Migration has been at the centre of political debate across the OECD in recent years. Drawing on data from the OECD Survey of Adult Skills (PIAAC), this report provides new evidence on differences in migrants’ characteristics and contexts and considers how these relate to the skills migrants possess. It also examines the relationship between migrants’ skills and their labour and non-labour market outcomes in host countries. Finally, it sheds new light on how migrants’ skills are developed, used and valued in host country labour markets and societies. Results and lessons gleaned from analysis highlight the way forward for future research on this topic.
The report represents an invaluable resource for policy makers across different sectors as they design and implement strategies aimed at promoting the long-term integration of foreign-born populations in the economic and social life of their countries. The analyses presented allow us to identify the skill composition of foreign-born populations, the labour market and broader social outcomes associated with such skills, and the factors that can promote skill acquisition and skill use.
Skills on the Move - Migrants in the Survey of Adult Skills
1. SKILLS ON THE MOVE
Andreas Schleicher – OECD Director for Education and Skills
10 November 2018
Migrants in the Survey of Adult Skills
2.
3. The public debate on immigration is sometimes based on a
misleading view of migrants as a homogenous group of low
skilled individuals, with little chance of integrating
Typically, skills are proxied by the level of education attained
imperfect proxy, especially for migrants
New information from the OECD Survey of Adult Skills allows
an in-depth analysis of the skills of migrants, their use and
development in the host countries.
This information is particularly valuable in the policy context:
effective integration policies must build upon migrants’ existing
skills and experiences in order to help them recognise, develop, and
use their skills in a tailored and individualised fashion
PIAAC: a tool to study migrants’ skills
4. The share of migrants varies widely across
countries
0
5
10
15
20
25
30
35
Foreign-born
%
Proportion of foreign-born among the 16-65 years old population
The share of migrants in the population is very small (less than
3%) in Indonesia, Japan, Korea, Poland, the Slovak Republic and
Turkey.
5. On average, migrants have lower skills
proficiency than natives
Migrants are less proficient in literacy, numeracy and problem solving than
native-born adults in all participating countries, except for Chile
Wide cross-country differences are driven by differences in the
composition of the migrant populations between countries.
Literacy proficiency, by place of birth
200
220
240
260
280
300
320
Foreign-born Native-born
score points
6. Migrants are more likely to have very
low levels of literacy skills…
0
10
20
30
40
50
60
70
80
Foreign-born Native-born
%
Adults with very low literacy proficiency (Level 1 or below), by place of birth
7. … but many migrants have high levels of skills, although
they are less likely than natives to have high skills
Shares of persons reaching levels 4 and 5 in literacy proficiency, by place of birth
0
5
10
15
20
25
Foreign-born Native-born
%
8. Migrants are a very heterogeneous
group
Migrant’ skills proficiency vary greatly both within and across countries: they depend strongly on
the place where they acquired their qualifications, the age of arrival in the host country, the
duration of stay and the languages they speak
Literacy proficiency of migrants, by personal characteristics
Need for countries to adopt a tailor-made
approach in terms of skills, training and
integration programmes.
180
200
220
240
260
280
300
All foreign-born Native-born average Foreign-born who are native speakers in the host-country language
score points
9. Migrant educational levels play a crucial
role in shaping skills
Levels of literacy proficiency, by place of birth and education level
Whatever the education levels, migrants have lower proficiency than native-borns
Low educated migrants are particulary at risk of having very low proficiency.
10. Origin of migrants: example of EU migrants
0
10
20
30
40
50
60
70
80
Share of EU migrants (%)
%
Migrants from EU countries have similar skills than native-borns
Within the EU, countries face very different situations
11. Migrant age at arrival matters
0
5
10
15
20
25
30
35
40
45
50
Share of immigrants who arrived before age 6 (%)
%
The share of migrants who arrived before 6 is actually related to whether migrants
come as a family or not.
All countries where this share is higher than 20% are countries in which there are
only few migrants
12. • Gaps in literacy skills between migrants and
natives are on average 25 points…
… but are down to only 13 points if migrants are native
speakers!
• Language proximity between the migrant’s mother
tongue and the host country language also matters
• Language penalty is related to the age at arrival,
but not to the length of stay in the host country
Language matters!
13. Migrants and host country langage
0
10
20
30
40
50
60
70
80
90
100
Share of immigrants who speak the host country language (%)
%
14. Literacy gaps between natives and
migrants
-10
0
10
20
30
40
50
60
Migrant gap Migrant gap adjusted for age, education and gender
Migrant gap adjusted for age, education, gender and languageScore-point difference
15. Language distance in PIAAC participating
countries
0
20
40
60
80
100
120
Language distance
Average language distance Interquantile range (75th minus 25th percentile)
16. • Adult training particularly important for
migrants
• Differences in participation rates within-
countries are small. What matters are
between-country differences
• Migrants do express higher demand for
training, and face higher barriers => higher
unmet demand
Migrants and lifelong learning
17. Rates of participation in lifelong learning
0
10
20
30
40
50
60
70
80
90
100
Participation rate (in %)
Foreign-born Native-born
Light grey bars/diamonds indicate countries for which the unadjusted difference in participation rates
between natives and migrants is not statistically significant
19. Labour market outcomes
of migrants tend to lag
behind those of the native-
born
Migrants earn significantly less than
natives
Median hourly earnings
Migrants are more often unemployed
or inactive, and those who are in
employment tend to have lower
returns to education – in terms of
earnings – than their native-born peers
0
5
10
15
20
25
USD Natives Migrants educated at destination Migrants with foreign qualification
20. Migrants and the use of reading skills at
work
Differences between foreign- and native-born adults in the use of reading skills at work
-100
-80
-60
-40
-20
0
20
40
60
80
100 Further control for occupation Unadjusted gap Adjusted gap
differencesin%ofa
sttanarddeviation
21. Migrants and low skilled occupation: the
importance of education location
Share of workers in low skilled occupations
0
5
10
15
20
25
30
35
40
%
Migrants educated at destination Migrants with foreign qualification Natives
22. Reported health by immigrant status
40
45
50
55
60
65
70
75
80
85
90
95
%
Natives Migrants
24. A large part of the difference in the returns to education reflects
different patterns in occupational placement:
– Migrants are concentrated in jobs that are associated with a lower socio-
economic status in comparison with natives.
Yet, migrants are often paid less than the native-born even when
operating in similar roles
– This is largely (but not entirely) explained by differences in the skills held and
language spoken by migrants and natives, as well as by the country where their
education was acquired.
Skills play an important role in explaining labour
market outcomes gaps between migrants and natives
These results stress the key role played
by language proficiency and skills as
drivers of the labour market performance
of migrants
Notes de l'éditeur
This is the result of joint work between the Directorate for Education and Skills and the Directorate for Employment, Labour and Social Affairs.
In recent years, migration surged at the forefront of the policy debate across all OECD countries, mostly in response of the refugee crisis.
Although governments have put in place several policies to facilitate migrants’ integration, these have received mixed reactions from the civil society. One of the reasons is that the public debate is sometimes based on the perception that migrants are a homogenous group of low skills individuals, with little chance of integrating.
Yet investing in the integration of migrants is of primary importance. Migrants lag behind natives in terms of labour market outcomes and although they do converge, this process may take a long time. Effective integration policies should be in place to ensure migrants can enter the labour market and are able to develop and fully use their skills.
In the past, skills have been typically proxied by educational level, but this is particularly problematic in the case of migrants, since they come from a different schooling system where skills may be taught differently.
Thanks to PIAAC, the Survey of Adult Skills, it is now possible to address many policy-relevant questions related to skills.
It should be noted however, that the skills in PIAAC are assessed in a given language, hence the proficiency in the test language affects the skills assessment. Indeed, more than half of migrants in PIAAC are not native-speakers of the text language.
The results of this report are particularly valuable in the policy context: in fact, effective integration policies must build upon migrants’ existing skills and experiences in order to help them recognise, develop, and use their skills in a tailored and individualised fashion. The results found in this report will hence help policy makers to better design and implement strategies that can promote the long-term integration of migrants.
Migrants can account for up to 30% of the adult population in PIAAC countries
This share is reached in countries where migration play a foundamental part in their origins: New Zealand, Canada, Australia
In other countries, the migrant share is negligeable (Korea, Japan)
European countries face less extreme but still important difference: in Sweden and Lithuania, the share of migrants reach 20% but only a few percent in most east european countries
On average, migrants are less proficient in literacy, numeracy and problem solving than native-born adults in all participating countries, except for Chile.
But the skill gaps between migrants and natives vary greatly across countries.
The differences across countries are driven by differences in the composition of the migrant populations which are related to migration policies, historical ties and geopolitical factors.
In half of the countries, the skill gap between migrants and natives becomes zero or statistically insignificant when language and foreign qualifications are taken into account in addition to demographic characteristics and educational attainment.
In most countries, part of the remaining negative effect can be related to cultural differences, as migrants from different countries might interpret the assessment questions differently.
In most countries, migrants are a more heterogeneous group than natives.
In addition to the level of education, several key characteristics are associated with migrants’ skills. Migrants who had acquired their highest qualification in the host country have for example a level of literacy proficiency of 24 score points higher than migrants who had earned their highest qualification elsewhere. Migrants who arrived before the age of 6 also have a level of literacy proficiency of 27 score points higher than migrants who arrived after 6 years old. There are also large differences in literacy proficiency related to migrants’ region of birth, with migrants from EU countries having higher levels than other migrants. The duration of stay and the languages they speak also play key roles on their literacy skills.
We can see here how differences in migration experience partly explains the heterogeneity within the migrant population. It also demonstrates the need for countries to develop a tailor-made approach in terms of skills, training and integration programmes to address the very different needs of the migrant population.
(Migrants with very low education level and poor literacy and numeracy skills will need intensive support as early as possible to facilitate their access to the labour market and value their skills. On the contrary, highly educated migrants would benefit from advanced language and job-specific skills required for high-skilled jobs, and a recognition of their qualifications and skills.)
-Education is strongly associated with proficiency and this remains true for migrants.
-Whatever the education levels, migrants have lower proficiency than native-borns
-This could be partly driven by the place where migrants were educated, as shown in the previous slide
-Importantly, low educated migrants are particulary at risk of having very low proficiency: 25% do not reach level 1 in literacy.
In each country, migrant gaps are shaped by the composition of the migrant population. The country of origin of migrants play an important role here.
migrants from EU countries have similar skills than native-borns.
But even among EU countries, it varies between almost 70% in Ireland (where most of migrants are English born), and as low as 15% for Netherlands.
Migrants who arrive as young child in the country of destination have skills comparable to native born.
The share of migrants who arrived before 6 is actually related to whether migrants come as a family or not.
In all countries that matter, safe Israel, this share is quite steady between 7% and 20%
Countries where this share is high are not countries where migrants are numerous
The report provides unique evidence on the importance of language in shaping differences between migrants and natives in literacy and numeracy. So far most analyses considered migrants whose mother tongue was different from the language used in the PIAAC assessment as a “homogeneous group”: non-native languages speakers. But this did not recognise, for example, that for a Romanian speaker learning Italian is easier than it is for an Albanian speaker, because the Romanian language is considerably more similar to Italian (because of common Latin roots) than Albanian (a Slavic language).
Analyses developed in the report identify the unique role played by the dissimilarity in the native language spoken by individual migrants and the language in which they sat the PIAAC test in explaining both differences across countries in the language composition of migrant groups and differences within countries in migrants’ proficiency.
Analyses reveal that a large portion of the gap observed between migrants and natives in a country like Sweden is due to the fact that most migrants who settle in Sweden have native languages that are very different from Swedish.
This slide shows the observed difference in literacy scores between native born and foreign-born individuals (blue bar) as well as the difference when migrants and natives of similar ages education and gender are considered (the diamond). Results show that the gap in most cases remains either similar, or is even larger between the two groups. However the gap is considerably smaller when comparing the gap between migrants and natives who are both native language speakers of the language of the test (or, conversely who are both non-native speakers). The black dot shows that in the vast majority of countries the migrant gap is considerably smaller when controlling for the mother tongue language of individuals. In some countries the gap becomes very small or is not statistically significant (for example Greece and Finland). In others the gap remains seizable, albeit smaller, such as Germany, Canada and Sweden.
The bars show the average language distance between the languages spoken by migrants and the languages in which they sat the PIAAC test in each country. The diamond illustrates for each country the variability (measured using the interquartile range) in the language distance spoken by migrant populations in each countries.
These results show how migrants in Sweden, for example, generally have a mother tongue that is very different from the language in which they say the test (Swedish) (the blue bar is very big in comparative terms) but also that linguistic barriers affect all migrants equally (low level of interquartile range).
By contrast, in a country like Spain, linguistic barriers tend to be low, by international standards. However, some migrants speak languages that are very similar to the language in which they sat the test, while others speak languages that are very distant and therefore differences in proficiency among migrants in Spain tend to be more accentuated because of differences in mother tongue languages.
Considering the additional barrier determined by the specific languages migrants speak is crucial in order to determine the training needs (both in terms of intensity and length that individuals may require in order to become proficient and be able to express their skills and potential in the labour market and their communities).
Lifelong learning and adult training are becoming increasingly important in a rapidly changing world. Training arguably even more important for migrants, who need to bring their skills in line with the requests of the host country’s labour market.
Still, data show that the rates of participation of migrants in adult training are LOWER than those of natives. The differences are however rathers small, especially after account is taken for differences in observable characteristics (eg. more skilled and educated adults tend to participate more in training)
But there is evidence that migrants, although they participate LESS, express a higher DEMAND for training:
When they do participate, migrants tend to spend MORE HOURS in training than natives
Migrants are more likely to report that they have not been able to participate in training despite their desire to do so.
The graph shows the rates of participation in adult training activities, i.e. whether the respondent participated in any kind of training in the 12 months preceding the survey.
Lighter colors indicate countries for which the unadjusted difference in participation rates between natives and migrants is not statistically significant.
Here it is important to note that between-country differences in participation rates are much more pronounced than within-country differences between migrants and natives – eg. Migrants in the Nordic countries are much more likely to participate in training than natives in Italy and France, because adult training is much more widespread in the Nordic countries across the board
As pointed out before, migrants are more likely to report that they were not able to participate in training in spite of their willingness to do so. This graph reports differences between migrants and natives in the REASONS given for such inability to participate . On average, migrants seem to face important barriers in terms of the COST of training (which is expected). They are also more constrained by family responsibilities, possibly because they cannot rely on extended family networks that could provide support with childcare, or because they have lower access to services that could help them in this respect.
Across almost all OECD countries, migrants have worst labour market outcomes than natives. In particular, migrants are more often unemployed or inactive, and those employed have lower returns to education – in terms of earnings – than natives.
As the figure shows – foreign-born adults earn significantly less than their native-born peers. However, some migrants groups are doing better than others. Across the participating countries, the hourly wage of natives is only 5% higher than that of migrants educated in the host country, and 16% higher than the hourly wage of migrants with foreign qualification.
Several factors explain this wage gap between natives and migrants. In particular, the report finds that an important share of the difference in returns to education is due to different patterns in occupational placement, with migrants concentrated in less prestigious jobs, that is occupations associated with a lower socio-economic status.
Nevertheless, migrants are often paid less than natives even within the same occupation. Thanks to the richness of the PIAAC data, the analysis in this report shows that the wage gap is largely explained by differences in the skills held and language spoken by these two groups.
The country where the worker completed his education also matters: migrants who completed their studies in the host country, tend to have labour market outcomes which are quite similar to those of the native-born.
Overall these results all point at the key role played by language and skill proficiency as important drivers of the labour market performance of migrants. Policies that help migrants improve their language skills have thus a major role to play here.