, reaching 2.7 million in 2015. However, the flows have been rather modest (bottom panel of Figure 18), even in 2006, when the largest change during the 2000-15 period was recorded. The approximate 140,000 increase in 2006 was much lower than annual outflows from central and eastern Europe during the whole period (see the bottom panel of Figure 15), and more recently from southern Europe (see the bottom panel of Figure 17), even though the total population of north-west EU countries is more than twice that of the central and eastern European countries or the southern European countries.
Figure 18: North-west EU citizens living in another north-west EU country, thousands, 2000-15
Source: Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset, dataset code: migr_pop1ctz and UK Office for National Statistics ‘Population by country of birth and nationality underlying datasets’ dataset. Note: countries included: Belgium, Denmark, Germany, Ireland, Luxembourg, the Netherlands, Austria, Finland, Sweden and the United Kingdom. Some missing data was interpolated: Belgium: Luxembourgish growth rates applied where necessary; Ireland: UK growth rates applied where necessary; Luxembourg: Belgian growth rates applied where necessary.
Table 4: The main EU destination countries for migrants from north-western Europe, stock of migrants as of January 2016 (thousands)
|
Austria
|
14.2
|
n.a.
|
166.3
|
4.6
|
2.8
|
-
|
4.1
|
3.3
|
0.4
|
6.1
|
0.9
|
1.4
|
0.6
|
1.3
|
0.5
|
|
Belgium
|
26.0
|
32.4
|
24.5
|
96.3
|
-
|
2.2
|
30.6
|
1.5
|
0.8
|
5.5
|
19.4
|
1.1
|
2.4
|
1.2
|
0.4
|
|
Denmark
|
30.9
|
11.7
|
19.6
|
5.4
|
2.7
|
1.1
|
3.0
|
37.1
|
0.5
|
2.0
|
2.0
|
-
|
0.6
|
0.7
|
0.7
|
|
Finland
|
9.6
|
12.3
|
12.9
|
2.7
|
2.8
|
1.5
|
2.8
|
57.6
|
0.5
|
1.6
|
1.2
|
2.8
|
0.8
|
0.5
|
-
|
|
France
|
167.6
|
100.7
|
116.6
|
-
|
161.8
|
7.8
|
20.9
|
7.6
|
6.1
|
28.6
|
41.7
|
6.1
|
8.4
|
4.4
|
2.0
|
|
Germany
|
136.5
|
142.1
|
-
|
88.2
|
39.3
|
176.4
|
72.3
|
28.2
|
7.7
|
36.7
|
12.8
|
23.7
|
9.0
|
11.2
|
4.1
|
|
Ireland
|
336.0
|
15.6
|
12.4
|
8.5
|
4.1
|
1.4
|
5.3
|
2.4
|
-
|
2.7
|
1.7
|
1.8
|
0.9
|
0.5
|
0.6
|
|
Luxembourg
|
0.1
|
n.a.
|
16.3
|
6.3
|
4.3
|
0.8
|
0.5
|
0.0
|
0.0
|
0.2
|
-
|
0.1
|
0.1
|
0.0
|
0.0
|
|
Netherlands
|
81.7
|
46.5
|
134.5
|
37.9
|
151.7
|
8.5
|
-
|
9.9
|
3.0
|
8.1
|
4.0
|
7.5
|
5.9
|
2.7
|
1.3
|
|
Sweden
|
34.4
|
21.0
|
17.6
|
8.2
|
3.8
|
2.9
|
4.2
|
-
|
1.1
|
3.0
|
1.9
|
14.9
|
2.0
|
1.2
|
8.2
|
|
UK
|
-
|
296.4
|
97.0
|
148.8
|
23.5
|
10.0
|
44.2
|
19.8
|
121.2
|
26.6
|
6.1
|
16.7
|
17.2
|
16.0
|
4.4
|
|
All countries
|
1144.1
|
911.5
|
608.0
|
293.4*
|
279.3
|
121.5
|
89.7
|
40.4
|
35.4
|
28.1
|
19.3
|
16.4
|
11.1
|
6.5
|
6.4
|
Source: Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset, dataset code: migr_pop1ctz. Note: blue = main destination country, red = second main destination country. * The total for Spain as a destination country does not include Luxembourgish citizens.
Posted workers
In EU terminology, posted workers are EU citizens with an employment contract in their home country, who are temporarily posted to a host EU country by their employer when their employer provides a certain service. For example, if a Polish construction company builds a house in Germany it might post workers from Poland who have Polish contracts. Posted workers are different from labour migrants, who reside and are employed in the host country.
Posted workers receive a great deal of attention in EU policy discussions (Darvas and Vaccarino, 2016). The revision of the EU directive setting the rules for posted workers is at time of writing under negotiation between the European Parliament and the Council. In 2015, there were about 2 million work postings in the European Economic Area (EEA) and Switzerland, representing a small share of total employment (0.65 percent of the labour force and 0.9 percent of total employed people in the EU). These numbers are small compared to the 11.3 million long-term migrants. Furthermore, the average duration of a posting is only 98 days, so in full-time-equivalent terms, the share of posted workers in total employment is only 0.4 percent (Darvas, 2017b, and De Wispelaere and Pacolet, 2016).
Posting is concentrated in three countries, on both the sending and receiving ends. Poland (22.8 percent), Germany (11.7 percent) and France (6.9 percent) are the largest sending countries in absolute terms, while the main destination countries are Germany (28 percent), France (11.9 percent) and Belgium (10.5 percent). When analysing posted workers as a share of domestic employment, however, the picture is quite different. For instance, Poland sends out almost a quarter of all posted workers, yet these account for only 2.5 percent of domestic employment. On the other hand, Luxembourg (24.7 percent) and Slovenia (14.2 percent) have the highest shares of sent posted workers as a share of domestic employment. From the receiving countries’ perspective, Luxembourg (9 percent), Belgium (3.8 percent) and Austria (2.5 percent) have the highest shares of posted workers relative to domestic employment (European Commission, 2016d).
The dominant sector for posted workers is construction, which accounts for about 42 percent of total postings, but there are major differences between EU countries (Figure 21). Non-construction industry, finance, and education and social work also account for relatively large shares.
Figure 19: Posted workers by sending member state, 2010 and 2015
Source: De Wispelaere and Pacolet (2017) Note: for Croatia and Switzerland data is from 2013 instead of 2010, for Bulgaria data is from 2012 instead of 2010, and for Sweden data is from 2011 instead of 2010. The data for Norway for 2015 is from 2014.
Figure 20: Posted workers by receiving member state, 2010 and 2015
Source: De Wispelaere and Pacolet (2017) Note: Because of a lack of data, these numbers do not include posted workers active in two or more member states; as such the numbers given are an underestimate. For Croatia, 2013 data is substituted for 2010 data.
Figure 21: Sector of posted workers in receiving member states, 2015
Source: European Commission (2017) ‘Posting of workers - Report on A1 portable documents issued in 2015’, http://ec.europa.eu/social/main.jsp?catId=738&langId=en&pubId=7980&furtherPubs=yes.
De Wispelaere and Pacolet (2015) argue that there are several reasons to favour posting over permanent migration. For example, because posted workers pay tax in their countries of origin (if the posting does not exceed half a year), the posting mechanism can potentially serve as a stabilisation tool, somewhat serving as a form of fiscal transfer, when the country of origin experiences an asymmetric shock. However, the specific conditions under which posted workers are employed in host countries, their relatively small number and the limited durations of postings mean the labour market impact and/or local displacement effects are likely to be very small.
Human capital flight
Human capital flight, or brain drain, is the emigration of skilled individuals. The literature has extensively examined the brain-drain hypothesis and its detrimental effects on sending countries, mainly focusing on movements from developing countries to the OECD (Docquier, 2014). For the EU, two avenues can be analysed. The first is the EU as a whole and the movement of people from the EU to other economies, such as the United States or Canada (Choi and Veugelers, 2015). Second, movements within the EU can be examined, mainly the outflows of skilled migrants from central and eastern Europe and southern Europe to the northern and western EU countries.
While the academic literature focuses on highly skilled people, such as those who completed tertiary education, in the European context other skilled professions, such as construction and other industrial workers, are similarly important.
Brain drain can result from push and pull factors. Low wages and GDP per capita in home countries, combined with the prospect of better living standards in other countries, lead people to move. High rates of youth unemployment and the EU rights of free movement have also been highlighted as push factors (Andor, 2014).
On the economic impact of brain drain, as expected, declining productivity, the lack of innovation and increasing inequality are mentioned as consequences for income, welfare and employment in sending countries (Grubel and Scott, 1966; Bhagwati and Hamada, 1974), However, the effects depend on the “magnitude, the speed, the intensity and the structure of immigration flows with regard to age and qualification”, and on “the business cycle of the receiving economy” (Straubhaar, 2000).
Figure 22 shows that on average, about 10 percent of the EU’s high-skilled people are living abroad, but there are huge differences between countries. About half of Malta’s high-skilled people were abroad from the 1980s to the 2010s.
Figure 22: Share of high-skilled nationals living abroad relative to total high-skilled nationals, 1980-2010
Source: Brücker et al (2013) dataset. Note: for each skill level and year, emigration rates are computed as the proportion of the stock of migrants over the stock of pre-migration population (defined as the sum of residents and migrants in each source country). Age group: 25 years and older. The database is built merging national census and population data for 20 OECD destination countries concerning 195 origin countries. For details, see methodological note, available at: http://doku.iab.de/daten/brain-drain/iabbd_8010_v1_methodology.pdf.
The share of high skilled workers from in Austria, Cyprus and Ireland working abroad was similar to the Maltese share in the 1980s, but these shares have reduced more recently. In contrast, Bulgaria, Croatia, Estonia, Portugal and Romania are now sending significantly more of their high-skilled people to other countries.
Figure 23 reports shares of low, medium and high-skilled nationals living abroad for the three main EU country groups. Clearly, emigration of high-skilled individuals is highest from the group of central and eastern European countries (14.5 percent in 2010). The shares are lower for the group of north-west European countries (9.6 percent) and southern Europe (6 percent). It is notable that, for the central and eastern European and the north-west European country groups, much larger shares of high-skilled people have left compared to medium and low-skilled people, but for southern Europe the emigration rates of all three skill classes are rather similar.
Figure 24 shows that labour shortages in central and eastern European countries increased significantly after the 2004 EU enlargement, but there was also an increased shortage of labour in north-west EU countries in this period. These finding have two implications. First, emigration after EU enlargement had a negative impact on central and eastern European labour markets and created labour shortages.
Human capital flight can harm source countries by creating labour shortages (Docquier and Bhargava, 2007). Figure 24 shows the share of respondents to the European Commission’s Business Survey that mentioned labour shortages as a factor impeding production for the industry, services and construction sectors.
Figure 23: Share of low, medium and high-skilled nationals living abroad as a percentage of total number of nationals with the same skill level, 2005-10
Source: Bruegel based on Brücker et al (2013) dataset and Eurostat ‘Population by educational attainment level, sex and age’ dataset, code: [edat_lfs_9901]. Note: the figure shows data for people aged above 25. See the note to Figure 22 for further details.
Figure 24: Labour shortages as a factor impeding production
Source: European Commission, European Business and Consumer Surveys: Joint harmonised EU industry survey (Question 8 Answer F3S – ‘Factors limiting the production: labour’), Joint Harmonised EU Construction Survey (Question 2 Answer F4S – ‘Main factors currently limiting your building activity: shortage of labour force’) and Joint harmonised EU services survey (Question 7 Answer F3S – ‘Factors limiting the business: labour forces’). Note: CEE10 = 10 central and eastern European countries; NW11: 11 north-west EU countries; SE4: 4 southern European countries.
Second, the immigration of these central Europeans to north-west EU countries did not take jobs from local workers at any significant rate (at least in the construction, other industry and services sectors for which data is available), because labour shortages in those countries were on the rise in parallel to the arrival of central and eastern European workers.
Figure 25: Countries’ capacity to retain and attract talent, 2017
Source: World Economic Forum Global Competitiveness Index, 2017-2018 Edition. Indicator codes [EOSQ399] and [EOSQ400]. Note: the indicators are presented as a weighted average between 2016 and 2017. The indicators measure to what extent the country retains/attracts talented people domestically/from abroad, respectively. Both indicators are scaled from 1 to 7, with 1 indicating not at all and 7 = to a great extent. A low score in the country’s capacity to retain talent, therefore, indicates that the best and brightest might leave to pursue opportunities abroad; a low score in the country’s capacity to attract talent indicates that the country is not able to attract the best and brightest from around the world. For details on the computation methodology see: ‘Technical Notes and Sources’ from the Global Competitiveness Report 2017-2018.
With the recession and increased unemployment after 2008, labour shortages became a minor factor. But with the recovery after 2012, labour shortages began to appear again. More recently the severity of this problem even exceeded its pre-crisis peak, especially in central and eastern European countries, and also in north-west European countries. Therefore, the conclusions reached for the pre-crisis period continue to apply: emigration from central and eastern European countries had a major negative impact on their labour markets and created labour shortages, while the migration of these central Europeans to north-west EU countries did not take jobs from local workers.
It is also noticeable that in the four southern European countries, labour shortages have not posed a major problem, either before the crisis or more recently.
The ability of a country to attract and retain talent is an important factor that can influence human capital flight and brain drain. World Economic Forum indicators (2017-18) show that countries’ capacities to retain and attract talent are similar (Figure 25). For instance, the United Kingdom has a high capacity to both retain and attract talent, whereas Romania has a small capacity in both areas.
Figure 26: Change in countries’ capacities to retain and attract talent, 2014-17
Source: Bruegel based on World Economic Forum Global Competitiveness Index, 2017-2018 and 2014-2015 Editions. Note: the variation is computed by taking the difference between the scores attained in 2017-2018 and the scores attained in 2014-2015. See Figure 25 for infomation about the indicators.
From 2014-17, the central and eastern European countries that suffer the most from labour shortages experienced mixed fortunes in terms of their ability to retain and attract talent. Bulgaria, Poland, Estonia and the Czech Republic became more attractive to talent from 2014-17 (though their scores remain below the scores of the main north-west EU destination countries), while Hungary, Romania and Latvia became less attractive to talent.
More recently, some studies have posited that the effects of remittances and returning migrants could overcome the negative impacts of emigration – according to the brain-gain hypothesis, sending countries could also benefit from the exit of the brightest, either because it creates incentives for them to invest more in human capital (see Poutvaara, 2004; Mayr et al, 2009; Beine et al, 2003) or through positive diaspora externalities on foreign direct investment (Docquier and Lodigiani, 2006).
3.3 Immigration from outside the EU
Economic, social, political, historical and cultural factors explain the large numbers of extra-EU migrants in some member states. Migali and Natale (2017) find that the existence of a previous colonial link, a common legal system and a common official language are all significant determinants of total first residence permits, regardless of the form of migration. Commonwealth citizens who are either married to, or are the children of, someone with the right of abode might also have the same right to reside in the UK for an unlimited time. Commonwealth citizens are also eligible for a UK ancestry visa, which allows them to stay in the UK for five years. In some cases, bilateral agreements can influence migration flows. Germany and Turkey, for instance, have a strong migration link that dates back to a bilateral labour agreement signed in 1961. Following an initial influx and settlement in Germany of contracted workers from Turkey, immigration flows during the 1970s were linked to family reunification. In the 1980s migration from Turkey to Germany had political motivations, with migrants seeking asylum (Aydin, 2016).
Figure 27: EU country residence permits by main sending countries and reason for issuance, 2008 and 2016, thousands
Source: Eurostat ‘First permits by reason, length of validity and citizenship’ dataset, code: migr_resfirst. Note: country names indicated on the horizontal axis are source countries from where immigrants arrived in the EU. China includes Hong Kong. Other reasons include refugee status, humanitarian reasons, residence only permits without the intention to work, study or reunite with family (eg residence permits issued to pensioners).
Figure 28: First residence permits issued by EU countries, type and duration of the permit (share of total permits for a particular duration), 2008 and 2016
Source: Eurostat ‘First permits by reason, length of validity and citizenship’ dataset, code: migr_resfirst. Note: the category ‘Others’ includes permits issued to refugees and beneficiaries of various other protection schemes (subsidiary protection, international protection, humanitarian reasons, unaccompanied minors, victims of trafficking in human beings), but also for unspecified reasons and residency-only permits. Eurostat’s metadata file notes that there have been improvements in statistics collected, but problems of missing and incomplete data occur: http://ec.europa.eu/eurostat/cache/metadata/en/migr_res_esms.htm.
Migrants’ motives
The type of residence permit issued varies with the length of the stay (Figure 28). Immigrants who come for less than a year arrive predominantly for work: 63 percent of short-term residence permits in 2008 and 44 percent in 2016 were issued for this reason. In contrast, in 2008, 24 percent of migrants staying more than one year received a work permit, declining to 13 percent in 2016. People migrating for education accounted for about a quarter of both short and long-term residence permits. This share did not change much between 2008 and 2016. Family reasons are less important for short-term migrants (10 percent of residence permits issued), but about one-third of longer-term residence permits are issued to migrants for this reason. Residence permits issued in the category ‘others’, which includes refugees, beneficiaries of other protection schemes and residence permits issued for unspecified reasons (which also could include granting of refugees status), saw a significant increase both in terms of short-term and longer-term permits. In terms of absolute numbers, 345,000 short-term and 684,000 long-term permits issued to immigrants for ‘other’ reasons in 2016, altogether slightly over than one million people.
High-skilled migration into the EU
Between 2010 and 2011, 28 percent of migrants into the EU were tertiary-educated. High-skilled migrants (HSM) can be broadly defined as being tertiary educated, but the OECD provides three definitions (Chaloff et al, 2009) of HSM based on education, occupation or wage level: ‘highskilled’ is considered to “include post-secondary education that is university-level but that may involve a vocational, technical or professional qualification of shorter duration than a bachelor’s degree”, corresponding approximately to tertiary education. From an occupational standpoint, “occupations including managers, professionals and associate professionals” are included in the definition. The wage-level is used in some countries as a proxy for skill level.
High-skilled migration offers added value to recipient countries. Benefits range from increasing the entrepreneurial profile of the country to filling the skills gap in host countries. Desiderio and Mestres-Domènech (2011) note that “the proportion of new migrant entrepreneurs in the labour force is much higher than among natives”, a significant share of migrant entrepreneurs (30-40 percent) have tertiary education and a “higher average educational level than their native counterparts”. An influx of high-skilled migrants is also useful in filling specialist shortages. HSM “complement physical capital and technology and the human capital of both low- and high-skilled native workers” (Constant, 2014) and can be an incentive for technological progress (Kerr and Lincoln, 2011).
In the EU, the main common mechanism to attract HSM is the European Blue Card, which is regulated by Council Directive 2009/50/EC of 25 May 2009. This directive defines highly qualified employment as the “employment of a person who in the Member State concerned, is protected as an employee under national employment law and/or in accordance with national practice, irrespective of the legal relationship, for the purpose of exercising genuine and effective work”, “is paid” and “has the required adequate and specific competence, as proven by higher professional qualifications”.
However, there is a general consensus that the EU needs to improve its policies toward high-skilled migrants (European Commission, 2016a). The Blue Card has not appealed enough to either employers or high-skilled workers themselves and – with the exception of Germany – has been severely underused by EU countries. This means that the EU issues fewer permits to highly-skilled workers than other competitors for HSM. In 2015, 17,104 visas were issued in Europe to high-skilled workers, compared to the US total of 275,317.
Discussions are underway on reform to improve the Blue Card system and address some of its major shortcomings (see González et al, 2013, and Kahanec and Zimmermann, 2011). To begin with, the EU Blue Card sets a rather high salary requirement threshold. The thresholds vary in different EU countries but are still higher than HSM scheme thresholds in other countries. It is therefore difficult for a HSM to meet the minimum salary requirement in order to apply for the card in the first place; only one in three highly-educated migrants meet the salary requirements. The system also greatly restricts mobility: in-country status changes are not allowed and Blue Card holders cannot move from one member state to another without applying for a new EU Blue Card. Intra-EU mobility under the current directive is considered relatively unattractive for migrants. More importantly, many employers are reluctant to request Blue Cards for potential employees, seeing the process as too problematic. Offering incentives, such as the possibility for cardholders to apply for a permanence residence permit, would improve the attractiveness of the Blue Card scheme.
In the face of an increasing skills shortage, it is crucial for the EU to begin to take more of a lead in the global race for talent. According to a Gallup survey (cited in European Commission 2016a), 33 percent of highly-educated workers worldwide who intended to migrate preferred the EU (or European Economic Area) as a destination, compared to 19 percent who preferred the United States. However, in the United States, 37 percent of working age migrants have at least a college degree against 26 percent in the EU. This could suggest that while the EU might seem attractive to a potential migrant, its system does not facilitate high-skilled migration in the same way that those of other OECD countries do.
Inflows of asylum seekers into the EU
Both in 2015 and 2016, about 1.2 million to 1.3 million first-time asylum applications were registered in the EU with changing within-year intensity (Figure 29). In all EU countries, asylum applications from asylum seekers peaked in the autumn of 2015. In addition to the notable case of Germany, high numbers of applications were filed in Hungary, Sweden and, to a lesser extent, Austria and Italy. From the end of 2015 to the autumn of 2016, Germany received the overwhelming share of applications; thereafter, the number of applications abruptly fell. Since the beginning of 2017, the asylum application wave has abated considerably, with the notable exception of Italy, where the number continues to rise at time of writing.
Figure 29: First-time asylum applications in the EU, by country of application, January 2008 – September 2017
Source: Eurostat ‘Asylum and first time asylum applicants by citizenship, age and sex Annual aggregated data (rounded)’ dataset, code: migr_asyappctza. Note: for some countries, data on first-time asylum applications is not available for certain periods at the beginning or our sample period, but data on total applications is available. We approximate the missing first-time applications data by assuming that the ratio of first time to total applications was the same in the missing periods as the actual value of this ratio in the first full year when both indicators are available.
The large inflow of asylum seekers posed immense difficulties in dealing with their applications, which is also reflected in the number of pending applications. The number of applicants waiting for a decision increased from approximately 300,000 in 2010-13 to 1.2 million by September 2016 (Figure 30). Since then, along with the decline in applications (Figure 29), the number of persons with pending applications has declined. Nevertheless, as of September 2017, there were almost one million asylum seekers awaiting decisions, a rather large number.
Large numbers of asylum applications do not translate into equally large acceptance rates. In the EU as a whole, the acceptance rate was about 50 percent, but there are very significant differences between member states (Figure 31). In Hungary, Poland and Croatia, the acceptance rate is only about 20 percent or even lower, while Malta accepted about 80 percent of asylum applications and Slovakia more than 60 percent.
Figure 30: Asylum seekers with applications pending at the end of the month, January 2008 – September 2017
Source: Eurostat ‘Persons subject of asylum applications pending at the end of the month by citizenship, age and sex Monthly data (rounded) [migr_asypenctzm]’ dataset. Note: Data for some countries was missing for 2014 (Austria and Netherlands), 2013 (Netherlands and Cyprus), 2011-12 (Cyprus and Croatia), and some months in 2010 for Belgium, the United Kingdom, Cyprus and Croatia. Whenever it was possible, we approximated these missing values by interpolating available data, eg for Austria we assumed a constant growth rate between December 2013 and January 2015. For the UK, we used the percent growth rate observed in Sweden to approximate the missing UK values in January-May 2010.
There have been large swings in the acceptance rate through the years. For example, in Latvia it increased from 9 percent in 2015 to 78 percent in 2017, while in Bulgaria it declined from 91 percent in 2015 to 31 percent in 2017. In contrast, acceptance rates remained fairly stable in, for example, France, Italy and the UK.
The composition of flows of asylum seekers (eg whether they have risked their lives or migrated for economic reasons but used the current inflow of refugees as an opportunity to enter the EU) might explain some of the differences in acceptance rates, but is unlikely to be the only factor.
Figure 31: Positive first instance decisions on asylum applications, % of applications, 2015-17
Source: Eurostat’s ‘First instance decisions on applications by citizenship, age and sex Quarterly data (rounded) [migr_asydcfstq]’ dataset. Note: the first half 2017 is reported for 2017.
Leerkes (2015) analyses the composition of flows of asylum seekers by country of citizenship, age and gender and finds that some differences in acceptance rates are explained by these compositional factors, but nevertheless substantial differences persist. Drawing on the literature, he concludes that the differences in composition-adjusted acceptance rates are likely explained by differences in the willingness to admit asylum seekers, which is in turn can be related to differences in unemployment or differences in the popularity and political influence of anti-immigration parties. Therefore, it seems that variable implementation of the EU’s asylum rules also explains the differences between countries’ acceptance rates.
There is a huge difference within the EU in terms of where asylum applications are made, where the asylum applications are accepted and GDP (Figure 32). Disproportionate distribution of asylum seekers and refugees within the EU relative to the distribution of GDP entails disproportionate burdens and socio-economic implications for EU member states.
- Germany received disproportionally more applications and had a higher positive decision rate relative to its share of EU GDP, in all three years of 2015, 2016 and 2017.
- Sweden’s shares of applications and of positive decisions were also disproportionally high in 2015, but by 2017 its asylum seeker and refugee shares became more similar to its GDP share.
- Greece has received a disproportionately high share of applications, but because it ultimately accepts fewer asylum seekers than the EU average, its share of positive decisions remained close to its GDP share.
- Hungary received 14 percent of applications in 2015, a huge share compared to Hungary’s 1 percent share of EU GDP. But because of a very low acceptance rate (Figure 31), Hungary’s share of positive decisions was a mere 0.2 percent. After building a fence along its southern border, Hungary’s share of applications declined to very close to zero.
- Italy’s share of applications was lower than its share of EU GDP in 2015, but this relationship had reversed by 2017. However, because of a lower than the EU average acceptance rate, Italy’s share of positive decision remains well below its share of EU GDP.
- For many other countries, including Poland, the Netherlands, Spain, France and the United Kingdom, the shares of asylum applications and positive decisions remained well below their GDP shares.
The EU has financial instruments dedicated to supporting the management of migration flows and safeguarding the Schengen area, including the Asylum Migration and Integration Fund for 2014-20, with a total budget of €3.137 billion, and the Internal Security Fund, for which €3.8 billion was allocated in the 2014-20 multi-annual financial framework. We discuss these instruments, along with the various other initiatives to manage the refugee crisis, in chapter 7.
Figure 32: Total asylum applications, positive first instance decisions, and GDP, % of EU totals, 2015, 2016 and 2017
Source: Eurostat’s ‘Asylum and first time asylum applicants by citizenship, age and sex Monthly data (rounded) [migr_asyappctzm]’ and ‘First instance decisions on applications by citizenship, age and sex Quarterly data (rounded) [migr_asydcfstq]’ datasets, and the November 2017 AMECO dataset. Note: the 2017 distributions were calculated using the following data: January-August 2017 for first time applications, January-June 2017 for positive decisions, and the whole of 2017 for GDP.
Figure 33: Asylum applicants by age and gender, 2014-16
Source: Eurostat ‘Asylum and first time asylum applicants by citizenship, age and sex’ dataset, code: [migr_asyappctza].
3.4 The share of foreign citizens and foreign-born people resident in the population
While defining an immigrant is relatively straightforward in relation to flows of people, defining immigrants within the resident population is much more complicated (see Box 1 in chapter 1). There are two standard indicators of stocks of immigrants in the resident population of a host country: the share of foreign citizens and the share of foreign-born people. Foreign citizens are people who retain the citizenship of their home countries. This category might include people born in the host country. Foreign-born persons were born outside of the country of their usual residence, regardless of their citizenship. This category might include home-country citizens who were born to home-country parents abroad. There is some but imperfect overlap between the two categories, because some foreign-born people obtain the citizenship of their host country, while some foreign citizens were born their host country.
While these two alternative indicators of immigrants in resident population are straightforward, they might not coincide with the perceptions of the native population. For example, native people might still consider a naturalised former immigrant as an ‘immigrant’, even if this person has obtained the citizenship of the host country.
Foreign citizens
We calculate the share of foreign citizens in each EU country at the end of 2015. We make two adjustments to population statistics. First, at the end of 2015 there were 1 million asylum seekers in the EU who had submitted their asylum applications and were waiting for decisions. These people were in EU member states, but were not included in population statistics. Second, we also consider posted workers. Most posted workers are not included in the resident population of the host country, when the duration of the posting is short.
Therefore, we calculate an adjusted population for each EU country at the end of 2015. We start with the reported population number, add the number of asylum seekers waiting for decisions on their applications, add the posted workers received from other EU countries and subtract the posted workers sent to other EU countries. Since no data is available on the number of posted workers at the end of 2015, we had to approximate their number (Box 2).
Box 2: Calculation of the number of posted workers at the end of 2015
The calculation of posted workers present in each EU country at the end of 2015 necessitates some approximations, for three reasons.
First, data is not available on the number of posted workers at the end of 2015, only on the total number of people involved in posting in 2015. Most of them, for example those people who go for a posting period for one month between January and November, are not present in the host member state at the end of the year. We therefore calculate the average number of posted workers for each day in 2015, which is the same indicator as the ‘full time equivalent job indicator’. We calculate this indicator from the perspectives of both sending and receiving countries. To do that, we need information on the average duration of postings. This information is available for 13 sending countries. For the other 15 EU countries, we used approximations by considering neighbouring countries or countries with similar characteristics, as indicated in Table 5. For example, the average duration of postings by Belgian companies is 33 days, while it is 29 days for French companies. We used the average of these two numbers, 31 days, for neighbouring countries and other western European countries that might have similar characteristics, such as Austria, Denmark, Finland, Germany, the Netherlands and Sweden. Table 5 shows that the duration of postings is typically much longer for workers posted from central and eastern European countries than from western European countries. We assume that the duration of postings from a particular source country to all destinations countries is the same, given that duration data is not available on a bilateral basis. For example, we assume that Belgian posted workers spend on average 33 days in Germany, the same 33 days in the Netherlands, and the same 33 days in all other EU countries.
Second, while all people holding the A1 document are considered ‘posted workers’ in policy discussions (their total number was 2.05 million for the European Economic Area (EEA) plus Switzerland in 2015), in fact slightly less than 1.5 million people are actually posted workers, while the remaining half a million are either ‘active in two or more member states’ or ‘others’, such as civil servants, sailors or flight crew. Still, we consider people belonging to all three categories as ‘posted workers’, similarly to the use of this term in policy dialogues. The detailed bilateral matrix for source and destination countries is available only for the 1.5 million actual posted workers: for all others, we assumed that the composition of destination countries (from a given source country) is the same as the composition of actual posted workers. Furthermore, we only consider EU members, but not non-EU EEA members or Switzerland, and thus we do not consider people who are posted from or to Norway, Iceland, Lichtenstein and Switzerland.
Third, the detailed decomposition of destination countries is available for all EU member states except the United Kingdom. For the UK, we assumed that the share of Ireland in destination countries is 15 percent (this is an ad-hoc assumption, based on an overview of the shares of posted workers of other EU countries in their neighbouring countries), while the share of all other EU countries is proportional to the average share of each country in total EU postings. Given that the UK posts relatively few workers, less than 32,000, who account for about 12,000 full-time equivalent jobs, this approximation hardly distorts our final results.
Table 5: Average duration of postings of workers by sending countries, actual data (white box) and our approximations (gray box), 2015 (days)
|
|
|
|
Austria
|
31
|
Belgium, France
|
|
Belgium
|
33
|
|
|
Bulgaria
|
188
|
Croatia, Hungary, Slovenia
|
|
Croatia
|
225
|
|
|
Cyprus
|
298
|
|
|
Czech Rep.
|
148
|
|
|
Denmark
|
31
|
Belgium, France
|
|
Estonia
|
227
|
|
|
Finland
|
31
|
Belgium, France
|
|
France
|
29
|
|
|
Germany
|
31
|
Belgium, France
|
|
Greece
|
195
|
Cyprus, Italy
|
|
Hungary
|
198
|
|
|
Ireland
|
240
|
|
|
Italy
|
92
|
|
|
Latvia
|
266
|
|
|
Lithuania
|
266
|
Latvia
|
|
Luxembourg
|
16
|
|
|
Malta
|
61
|
France, Italy
|
|
Netherlands
|
31
|
Belgium, France
|
|
Poland
|
173
|
Czech Republic, Hungary
|
|
Portugal
|
61
|
France, Italy
|
|
Romania
|
188
|
Croatia, Hungary, Slovenia
|
|
Slovenia
|
142
|
|
|
Slovakia
|
173
|
Czech Republic, Hungary
|
|
Spain
|
61
|
France, Italy
|
|
Sweden
|
31
|
Belgium, France
|
|
UK
|
135
|
France, Ireland
|
Source: Bruegel based on De Wispelaere and Pacolet (2017).
Our results (Table 6) show that intra-EU mobility has not reached very high levels as a percentage of host country population, with the exception of Luxembourg. For example, central and eastern European citizens account for only 1.88 percent of the population of north-west EU countries, while posted workers add a mere additional 0.15 percent. Six southern EU members account for 1.18 percent of the population of north-west EU countries, while citizens of other north-west countries account for an additional 1.13 percent (eg Germans who live in Belgium and Belgians who live in Germany). Altogether, other EU citizens account for 4.4 percent of the population of north-west EU countries. Central and eastern European countries host very few other EU citizens, while in the South 6 countries, the share of central and eastern European citizens is almost 2 percent of population.
The share of central and eastern European citizens in the resident population of host countries is highest in Austria and Ireland (over 4 percent). Their share at about 2.5 percent in Germany and the United Kingdom is also above average, while their share is especially low in France at 0.3 percent. Luxembourg is really special: 41 percent of the population is citizen of another EU country, mostly from north-west and southern European countries. With 13 percent of population from other EU countries, Cyprus hosts the second largest share of other EU citizens, while Belgium and Ireland are in third place with shares slightly higher than 8 percent.
The shares of foreign citizens in the population are rather low, below 2 percent, in a number of central European countries: Bulgaria, Croatia, Hungary, Lithuania, Poland, Romania and Slovakia. These low shares are not surprising because wage levels are much lower in these countries than in western European countries and wages are key drivers of migration decisions. In two other central European countries, Estonia and Latvia, the rather high share of about 15 percent is explained by historical reasons related to their former inclusion in the Soviet Union – Russian nationals continue to live in these countries (see chapter 2). However, the share of other-EU nationals is extremely low even in Estonia and Latvia, highlighting again the major role of wages in migration flows.
Table 6 also shows that posted workers account for an especially minor share of population. For example, the shares of posted workers from central and eastern European countries in the populations of France, the Netherlands and Belgium (three countries that have been especially vocal in the posted workers debate) are 0.05 percent, 0.13 percent and 0.32 percent, respectively – very low shares. Therefore, it is surprising that the revision of the posted workers directive has been so prominent in EU policy debates (see chapter 7).
The share of non-EU citizens in resident populations varies from half percent in Lithuania to 14 percent in Estonia and Latvia. The EU average is slightly above 4 percent. Pending asylum applicants add a further 0.2 percent, which, as we note, is distributed rather unevenly between EU countries, with Sweden and Austria having the largest shares relative to population.
Table 6: Composition of EU country populations according to citizenship, end-2015 (percent of adjusted population)
Source: Bruegel. Note: North-West 11: first 15 EU members without Italy, Greece, Portugal and Spain. South 6: Cyprus, Italy, Malta, Greece, Portugal and Spain. CEE 11: the central and eastern European members that joined the EU in 2004-13. Adjusted population: population as reported in population statistics plus pending asylum seekers plus posted workers received minus posted workers sent.
Foreign-born people
Since some immigrants obtain the citizenship of their host country, the number of foreign citizens within a population might not correspond with all people who are viewed as ‘immigrants’. Foreign-born people might not correspond either to people who are viewed as ‘immigrants’, because the category of foreign-born people does not include people whose parents or grandparents were immigrants but who were born in the host country – quite often, such people are also considered ‘immigrants’. Moreover, the foreign-born population includes those people who were born to native parents abroad but moved to the home country of their parents.
EU countries generally have higher shares of foreign-born residents than foreign citizens, with the exceptions of Luxembourg, Estonia and Greece (Figure 34). Switzerland, Australia, New Zealand and Canada host much larger shares of foreign-born people than EU countries with the exception of Luxembourg. Nevertheless in almost half the EU countries (for which OECD data is available), the share of foreign-born residents is similar or even higher than in the United States. However, the way intra-EU mobility is taken into account has a major bearing on the comparison with the US. For the US, we do not consider intra-US mobility, but only mobility from other countries. But for the EU, foreign-born people and foreign citizens also include people from other EU countries. If we view the EU as a single entity, the share of non-EU people in the EU is much lower than non-US people in the United States. But if people from other EU countries are considered to belong to the same category as non-EU immigrants, then in about half of the EU countries, the share of ‘immigrants’ has reached the share observed in the United States.
Neither the share of foreign-born, nor the share of foreign citizens, is high in the United Kingdom relative to other EU countries, and the UK shares are well below the shares in Canada, Australia and New Zealand. Therefore, it is notable that even though an international comparison does not suggest unusually high share of foreigners in the UK, immigration was a key topic in the campaign ahead of the UK’s referendum on EU membership in June 2016.
Figure 34: Foreign-born and foreign citizen population in 2013 (percent of total population)
Source: OECD International migration database. Note: 2012 for Ireland, France and Greece; 2011 for Poland and Canada. Data on foreign citizens is not available for Australia and New Zealand, while data on foreign born residents is not available for Japan.
Source: Bruegel based on Eurostat ‘Population on 1 January by age group, sex and citizenship’ dataset , code: migr_pop1ctz and OECD International Migration Database. For the foreign-born panel, support was computed as the weighted average of support for intra-EU and extra-EU immigration, using weights proportional to the share of other-EU and non-EU citizens in population.
