Selected Issues

Abstract

Selected Issues

Migrant Integration in Denmark and Europe: Evidence Using Micro Data1

This paper provides evidence on historical patterns of migrant integration in Denmark and Europe using a rich dataset and unified empirical framework. Employment gaps between natives and migrants in Denmark are persistent after many years of residence, in particular for nonwestern migrants, but their integration over time appears more effective in Europe. Upon arrival, female migrants exhibit wider initial employment gaps relative to natives compared with male migrants but they catch-up faster over time. Our findings also indicate that domestic education of migrants is important for successful integration. The extent of migrant education in Denmark is lagging the average for other European countries, suggesting room for improvement.

A. Introduction

1. The refugee surge in Europe has subsided, but integrating the large stock of refugees and other immigrants remains challenging. In Denmark, unemployment among the foreign-born is high and it is most pronounced for non-EU migrants. For integration policy to be most effective, it is important to understand historical patterns of integration in receiving countries.

Figure 1.
Figure 1.

Unemployment among the Foreign Born

Unemployment Rates of Foreign-Born Persons

(Percent of respective cohort)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

Sources: Eurostat and Fund staff calculations.

2. This paper addresses three main research questions related to labor market outcomes for migrants in Denmark and Europe at large:

  • How successfully do migrants integrate in the labor market in Denmark and in Europe?

  • How does integration speed in Denmark compare with other countries?

  • What is the role of education in improving the probability of having a job?

3. To answer these research questions, this paper first explores in section B the micro data from the Labor Force Survey (LFS) by Eurostat. Then, section C lays out the empirical framework and section D illustrates the baseline results for Denmark and Europe. Section E presents examines the role of foreign and domestic education in improving migrant employment probability. Section F concludes.

B. A First Look at the Micro Data

4. Eurostat’s LFS is an extensive database with large coverage of European countries spanning all the way back to 1983, with rich information both at the individual and household levels.2 Key variables of interest include country of birth, age, education, marital status, labor market status, employment characteristics, and survey year. For migrants—whom we identify by country of birth—the length of residency in the receiving country is also of prime relevance. The limitations of the dataset are that it does not allow tracking individuals over time; data on wages or income are missing; it lacks information on language skills; and data for Germany are not provided.

5. In this section, we first provide descriptive statistics for the latest survey year (2014), followed by summary descriptive charts on labor integration outcomes based all survey years considered in the empirical section (2010–2014). For the last survey year (2014), the Denmark sample includes more than 44,000 individuals with a 12 percent share of foreign-born, 45 percent of whom are from a nonwestern origin (Table 1).

Table 1.

Denmark: Distribution by Country of Birth, Eurostat’s Labor Force Survey

article image

Western refers to migrants from Europe, North America, and Oceania; Nonwestern refers to migrants from Asia, Latin America, and Middle-East and North Africa.

Source: Eurostat’s Labor Force Survey and Fund staff calculations.

This sample is evenly split across male and female, with the majority being aged above 45 years of age and a small share of people with low level of education (Table 2).

Table 2.

Denmark: Distribution by Demographics, Eurostat’s 2014 Labor Force Survey.

article image
Source: Eurostat’s Labor Force Survey and Fund staff calculations.

From Table 3, the probability of being employed in 2014 in Denmark is much lower for foreign-born persons (66 percent) than for natives (82 percent). In contrast, there is only a 9 percentage points gap between the employment probability of natives and migrants across all Europe, on average. In Denmark, unemployment among the foreign-born in the 2014 survey sample is twice as high as for natives (8 percent versus 4 percent), and inactivity is much higher among the foreign-born. The micro data also allow us to examine employment rates of migrants after several years of residence in the receiving country. Considering the 2010–2014 surveys, we present summary charts in Figure 2 on the integration outcomes in some countries in Europe. For most countries, there are positive employment and participation gaps between natives and migrants, and these gaps are higher for immigrants born in nonwestern than in Western countries. In Denmark, employment and participation gaps between natives and migrant from nonwestern origin are quite pronounced. Further, for European countries, these employment and participation gaps generally decline after several years of residency in the country, whereas they are persistent in Denmark after 30 years (mostly for nonwestern migrants as the next section will show). These preliminary findings suggest that labor market prospects for migrants could be more favorable in other European countries than in Denmark.

Table 3.

Denmark: Labor Market Outcomes, 2014 LFS

article image
Source: Eurostat’s Labor Force Survey and Fund staff calculations.
Figure 2.
Figure 2.
Figure 2.

Integration Outcomes and Years of Residency

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

These gaps generally decrease with greater years of residence in the country, but they remain open in Denmark after 30 years of residence.

C. Empirical framework

6. To investigate employment integration in Europe, we follow the established literature on earning assimilation of migrants (Chiswick, 1978; Borjas, 1985; Borjas, 2015; Frieberg, 2000) and estimate the following baseline probit regression specification:

Pr(Employed)=α0+α1Mig+α2Ysm+α3Ysm2+α4Age+α5Age2+α6Edu+α7X+ε(1)

Where Pr(Employed) is the probability of having a job, which can assume either a value of 0 or 1; Mig is a dummy variable for migrants; Ysm refers to years since migration; Age and Edu are the age and education level of the person surveyed; and X includes other controls such as marital status, country, and survey year indicators. In equation (1), α1 provides the initial migrant-native employment gap holding key person characteristics constant and α2 proxies the catch-up integration speed of migrants (which could be non-linear).

7. We allow for the integration profile of migrants to vary by origin and gender, and run the baseline specification in (1) for Denmark only and for all 13 European countries. We only consider the period 2010–2014 to avoid the crisis period and ensure a more balanced data coverage, as there is less coverage for surveys from earlier years. We also split the sample by Western and nonwestern origin of migrants to investigate differences in employment probabilities and integration speeds among the two categories of migrants.

8. Further, we investigate an extension of the baseline specification to examine the role of education, both foreign and domestic, in shaping the probability of being employed. To do so, we decompose total education into that accumulated abroad in the foreign country (Eduf) and in the receiving or domestic country (Edud), as in Friedberg (2000). We similarly decompose Ysm into Edud and work experience accumulated in the receiving country but, since information on the latter is not available, we use Age as proxy. To impute the amount of foreign and domestic education in the micro data, we use information on year of migration and year of completion of education, assuming education (in years) is continuous. We obtain equation (2), in which coefficients of interest are β2 and β3:

Pr(Employed)=β0+β1Mig+β2Eduf+β3Edud+β4Age+β5Age2+β6X+ε(2)

We also explore the possibility of different returns to domestic education for natives and migrants. To that end, we incorporate an interaction term between Mig and Edud, allowing for the impact of domestic schooling to differ between natives and migrants. In equation (3) below, we are interested in the parameter estimate δ4.

Pr(Employed)=δ0+δ1Mig+δ2Eduf+δ3Edud+δ4Mig×Edug+δ5Age+δ6Age2+δ7X+ε(3)

We use the above multivariate framework to estimate the (a) employment gap between natives and migrants by origin upon arrival, (b) integration speed of migrants, and (c) marginal effect of domestic education on natives and migrants.

D. Analysis of Integration Speed

9. Table A1 (see Appendix 1) presents summary statistics for key variables. Out of close to 3.5 million individuals aged between 22 and 62 years across Europe, the probability of having a job is 72 percent. Immigrants represent 13 percent of total people surveyed and they have been residing in the receiving country for an average of 19 years. A surveyed individual is on average 45 years old, has completed 12 years of education, and has a 60 percent probability of being married.

10. Table A2 displays the estimation results of equation 1 for the Denmark and European samples, showing only the key variables of interest, Mig and Ysm. The estimated parameter α1 from equation 1 is negative and significant indicating that, conditional on having the same age, education, and other key characteristics, a migrant upon arrival has a lower probability of being employed relative to a native. Considering all migrants together, the parameter estimate on Mig in the upper panel on Denmark is larger in absolute terms than for the average all European countries in the lower panel of Table A2. These results suggest wider initial employment gaps in Denmark relative to Europe for all migrants. However, the initial employment gap between nonwestern male migrants and natives is lower in Denmark (α1 = −0.565) than in Europe (α1 = −0.842). More broadly, initial employment gaps are wider for migrants from nonwestern origin than for migrants of western origin, and they are mostly pronounced among female.

Also from Table A2, the estimated parameter α2 from equation 1 is positive and significant for the Europe sample, suggesting that spending more time in the receiving country improves the probability of migrants finding a job. However, in the case of Denmark, these results are only significant for female migrants, whether they have a Western or nonwestern origin.

11. Using the results from equation 1, we calculate migrant employment probabilities upon arrival and after 10 years of residence in the country for Denmark and all Europe and illustrate the change in these probabilities in Figure 3. In both Denmark and Europe, female migrants are more likely to be employed after spending 10 years in the receiving country. However, for nonwestern female migrants, this probability is twice as high in Europe as in Denmark. For nonwestern male migrants, their employment probability in Europe increases by close to 9% percentage points after 10 years, but it declines in Denmark by 10 percentage points.

Figure 3.
Figure 3.

Change in Employment Over Time

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

Sources: Eurostat's labor Force Survey and Fund staff estimates.

12. We also use the results of the baseline regression from equation 1 to simulate the integration profile of a 30-year old migrant (of Western and nonwestern origin) in Denmark and all Europe, conditional on having the same age, education, and other key characteristics as the average representative individual in each sample. We illustrate the findings in Figure 4 and summarize them as follows. First, for both female and male in Denmark and Europe, the integration of migrants from nonwestern origin persistently lags that of migrants from western origin. Second, upon arrival to the receiving country, all female migrants exhibit a wider initial employment gap compared with male migrants and their catch-up speed is faster over time. Third, the employment probability of female nonwestern migrants after 20 years of residence in Denmark is close to the corresponding average for Europe (at about 60 percent). Fourth, whereas the employment probability of a nonwestern male migrant in Europe gradually catches up with that of other male migrants and natives, it gradually drops over time in Denmark, despite an initial employment gap upon arrival that is similar to that of western migrants.

Figure 4.
Figure 4.

Simulation of Integration Profile, 30-Year Old Migrant

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

13. Next, we consider country-specific differences in initial employment gaps and integration speed, re-running equation 1 separately for each country. For each country, we calculate marginal effects for the conditional employment gap between natives and migrants upon arrival. The results (sorted for male migrants in Figure 5) confirm that the initial employment gap between natives and migrants in Denmark, both for female (about 25 percent) and male (about 10 percent), is not as pronounced as for other countries. Using the country-by-country results, we also calculate the marginal effects of the employment probability with respect to years since migration, showing the results sorted for male migrants in Figure 6. Each additional year of residence in Denmark increases the probability of employment for female by 1 percentage point, which is larger than for many other countries in Europe. However, the employment probability of male migrants in Denmark declines by ¼ percentage point with each additional year of residence in the country, although this finding is not statistically significant.

Figure 5.
Figure 5.

Initial Employment Gap between Migrants and Natives, By Country

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

Figure 6.
Figure 6.

Estimated Integration Speed, By Country

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

Sources: Eurostat and Fund staff estimates.

E. The Role of Education

14. Studies by Borjas (1985, 2015) contend that the economic impact of immigration on a receiving country ultimately depends on the skills composition of the immigrant population. In this section, we analyze the role of both foreign (acquired prior to migration) and domestic (acquired in the receiving country) education in improving migrant integration outcomes. Tables A3 and A4 present the estimation results of equations 2 and 3, respectively.3

15. The positive and significant coefficients in Table A3 (β1, and β2 from equation 2) indicate that both foreign and domestic education increase the probability of employment for all migrants, regardless of their country of origin. The size of the parameter estimate on domestic education is also greater than for foreign education, suggesting that the pay-off from domestic education—in terms of a higher employment probability—is greater than for foreign education. To assess the economic significance of differences in employment probabilities conditional on other individual characteristics, Figure 7 shows the marginal effect of having one more year of domestic and foreign education for female and male in Denmark and Europe, also broken down by migrant origin.

Figure 7.
Figure 7.

Foreign vs. Domestic Education

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

16. We summarize three key takeaways from Figure 7. First, domestic education raises the probability of employment more than foreign education. Second, in all countries, the employment gains for female individuals are much higher than for male after 10 years of education in the country. In Denmark, 10 years of domestic education for female raise their chances of having a job by 30 percent whereas the corresponding probability for male is 20 percent. Third, there does not appear to be significant differences in the returns to foreign education of nonwestern male migrants to Denmark or other countries in Europe: 10 years of foreign education boost their employment probability by around 20 percent. For nonwestern female migrants, however, 10 years of foreign education raise their chances of having a job by above 30 percent in Europe whereas the corresponding gain in Denmark is 20 percent.

17. After coming to a country, migrants also receive domestic schooling. However, the extent of such education in Denmark—measured in years of education received—is lagging the average for other European countries (Figure 8).4

Figure 8.
Figure 8.

Migrant Domestic Education

Average Years of Migrant Education in Receiving Country

(Years per migrant)

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

Sources: Eurostat Labor Force Survey (2010-2014) and Fund staff calculations.

18. Table A4 shows estimation results of equation 3, where the parameter estimate δ4 on the interaction term (Domestic education* Migrant) allows for different returns to domestic education for natives and migrants. For the results in the upper panel on Denmark, δ4 is insignificant and it flips sign based on the subsample considered. In contrast, the results for Europe in the lower panel show a negative and significant interaction term, suggesting that the employment probability of a migrant is lower than that of a native, after they each receive domestic schooling.

19. Figure 9 illustrates the marginal effects of each additional year of domestic education, which results in lower employment probability for migrants than for natives. There are also differences among migrants of difference origin as well as among female and male. Except for female migrants in Denmark, the employment probability increases more for western than nonwestern migrants after receiving domestic education. The employment probability of nonwestern female migrants in Denmark (Europe) rises to close to 55 percent (slightly below 50 percent) after 10 years of domestic schooling, compared with around 70 percent for natives. For nonwestern male migrants in both Denmark and Europe, the probability of being employed hovers around 60 percent after 10 years of domestic schooling; the corresponding figure for natives is about 80 percent.

Figure 9.
Figure 9.

Domestic Education, Natives vs. Migrants

Citation: IMF Staff Country Reports 2017, 159; 10.5089/9781484304617.002.A002

F. Conclusion

20. Understanding the complex process of migrant integration could benefit policy assessment on how to maximize the economic impact of immigration on a receiving country. Using a rich dataset and unified empirical framework, this paper provides new evidence on the labor market integration of migrants in Denmark and across Europe.

21. The main findings are summarized as follows:

  • Overall employment and participation gaps between natives and migrants from nonwestern origin are high in Denmark. They are also persistent after 30 years of residence, unlike for most other European countries.

  • Initial employment gaps between natives and migrants are higher among female than male individuals and they are more pronounced for migrants from nonwestern origin.

  • Despite initial better conditions for nonwestern male migrants in Denmark relative to Europe, their employment probability does not improve with years of residence, unlike for female migrants to Denmark and for other migrants in Europe.

  • Education acquired by migrants prior to arrival matters, albeit to a lesser extent than domestic schooling.

  • Domestic education of migrants to Denmark—which lags other countries in Europe—is key to raising their probability of employment.

In view of these individual factors determining employment outcomes of migrants. integration policy should pay specific attention to vulnerable groups such as female and migrants from nonwestern countries, given different initial conditions. Integration policy could also aim to boost the return to foreign education, such as by improving the validation of foreign degrees. Further, it is important to enhance the pay-off to domestic education for migrants, including by improving language training.

References

Appendix I. Additional Tables

Table AI.1

Variables Entering the Baseline Specification

article image
Eurostat’s Labor Force Survey and Fund staff calculations.
Table AI.2

Estimation Results of the Baseline Specification

article image
Note: All regressions include controls (not show n) for education, age, age squared, marital status, and year and country fixed effects. Robust standard errors in brackets; *** p<0.01, ** p<0.05, * p<0.1.
Table AI.3

Estimation Results of Equation (2)

article image
Note: Regressions include all controls (not shown) specified in equation (2). Robust Standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1
Table AI.4

Estimation Results of Equation (3)

article image
Note: Regressions include all controls (not show n) specified in equation (3). Robust Standard errors in brackets. *** p<0.01, ** p<0.05, * p<0.1
1

Prepared by Rima A. Turk, based on a forthcoming working paper with Giang Ho. We are grateful to Eurostat for providing us with micro data from the Labor Force Survey. The results and conclusions are ours and not those for Eurostat, The European Commission, or any of the national authorities whose data have been used. We would also like to thank participants at the IMF European Department seminar for useful comments.

2

There are 13 countries included in the analysis: Austria, Belgium, Denmark, Finland, France, Ireland, Italy, Netherlands, Norway, Portugal, Spain, Sweden, and the UK.

3

In this paper, foreign and domestic education are imputed from the Eurostat data, as they are not provided by the LFS. It is noteworthy, however, that Statistics Denmark is currently improving the definition of foreign education, which would help better understand the skills composition of migrants.

4

The OECD PISA 2015 key findings similarly indicate that, while the average Danish performance of all 15-year old in science, mathematics, and reading exceeds the OECD average, immigrant students in Denmark do not perform as good as the average for other countries (OECD, 2016). Also, the mean literacy scores in 2012 for the 16-64 year-old foreign-born were lower for Denmark than for the EU and OECD averages (OECD, 2015).

Denmark: Selected Issues
Author: International Monetary Fund. European Dept.