Using EU Indicators of Immigrant Integration - European ...

Using EU Indicators of Immigrant IntegrationReport prepared for the European CommissionThis project was carried out between January 2012 and April 2013 by the European ServicesNetwork (ESN) and the Migration Policy Group (MPG) on behalf of the European Commission.ESN was responsible for the overall management of the project and the organisation ofseminars held in Berlin, Budapest and Lisbon. MPG was responsible for the research and forengaging policy-makers and practitioners. An ad hoc research team was created with asmembers: Jan Niessen, (Director), Thomas Huddleston (Coordinator) and Jasper Dag Tjaden(Researcher) and Professors, Pierre Desmarez, Dirk Jacobs, Síle Padraigin O’Dorchai andAndrea Rea of the Free University of Brussels (for quantitative analysis) and Albert Kralerand David Reichel of the International Centre for Migration Policy Development (for theassessment of data sources).DisclaimerThis report was prepared by external consultants and does not necessarily reflect the viewsof the European Commission.Brussels March 2013

IntroductionThis report on the further development and use of EU immigrant integration indicators in policydebates is prepared at the request of the European Commission by the European Services Network(ESN) and the Migration Policy Group (MPG). It is based on research undertaken by an ad hoc researchteam lead by MPG including scholars from the Free University of Brussels (ULB) and the InternationalCentre for Migration Policy Development (ICMPD) in Vienna.The initial research results were presented in discussions papers which were first discussed by theEuropean Commission and the National Contact Points on Integration (NCPIs) and subsequently byaround 300 governmental and non-governmental integration actors and academics from all 27 MemberStates and Norway attending three expert seminars that were organised in the course of 2012.Participants of the seminars were asked to make presentations on the development and use ofintegration indicators at local, regional, national and European levels. Concluding documents summarisedthe seminars’ main findings. In this way, the European Commission and the NCPIs helped torefine research questions and to test how results can be interpreted and used to inform policies. Theseminar participants got a better understanding of how policies and outcomes can be measured andmonitored, as is being done in an increasing number of countries.This final report has three parts which are preceded by this introduction, an executive summary anda background chapter.Part I of this report explores how three types of factors influence societal integration outcomes in fourareas and as such can inform integration policies. The three types of factors concern personal characteristicsof the immigrant population, the general context in the country and its specific migration andintegration policies. The four areas are employment, education, social inclusion and active citizenship.In those areas, the European Union selected an initial number of indicators (the Zaragoza indictors)which are considered in this project. Overall, the analysis reconfirms the relevance and usefulness ofthe Zaragoza indicators.Part II reconfirms the availability, accessibility and reliability of the main sources for the calculationof the integration indicators considered within the report. They include the EU-Labour Force Survey(EU-LFS), EU Statistics on Income and Living Conditions (EU-SILC), OECD’s Programme for InternationalStudent Assessment (PISA) as well as Eurostat’s migration statistics. These are well-establishedinternational and comparative data sources which build upon data that is gathered nationally, oftenby national statistical institutes.Part III presents different options reflecting the different ways in which indicators could be used to understandnational contexts, evaluate the outcomes of policies, and use targets to improve integration.It takes existing national and EU indicators as starting point for reflection. Indicators can be used todescribe the (constantly changing) situation in societies with citizens and residents with and withouta migration background. Indicators can also be used to clarify the link between integration policiesand societal outcomes, for example by monitoring the beneficiaries of policies and conducting robustimpact evaluations.The report is explorative and descriptive in nature. Considering the advantages and limitations ofinternational and comparative research, the results of this report represent a substantive contributionto the on-going debate and research on the development and use of integration indicators on whichintegration actors can build.4

Executive SummaryThis report is based on the project’s own researchand analysis of existing international quantitativeand qualitative research. Consultations withintegration actors from across Europe providedvaluable information and insights. The project’soutcomes can be summarised as follows:1.1 Analysis1. Time of residence matters: Integration is along-term process. Immigrants show betteroutcomes, the longer they are in the country.Just as for non-immigrants of working age,outcomes also improve generally with age.2. Gender matters: Across all Zaragoza indicators,foreign-born women and mothers areamong the most vulnerable groups, in particularthose born in non-EU countries.3. Origin country matters: Compared to immigrantsfrom EU or highly-developed countries,immigrants from non-EU countries, in particularless developed countries, have on averagedifferent reasons for migration, differentaspirations for the future, and different typesof problems. Non-EU immigrants face greaterchallenges on the labour market, the housingmarket and in schools. They are more likely tobe affected by poverty and over-qualification.Immigrants from lower developed countriesare also more likely to naturalise.4. Socio-economic background matters: Socialmobility remains modest for immigrants. Employmentand education outcomes still largelydepend on the parents’ socio-economic status.Immigrants are more likely to be concentratedin lower track, low-performing schools with alow average socio-economic status. Childrenof immigrants with low socio-economic statusface much greater difficulties advancing intohigher education.5. Quality matters: Integration is more than havinga job, housing, and basic education. Employedimmigrants more often work part-time,temporary, below their qualifications and withwages that are insufficient to protect themfrom poverty than non-immigrants. They aremore likely to live in overcrowded housing andpay more of their income for rent. Foreignqualifications are often devalued on the labourmarket or not formally recognised.6. Discrimination matters: Unequal treatment hampersintegration. Immigrants are less likely to behired even when their qualifications are similarto non-immigrants. Immigrant students areless likely to be referred to higher track educationeven when their grades are similar to theperformance of non-immigrants. In countrieswith larger gaps between immigrants and nonimmigrants,the public is more likely to say thatdiscrimination against migrants is a problem.7. Context matters: Structures of society shapeintegration. Immigrants tend to have betterlabour market outcomes, perform betterin schools and participate more in countrieswhere the general population has higher outcomes.Immigrants from the same countryof origin and with a similar social backgroundperform differently in different countries.Some welfare systems protect immigrantsfrom poverty better than others and someeducation systems are more favourable forlow-performing immigrants. Across the board,more research is needed on the direct impactof the welfare system, the education system,housing and general labour market structureson integration outcomes of immigrants.8. Policy matters: There is a lack of rigorous impactevaluations of policy effects in the EU.There is evidence that policies are directly relevantfor some EU integration indicators. Forexample, Employment rates tend to be higherin countries with a larger share of work migration.Welfare systems in some countries weremore successful in reducing poverty than others.Richer, more equal countries tend to adoptmore ambitious integration policies. More immigrantsnaturalise in countries where naturalisationpolicies are more open. While there is abetter understanding of which countries adoptpolicies and which outcomes seem to be associatedwith which policies, very little is knownabout the causal effects of specific policies orprogrammes.Employment9. Highly-qualified immigrants are most likelyto be overqualified for their job. Easier andmore accessible recognition procedures,equivalence courses and European cooperationcould facilitate the recognition of qualificationsand skills.10. Immigrants and their descendants are underrepresentedin the public sector which is5

6a major share of the job market in many EUcountries. Public sector employment targetsand targeted information campaigns can increaseapplication rates of eligible immigrants.11. Little is known about the impact of legal restrictionsfor employment of some migrant groupsand participation in labour market programmeson immigrants’ employment outcomes.12. Immigrants face severe discrimination on thelabour market. More accessible and coherentanti-discrimination legislation, stronger equalitybodies, more teacher and public sector trainingsas well as anonymous job applicationscould enhance the situation.Education13. Quality early childhood education and care isassociated with better education outcomesfor immigrants at the age of 15. Access andquality of early childhood education andcare can have an impact on immigrants’ longterm education careers.14. Education outcomes improve over time. Generaleducation policies can accommodate recentimmigrants by providing homework andother general school support for the young,language tuition for all, equivalence classesand access to life-long learning for adult immigrants.Targeted policies can tackle longersettled groups with lower achievement.15. Education outcomes still largely dependon the parents’ social background. There isevidence that decreasing socio-economicsegregation in schools, increasing the hoursspent in school, improving the quality ofteaching, delaying the age of tracking andsupporting students before and during thetransition into higher education can reducethat link. Smaller classes and more parentalinvolvement have proven to be effective forimproving immigrant children’s outcomes.16. Foreign trained immigrants’ qualifications areoften not recognised or the skills and qualificationsdo not fit current labour demand. Opportunitiesfor adult migrant learners to upgradeor equalise their qualifications, including byproviding easier access to lifelong learning canhelp to enhance employability of immigrants.17. Immigrant students with good potential faceobstacles of discrimination in schools. Away to tackle this is providing discriminationawareness trainings and support for teachersto deal with second language teachingand diversity in class rooms.Social Inclusion18. The results of the project’s seminar suggestthat social inclusion indicators are rarely partof the debate on migrant integration despitethe fact that social inclusion indicatorssuch as income, poverty, health and housinglargely affect other areas of integration, suchas education and employment. Immigrantshave on average lower incomes, higher povertyrisk, higher in-work poverty and worsehousing conditions. Some evidence suggeststhat immigrants actually are less likely touse social benefits than non-immigrantsif several factors are accounted for. Moreresearch is needed on the impact of socialbenefits on migrant integration.19. Immigrants are more likely to live in biggerfamilies and overcrowded housing. However,not enough is known about how housing policiesaffect integration outcomes. Evidence isneeded on accessing the housing market andits impact on the situation of immigrants.20. Especially immigrant women in large householdsare affected by poverty. There is a goodreason for reviewing compliance with andimplementation of gender equality legislationand the effects of family and unemploymentbenefits on migrant families.Active Citizenship21. Citizenship and long-term residence are onlytwo elements of active citizenship. Indicatorson other forms of civic participation of migrantssuch as voting, membership in organisations,running for or holding a political office,protesting or volunteering are needed to captureimmigrant’s political and civic involvement.22. The interaction between access to citizenshipand integration is complex. Naturalisation isboth a final step in a process and as a toolto further improve integration in several areasof life. Citizenship is a societal outcomeindicator, a policy indicator and a measureof openness of receiving societies, all at thesame time.23. In many EU countries, many immigrants inthe country more than ten years have notnaturalised. These people are still seen asforeigners and largely excluded from thedemocratic process.

24. Immigrants become citizens and long-termresidents more often in countries where theprocess is more inclusive and where dualcitizenship is accepted in both the country oforigin and destination. Beyond legal changesfacilitating naturalisation, support to paynaturalisation fees and minor changes ofadministrative procedures could facilitateacquisition of citizenship.25. Naturalised immigrants have on averagebetter integration outcomes than non-naturalisedin most countries, regardless ofwhether naturalisation policies are inclusiveor restrictive. It remains unclear whetherthis ‘citizenship premium’ is greater or lesserin certain countries due to their policies orto other factors. More research is needed onwho benefits most from naturalisation.The relevance of the Zaragoza indicators26. The project analysed relevant factors thatinfluence migrant integration with a view toinform integration policies at various levelsof governance. It reconfirms the relevance ofthe Zaragoza indicators and proposes a fewadditional ones.27. The Zaragoza indicators are relevant forachieving the goals of the Europe 2020 andET2020 Strategies and benchmarks. To illustratethe impact of effectively integratingimmigrants into the EU agenda, the projectdeveloped a ‘closing the gap-scenario’ bywhich equal outcomes of the migrant populationin comparison with the total populationare assumed.28. Currently, the total employment rate in theEU is 69 %. The employment rate for the foreign-bornis 64 %. The Europe 2020 target isto increase the overall rate to 75 %. Closingthe employment gap for foreign-born immigrantsaccounts for 10.7 % of meeting theEurope 2020 target across all EU countriesfor which targets and data are available. Giventhe ‘no gap scenario’, Austria, Germany,the Netherlands, and Sweden would halfwaymeet their national Europe 2020 target.29. Member States could prevent half a millionpeople from leaving school early, ifthey could close the gap for migrants. Thisaccounts for 8.7 % of all early school leaversin the EU. The EU as a whole would be30 % closer its headline target of reducingthe early school leaver rate from 14 to 10 %.The ‘no gap scenario’ accounts for more than50 % of reaching the target in Belgium, Cyprus,Denmark, Germany, Greece, and Italy.In fact, Sweden would exceed its nationaleducation target.30. 23 % of the EU population is at risk of povertyor social exclusion. The rate is 9 % higher forthe foreign-born population (32 %). If this gapwere closed, the EU could lift 3.3 million immigrantsout of poverty or social exclusion. Thisnumber accounts for 5 % of all people at riskof poverty or social exclusion in the EU. Thisstands for 17 % of all people at risk of povertyor social exclusion in Austria, 19 % in Belgium,and almost 19 % in Sweden. Closing the gapfor immigrants would bring the whole of theEU 16.2 % closer to reaching its headline povertytarget. The migrant gap represents morethan 50 % of the national targets in Austria,Belgium, Greece, and the Netherlands.1.2 Data sources31. The indicators of immigrant integration definedin the 2010 Ministerial Conference inZaragoza are based on high quality internationaldata collections. Data sources used arethe best available ones for monitoring the indicatorsdefined in the Zaragoza Declaration.32. As these data sources were not originally developedto measure outcomes for migrants,several improvements regarding data availabilityand reliability should be pursued, such as:Increasing sample sizes: The main problem withavailability and reliability of data for immigrantintegration indicators stems from small samplesizes. To better capture immigrants in the existingdata sources mainly two options are available:1. Oversampling: The sampling strategiescould be adapted to sample more immigrantsinto the samples. Oversampling does not onlyimprove the reliability and scope of analysis ofthe Zaragoza indicators that use the LFS, SILC,PISA and Eurostat data. It would also create theopportunity to include additional indicators. Forexample, boosting immigrant samples in theEuropean Social Survey would allow to measureimmigrants’ civic and political participation.2. Pooling of samples over years: Combining thedata from two years would double the samplesizes and will considerably improve availabilityof data in several countries, also allowing forfurther breakdowns in more countries. For increasingavailability and reliability, pooling datafrom three years might be considered, even if, asa result, migrant integration indicators could notbe produced every year. Pooling can be a cost-7

8effective short-term solution to make availablemore data to measure immigrant integrationwhile it has to be kept in mind that some qualityof data is lost in the process.Showing uncertainty in results: Since resultsbased on sample surveys are inevitably limited dueto sampling errors, the provision of confidence intervalswould increase the visibility of uncertaintyin the estimates. Showing confidence intervalsmight increase the credibility of results based onsmall samples. If minimum sample size requirementswere lowered, results could be published formore countries. Confidence intervals indicate rangesof the estimates, in which the true populationfigures are very likely to be (by convention 95 %confidence intervals). The estimation of these intervalsdepends on the sampling designs in thecountries, which are different across EU countries.Consequently, the confidence intervals for totalpopulation and for immigrants have to be providedby the National Statistical Institutes (NSIs).Harmonise methods of data production:Currently, different parameters for weightingsurvey estimates of the main surveys (especiallythe EU LFS) are employed across EU MemberStates. Some countries include estimates on theforeign born or foreign population to account fornon-response and some countries do not. It isimportant to harmonise weights across countriesaccording to country of birth or citizenship forproviding more reliable comparisons of the estimates.There is a lack of research on weightingmethods and practices by country of birth andcitizenship and its impact on data quality.33. If the previous recommendations were also appliedto other data sources, such as the EuropeanSocial Survey, they could be useful to exploreadditional indicators of migrant integration.34. Data harmonisation and improvement at theinternational level needs good coordination.Eurostat has set priorities for mainstreamingmigration statistics into general data collections.Eurostat should continue to take aleading role in reinforcing and coordinatingresearch on the issues mentioned above inclose cooperation with the NSIs.1.3 Using indicators35. There are three key policy purposes for usingintegration indicators: understanding integrationcontexts and immigrants’ integrationoutcomes, evaluating the results of policies,and mainstreaming integration into generalpolicies. These purposes are not one in thesame. Measuring the situation of immigrantsis a different exercise from the evaluation ofthe results of integration policies. The resultsof these policies cannot all be measured interms of immigrants’ outcomes on integrationindicators. For integration stakeholders,not all Europe 2020 targets are relevant forintegration. For policymakers working in otherfields, many of their policies are not significantlyaffected by the situation of immigrantsor the results of integration policies.36. The EU integration indicators make it easierto understand the integration context in theEU Member States so that policy actors canbetter learn from one other. The results showthe similarities and differences in national contexts,while further analysis reveals what factorsexplain these similarities and differences.Integration outcomes in different countries areoften related to the same key aspects of theimmigrant population, the general context, andnational policies. The more these factors arepresent in a country, the more likely are certainintegration outcomes in that country. Thisanalysis helps policy actors to appreciate theunique combination of factors influencing integrationin their country and other countries.Moreover, the use of indicators over time givespolicy actors a new long-term perspective forpolicy planning. The availability of these indicatorsis therefore a starting-point for more informedmutual learning across the EU.37. Policy actors have various options for usingintegration indicators at various levels, whichcan be implemented and combined in differentways and at different times. The mainoptions are summarised in the chart below.38. A more structured and regular integrationmonitoring at EU level can see the best resultsby building on existing national and Europeandata collection mechanisms. In this way duplicationis avoided. Existing data can be mademore accessible for policy actors and scientistswho can help to demonstrate how datacan be used responsibly and meaningfully.This may also incite further reflection about apossible expansion over time: from a simplerto a more sophisticated system. Experiencesat national, regional and local levels are importantin this respect as well.39. Integration policy makers and practitioners inMember States can learn from each other inhow indicators can be used in policy debates.The NCPI meetings and the European IntegrationForum are platforms for such exchange.

In addition, in some countries and in some integrationfields deeper analysis is undertakenfrom which others can learn.40. An annual or multi-annual report on integrationalong the lines of Eurostat’s StatisticalPortrait would be a promising start. This couldbe expanded to include more focused contextualinformation reflecting the specificities ofthe immigration population and the beneficiariesof specific policies in each country aswell as valuable information on integrationoutcomes and policies at national and, wherepossible, regional and local level. A next stagecan be when various types of analyses areused to measure impact.List of Zaragoza indicators and additional indicators to be consideredEmployment Education Social Inclusion Active Citizenship Welcoming SocietyZaragoza indicatorsProposed New IndicatorsEmployment rateHighest educationalattainmentAt-risk-of-poverty(and socialexclusion)Naturalisation rateUnemployment rate Tertiary attainment Income Share of long-termresidenceActivity rate Early school leaving Self-reported healthstatus (controllingfor age)Share of electedrepresentatives(research)*Self-employment Low-achievers (PISA) Property ownership Voter turnout(research)*Over-qualification Language skills ofnon-native speakers(LFS module)**Public sector employmentTemporaryemploymentPart-timeemploymentLong-termunemploymentShare of foreigndiplomasrecognised(survey)**Retention of internationalstudents(research)*Participation in earlychildhood education(SILC/PISA)**Participation in lifelonglearning(LFS, AES)Not in education,employment or training(LFS)Resilient students(PISA)**Concentration in lowperformingschools(PISA)**Child poverty (SILC)Self-reported unmetneed for medicalcare (SILC)Life expectancy(SILC)Healthy life years(SILC)Housing cost overburden(SILC)**Overcrowding (SILC)**In-work poverty-risk(SILC)Persistent povertyrisk(SILC)Participation in voluntaryorganisations(survey)*Membership in tradeunions (survey)*Membership in politicalparties (survey)*Political activity(survey)*Perceived experienceof discrimination(survey)*Trust in public institutions(survey)*Sense of belonging(survey)*Public perception ofracial/ethnic discrimination(Eurobarometer)Public attitudesto political leaderwith ethnic minoritybackground (Eurobarometer)Note: One star (*) marks indicators for which data needs to be collected or migrant sample sizes boosted.Two stars (**) mark indicators for which data is not available every year (ad hoc basis). The authors of this study propose a newcategory of indicators of the ‘welcoming society’. It includes the already proposed Zaragoza indicators ‘perceived discrimination’,‘trust in public institutions’ and ‘sense of belonging’.9

Proposal: An incremental approach to using integration indicatorsWork in Progress:Maintain the baselineImproved annual publicationof EU indicators.What: Indicator results on Eurostatwebsite & basic descriptive analysisin a ‘Statistics in Focus’ reportHow: Eurostat updates theresults. Commission & MemberStates consideration of additionalindicators; NSIs improvedata availability and qualitywhen possible (see Part IIAssessment)Existing resources: Eurostat& NSIs on-going work to mainstreammigrants in statistics,boost migrant samples,& develop ad hoc modulesfor surveysSecond option:Build on the baselineThird option: EvidencebasedpolicymakingMultivariate & longitudinalanalysisWhat: Analysis determines whichcontextual factors have most/leastinfluence on integration outcomesHow: Research analyses therelationship between EU integrationindicators and three setsof factors (differences in theimmigrant populations, generalpolicies & contexts, immigration& integration policies). Longitudinaldata can also be analysedwhere possible.Existing resources: Existinginternational datasets on thesecontextual factors, nationallongitudinal datasets, nationalmultivariate analysisUnderstanding the contextof integration (statisticians,researchers)Multiannual ‘integration report’based on indicatorsWhat: In-depth descriptive analysisof the integration situation& different national contextsHow: Data disaggregated forspecific groups, monitored overtime, & compared betweenimmigrants & non-immigrantswith same demographic characteristics(i.e. comparing ‘like withlike’ through statistical controls).Report can also include ad hocthematic chapters.Existing resources: Examplesinclude 2011 Eurostat ‘StatisticalPortrait’, 2012 OECD ‘SettlingIn’, as well as national and localreportsEvaluating the resultsof policies (researchers)Compare data on policiesand outcomesWhat: Analysis of the complexrelationships between integrationpolicies & outcomesHow: Define policies in terms ofintended outcomes on integrationindicators; conduct bivariateanalysis & mutual learningabout how policies influenceoutcomes and/or how outcomesare influencing policiesExisting resources: Up-to-date& comparable summaries ofpolicies through EU-fundedresearch, EMN, & EWSIGather contextual dataon who benefits from policiesWhat: In-depth descriptiveanalysis of the implementationof policies in the four indicatorareasHow: Gather and share statisticson policy implementation, the potentialand current beneficiariesExisting resources: Administrativeand official statistics, EUfundedresearch, EMN, & EWSIEconometric causalevaluations of policy impactWhat: Assesses prospective orretrospective impact of specificnational policies on integrationoutcomesHow: Evaluation studies done atnational level, sub-national level,or between countries, dependingon the availability of data.The Commission could provide areview of such studies, exchangeon methods, & quality standardsExisting resources: Examples ofcausal evaluations in EU and traditionalcountries of immigration.Using targets to improve& mainstream integration(policy actors)Keep integration indicatorsrelevant for EU targetsWhat: Europe 2020 targets arebasis for integration indicators& thus can now be monitoredfor immigrantsHow: EU integration indicatorscan be changed to reflectchanges in Europe 2020 targetsand statisticsExisting resources: Coordinationamong DGs & EurostatCalculate how integrationimproves EU targetsWhat: Calculations identify areas& countries where immigrantsare a major target group forgeneral policiesHow: Statisticians calculate how‘closing the gap’ for immigrantshelps EU countries meet theirtargets for smart, sustainable,and inclusive growth.Existing resources: EU OMCs& EurostatSet specific national goals& targets for integrationWhat: Member States assessindicator results & set their ownspecific national goals & voluntarytargetsHow: Results assessed basedon best available multivariateanalysis & policy impact evaluations;Member States learn fromcountries using targetsExisting resources: EU OMCs,cooperation structures betweenrelevant ministries, and withcivil society10

Chapter 1:Background on a pilotproject to developand use Europeanintegration indicatorsIndicators have gradually gained importance inEU debates on general socio-economic participationand later on immigrant integration. TheCommon Basic Principles for immigrant integrationpolicy in the EU, adopted by the Justice andHome Affairs Council on 19 November 2004 1 ,stated among other things that developing cleargoals, indicators and evaluation mechanisms arenecessary to adjust policy, evaluate progress onintegration and to make the exchange of informationmore effective.In 2009, a German government conference in Berlinsurveyed national experiences in monitoringintegration and integration policy, and the SwedishPresidency Malmö Conference Conclusions 2defined a list of core areas and indicators, whichare based on the Common Basic Principles and EUindicators in the EU2020 Strategy 3 , covering fourcore areas of integration: employment, education,social inclusion, and active citizenship. In eacharea, Member States identified an initial set of afew core indicators that are simple to understand,easy to communicate, comparable over time andfor which a certain outcome is desirable. The selectionof indicators was based on the availabilityand quality of comparable data. In total, 14 coreindicators were proposed together with a few indicatorsto be developed, which Member Statesalso considered important to monitor althoughcomparable data was still lacking.The 2010-2014 Stockholm Programme committedthe European Commission and MemberStates to develop a limited number of core indicatorsabout the results of integration policies.The 2010 Zaragoza meeting 4 of the ministersresponsible for integration agreed on these EUindicators, which were then approved at the Justiceand Home Affairs Council on 3-4 June 2010 5 .The Council also agreed for the Commission tolaunch “a pilot project with a view to the evaluationof integration policies, including examiningthe indicators and analysing the significance ofthe defined indicators taking into account the nationalcontexts, the background of diverse migrantpopulations and different migration andintegration policies of the Member States, andreporting on the availability and quality of thedata from agreed harmonised sources necessaryfor the calculation of these indicators.”Eurostat’s 2011 pilot study 6 addressed the availabilityand quality of data from harmonisedsources for the calculation of the 15 indicators,for which comparable data could be compiled. Italso published a statistical portrait of the firstand second generation 7 .The Commission’s July 2011 European Agenda forthe Integration of Third-Country Nationals 8 seescommon indicators as a way to systematicallymonitor the integration situation and EU2020targets, enhance policy coordination and makerecommendations in dialogue with the MemberStates. The Justice and Affairs Council of 13-14December 2011 also again invited the Commissionto further develop “a monitoring system includingappropriate data collection and analysisto monitor the degree of integration (outcomesof policies) based on agreed common indicators,recognising the principle of subsidiarity”. This hasbeen given a follow-up by this 15-month project,launched by the Commission to analyse the relevanceand results of these indicators, in line withthe aforementioned Council conclusions.(1) http://www.consilium.europa.eu/ueDocs/cms_Data/docs/pressData/en/jha/82745.pdf#zoom=100(2) http://www.se2009.eu/polopoly_fs/1.28600!menu/standard/file/Indicators %20and %20monitoring%20of %20outcome %20of%20integration %20policies.pdf(3) http://www.consilium.europa.eu/ueDocs/cms_Data/docs/pressData/en/ec/113591.pdf(4) http://register.consilium.europa.eu/pdf/en/10/st09/st09248.en10.pdf(5) Council document 9248/10. It should be noted that all Council conclusions stress that “there is currently no unifiedview among Member States on indicators in the area of active citizenship. Member States’ views differ in relationto the different views, goals and regulatory frameworks of integration policies in the respective Member States. Thearea of active citizenships is, however, an important area of development, considering that the participation of immigrantsin the democratic process as active citizens supports their integration and enhances their sense of belonging”.(6) http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-RA-11-009/EN/KS-RA-11-009-EN.PDF(7) http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-31-10-539/EN/KS-31-10-539-EN.PDF(8) http://ec.europa.eu/home-affairs/news/intro/docs/110720/1_EN_ACT_part1_v10.pdf11

PART 1: AnalysisChapter 2:What influences migrantintegration outcomes?In the analysis report we present three differenttypes of factors that influence integration outcomes10 . First we look at the immigrant populationand individual factors. 11 Secondly we considergeneral policies and factors associated withthe macro-level structure of receiving societieswith regards to the labour market, educationsystems, social policies and the political context.Thirdly, we look at migration and integrationpolicies and factors related to specific migrationand targeted integration policies. Individual-levelfactors have been explored most extensivelyby research as large-scale surveys and administrativedata make available suitable sourcesfor analysis. In comparison, general policies andcontext as well as migration and integration policiesremain under-researched as suitable data ismore difficult to obtain and statistical analysismore demanding.Clearly, the situation of immigrants across countriesis rather different, as well as their situationin comparison to non-immigrants. However,the project’s analysis has found some evidencethat countries with better social inclusion outcomes(e.g. income) are also the countries withbetter education outcomes. In countries wherethe difference in reading performance at age15 between foreign-born immigrants and nonimmigrantsis larger, the share of foreign-bornimmigrants below the median income level ishigher. We also find that in countries where ahigher share of foreign-born immigrants havebelow median incomes compared to non-immigrants,they tend to achieve basic education alot less than non-immigrants. 12 This also meansthat there may be an underexplored overlap betweendifferent areas of integration, such as employment,education, social inclusion and activecitizenship. It is important to consider all areasrelevant for integration and analyse the links betweenthe different areas. Outcomes in one areamay very well have positive or negative effectson the other.Not one single set of factors is able to explainthe situation of immigrants or all differencesbetween immigrants and non-immigrants in EUcountries. For example, socio-economic statusand residence influence education and employmentoutcomes to a large degree. However, thesocial status cannot explain everything. Thereading performance of 15 year old foreign-bornimmigrants from the same country of origin andwith similar background varies across differentEU countries. The second generation also showlarge education gaps even though they wereborn in the country of residence and have equalresidence periods. Accounting for social statusreduces the education gap between immigrantsand non-immigrants significantly in many countries,however, differences remain. This meansthat other factors such as general policies andnational context as well as specific immigrationand integration policies shape integration outcomes.2.1 Migrant populationThe first set of factors that directly influenceintegration outcomes relate to the immigrantpopulation. This is to say that the composition ofthe immigrant population in a country will havean impact on integration outcomes. The compositionof migrant populations is shaped by manyfactors (e.g. history, geography and migrationpolicies).Migrant population factors can be distinguishedas demographic (gender, age, family status, citizenship,country of birth (first or second generation),country of birth of the parents, length ofresidence/and age of arrival), socio-economic14(10) In the context of migrant integration indicators, the term ‘outcome’ describes a (statistical) result of a certainindicator, usually measured in rates. Outcomes are compared between immigrants and non-immigrants (gaps).In general, outcomes of indicators can describe the situation of integration of immigrants in a certain area ofsociety.(11) If not indicated otherwise, the term ‘immigrant’ generally includes both first or second generation and both EUnational or Third Country Nationals.(12) For a full list of correlations between indicators, see the annex. These calculations are provided by David Reichelfrom the International Centre for Migration Policy Development.

(education, employment, income, occupation,level of development of country of origin) andsocio-cultural characteristics (mother tongue,language acquisition). 132.1.1 Which migrant populationcharacteristics influenceemployment outcomes?With few exceptions from southern EU countriesand EU-12, the first and second generationis generally less likely to be employed thanthe non-immigrant population. In general, immigrantmen achieve similar employment ratesthan non-immigrants. The overall difference canlargely be explained by low rates for women, especiallyfrom outside the EU. Highly skilled immigrantsare more likely to be unemployed thanlow skilled immigrants. Highly skilled immigrantsare also more likely to be overqualified than nonimmigrants,especially if they were born outsidethe EU.Age, residence, genderJust like for the non-immigrant population, employmentoutcomes improve with working ageand residence. Nevertheless, equal years ofresidence do not erase the disadvantage of immigrants(migrant penalty). Second-generationimmigrants with equal residence than non-immigrantsshow relatively lower outcomes on thelabour market in many EU-15 countries.Women’s labour market outcomes are usuallylower than men’s, especially when they wereborn in non-EU countries. Foreign-born immigrantwomen between 20-29 years are morelikely to be married and have children than nonimmigrantwomen in most EU countries. Theproject’s analysis shows that countries wherethere are more foreign-born immigrant householdswith one or more children, foreign-bornmigrants are also more likely to be at risk ofpoverty. Countries where fewer foreign-bornmigrant women have primary education thannon-immigrants, foreign-born immigrant womenalso have lower employment and labour marketoutcomes. This is not the case for men. Childrenhave a greater negative effect on labour marketparticipation of foreign-born immigrant womenthan on non-immigrant women. Across some EUcountries, the risk of being unemployed and notin education is higher for female children of immigrantsthan for males, with the exception ofthe Scandinavian countries. This suggests thatmales continue to be the sole breadwinner moreoften in immigrant families than in non-immigrantfamilies. When they work, foreign-born migrantwomen are twice as likely to work in lowskilled professions as migrant men according tothe OECD. This gender gap does not exist for thenon-immigrant population in many EU countries.The policy implications could be that general labourmarket policies are made more suitable tosupport younger and more recent immigrants.Targeted labour market programmes and trainingscould focus on low skilled and long-term unemployedimmigrants, in particular women withchildren. A review of the impact of family benefitsand of gender equality legislation on immigrantwomen could lead to policy adaptations.EducationLower educational levels explain a large partof the differences between employment outcomesof immigrants and non-immigrants. Thegaps reduce when accounting for educationand socio-economic position of the parents. Forsome groups, the employment rate of non-EUimmigrants is up to five times higher than fornon-immigrants. Access to and quality of earlychildhood education, avoiding social segregationin schools, allowing high permeability betweenschool tracks and supporting immigrants beforeand during the transition phase into highereducation can help to enhance the qualificationsof immigrants and break the link between thesocio-economic status of immigrants with theirparents.Whereas unemployment tends to be higher forthe low-educated for both migrants and thenative-born, differences with the non-immigrantpopulation are most pronounced for the highlyeducated. In many EU countries, low-educatedimmigrants have a higher employment rate thantheir native-born peers. This is particularly visiblein countries that have had significant loweducatedlabour migration over the past decade.In contrast, in all countries with significant immigrantpopulations the highly educated immigrantshave lower employment rates than thehighly educated native-born. This could meanthat the migrant penalty is actually increasingwith higher skill levels.(13) More national research has also focussed on social capital (contacts, networks) and cultural characteristics (religion,attitudes towards receiving society’s norms and values) as relevant factors for migrant integration.15

16Job qualityEmployment is not enough to provide equal opportunitiesfor immigrants. Narrow employmentgaps can hide other issues of immigrant integration,such as the quality of work. Work qualityis usually measured by temporary employment,low-skilled employment, part-time-employment,public sector employment and over-qualification.In addition to age, educational attainment is clearlyan important determinant for accessing higherskilled, better paid jobs. However, higher skillsalso lead to increased risk of being overqualified.On average, there is virtually no difference in thelikelihood to be overqualified between immigrantsfrom high-income countries and the native-born.Eurostat showed that the risk of over qualificationis particularly high for recent immigrants fromnon-EU countries. Their formal qualifications arethus highly discounted in the labour markets ofhigh-income EU countries. The discount is mainlyobserved for those who have obtained their qualificationsin low-income countries. In contrast, immigrantstrained in the country of residence havesimilar over-qualification rates than the nativebornand always lower than those who haveacquired their qualifications abroad according tothe OECD. Easier and more accessible recognitionprocedures, equivalence courses and Europeancooperation could facilitate the recognition ofqualifications and skills for immigrants.2.1.2 Which migrant populationcharacteristics influenceeducation outcomes?Education outcomes vary considerably acrosscountry and across different indicators. In mostEU countries, the first and second generationhave on average lower educational attainment,leave school early more often and perform worsein reading at the age of 15. The educational disadvantageis less pronounced in terms of tertiaryeducation (e.g. university).ResidenceThe OECD found that years of schooling in thecountry of residence is a relevant factor for thereading performance of migrant students at age15. First-generation students who arrived in thecountry at a younger age outperform those whoarrived when they were older. Education systemsare better able to improve student performancewhen they have a longer opportunity to shapethe learning outcomes of immigrant students.The policy implication could be that general educationpolicies accommodate recent immigrantsby providing homework and other general schoolsupport for the young, language tuition for all,equivalence classes and access to life-long learningfor adult immigrants. Targeted policies can effectivelytackle longer settled groups with lowerachievement.Socio-economic backgroundParents’ education and socio-economic positionare one of the key explanatory factors of thelower outcomes of immigrants observed in EUcountries, especially for the second generation.Bivariate analysis shows that there is a clear relationshipbetween the average socio-economicstatus of the foreign-born population and underachievementin education. Immigrants performworse in countries where the foreign-bornimmigrant population is on average poorerthan the total population. Foreign-born residentsin richer and more equal societies oftenhave a lower socio-economic status than thenative-born; by extension, their children oftenhave higher rates of underachievement. Centraland Southern European countries tend to haveforeign-born populations with a similar – if nothigher – socio-economic status compared tothe native-born and, thus, little-to-no-gap inachievement between foreign- and native-bornstudents. Socio-economic status and educationallevel of parents appear to explain almostall the educational disadvantages of childrenof immigrants from high-income countries, butonly half of the disadvantage of the childrenfrom lower-income countries. To reduce the linkand decrease socio-economic segregation inschools, various measures can be introduced,such as increasing the hours spent in school,improving the quality of teaching, delaying theage of tracking and supporting students beforeand during the transition into higher education.Smaller classes and parental involvement projectshave proven effective in many cases toboost outcomes of immigrant children.Language spoken at homeThe language spoken at home has an influenceon the education of immigrants. Students thatspeak the language of instruction at home aremuch more likely to perform better in schools.In this regard, getting parents more involvedin their children’s education has proven to bean effective strategy to improve education outcomesof children.Other socio-cultural factors that influence educationoutcomes refer to different national, religiousor ethnic backgrounds (sometimes called

18Country of origin and language skillsSeveral studies have found that immigrants fromlower developed and politically unstable countriesare more likely to naturalise. Recent analysisshowed that foreign-born immigrants from low-ormedium-developedcountries are on average fivetimes as likely to naturalise as immigrants fromhighly-developed countries. In most EU countries,people from developed (especially EU) countriestend to naturalise less because they have less incentivesto acquire the citizenship of another EUcountry. In addition, educational attainment andspeaking the country’s language at home increasesthe likelihood to naturalise for immigrants fromdeveloping countries, but has no effect on immigrantsfrom developed countries. Low-educatedimmigrants from a high-income country are morelikely than their highly educated counterparts to benationals. On the contrary, among immigrants originatingfrom a lower-income country, those thatare highly educated are more likely to be nationalsthan their low-educated counterparts.ResidenceIn most EU countries, immigrant’s (self-reported)electoral participation increases the longer theyhave settled in the country. According to the OECD,in Finland, Greece, Ireland, Portugal, Spain and theUnited Kingdom, long-term residents’ voter turnoutis more than 10 percentage points higher than theturnout of immigrants in general. Furthermore, inseveral countries – Hungary, Israel and the UnitedKingdom –participation rates for long-term residentsappear higher than those of native-born.Subjective factorsMore subjective indicators, such as ‘sense ofbelonging’, ‘interest in politics’, ‘experience ofdiscrimination’ and ‘trust in political institutions’have been mentioned in the context of activecitizenship. More research is being done on thequestion how these subjective indicators may alsoinfluence more ‘objective’ active citizenship indicators,for example, voting and membership and/orparticipation in organisations. This research is stillat the beginning. So far, there are mixed resultsabout how these indicators are related to thevarious forms of political participation.2.2 General policiesand contextThe second set of factors that influence integrationoutcomes are ‘general policies and context’.This set of factors takes into account differentnational contexts across the EU. Broadly defined,they include labour market structures and economicgrowth, the education system, the welfaresystem, the housing market, and public opinion.In general, less is known about the relative importanceof macro-level factors compared towell-researched individual level factors.Explaining macro-level factorsLabour market factors include economic growth,occupational sectors, occupational conditions,minimum wages, and labour laws. Employmentrates in relation to overall economic growth givean indication whether immigrants benefit equallyin good times or suffer disproportionately in badtimes. The employment sectors can influenceimmigrants’ employment rate as sectors that requireless qualifications and social skills (i.e. agriculturalsector) are often easier for immigrants toaccess. However, skill-demanding labour marketsand the less accessible high skilled jobs increasethe risk of over-qualification for immigrants. Theproject’s bivariate analysis suggests that foreignbornimmigrants are disproportionately affectedby part-time and temporary work. They seem tobe the first one in and out of the labour market.In this regard, employment protection legislation– a measurement of how easy firms can hire andfire workers - can have an impact on immigrantemployment. In theory, tight labour laws decreasethe propensity of companies to hire workers. Thiscould disadvantage immigrants in competitionwith non-immigrants.Education systems are very diverse across theEU. Different systems impact the performanceof immigrants. Some of the better researchedcharacteristics of education systems are accessand availability of early-childhood education, theage of ability grouping (tracking), socio-economiccomposition of schools, the diversification ofschools tracks, and public spending on education.Others may be mandatory school years, graderepetition, class sizes, number of school hoursand centralised curricula. In some cases, theseaspects have helped immigrants to overcometheir often unfavourable socio-economic backgroundand discrimination in order to achievehigher qualifications.Basic income and housing are essential humanneeds. They are regarded as prerequisites forstructural integration in society. The link betweenwelfare systems, housing and migrant integrationremains under-researched. So far, considerablymore attention has been paid to employment andeducation. Some countries measure the uptake ofsocial services of immigrants. Other research haslooked at how social transfers affect the povertyrisk of immigrants compared to non-immigrants.

Social transfers include family related benefits,housing benefits, age-related benefits, unemploymentbenefits and social assistance. The extent towhich EU governments provide these benefits isreflected in overall social spending as percentageof GDP and general welfare generosity indices.The housing market is closely related to socialissues. Are immigrants living in overcrowded accommodation,how many own property, how relevantis social housing in a country, how muchdo immigrants spend on rent relative to their income?The answers to these questions have aninfluence on the social inclusion of immigrants.Lastly, public opinion can influence integrationoutcomes. Public opinion can be considered anumbrella term reflecting the more subjectiveindicators of receiving societies. This can bemeasured through public attitudes, (awarenessof) discrimination and media discourse. In somecountries, surveys indicate high levels of antiimmigrantattitudes. Surveys and experimentalstudies have shown high degrees of discriminationagainst immigrants, especially in schoolsand at the work place. Several internationalstudies have also analysed the media coverageof immigration and integration issues. They oftenfind a negative bias of migrant integrationissues in many EU countries.2.2.1 Which context factors influenceemployment outcomes?Presumably, a positive economic situation overallwill benefit immigrants. More growth is likelyto increase migrant employment. Generally, theproject’s bivariate analysis has shown that employmentrates of foreign-born immigrants arebetter in countries where non-immigrants alsohave higher employment rates. However, whileimmigrants take advantage of economic opportunitiesin good times, the financial crisis hasshown that immigrants are the first to exit thelabour market when times are more difficult. Immigrantswere affected most by the economicdownturn in once booming new countries of immigration.In addition, regional and local differencesin the labour market influence integrationas immigrants face very different opportunitiesin different places. Immigrants often move toplaces due to existing social networks, not necessarilydue to labour market needs.Employment sectorsThe project’s bivariate analysis shows that countrieswith larger agricultural sectors tend to havenarrower gaps in female employment rates andsomewhat narrower gaps in male employmentrates. Foreign-born non-EU immigrant womenare also less active compared to non-immigrantwomen in countries with larger service sectors.This suggests that service unrelated work (agriculture,industry) is favourable for the employmentof non-EU immigrant women. On averageacross the EU, immigrants are overrepresentedin low-skilled sectors such as construction, accommodationand food services and underrepresentedin higher skilled jobs including publicsector jobs. The second generation are less likelyto work in public administration, health and socialwork or education than non-immigrants. Insome countries, these public sector jobs representa large share of the labour market, offeringstable work conditions. Large differences inemployment rates with offspring of native-bornin Belgium and Spain are partly explained bythe low share of employment in the public sectoramong native-born offspring of immigrants.Public sector employment targets and informationcampaigns can increase application ratesof eligible immigrants. This has the potential toenhance the employment situation in addition toraising public awareness of diversity.Country of origin, legal restrictions, minimum wagesThe general level of development of the country ofresidence matters. There are greater differencesin labour market participation between the generalpopulation and the non-EU-born in countrieswith higher levels of human development. Legalrestrictions to access to labour markets for familyand humanitarian migrants in some countries canhave an effect on employment outcomes.There is some evidence that minimum wages settoo high or excessively restrictive employmentprotection legislation could increase the level ofstructural unemployment and make it especiallydifficult for new arrivals to find work.Discrimination and public opinionOne context factor that has yielded more researchresults is discrimination. The most convincingstudies of the occurrence of discriminationare field experiments, which test the actualbehaviour of employers seeking to fill job vacancies.Job seekers with ‘foreign’ names have tosubmit twice as many applications to be invitedfor an interview than other job seekers with theexact same qualification. Studies on discriminationin the labour market in Belgium, Denmark,France, Germany, Italy, the Netherlands, Norway,Spain, Sweden, Switzerland, the United Kingdomand the United States – following a standardprocedure for correspondence testing developedby the International Labour Organization (ILO) in1992, showed similar tendencies.19

20Based on the projects’ bivariate analysis, we findthat countries where foreign-born immigrantshave lower integration outcomes, the awarenessof discrimination among the general populationis higher. The project’s analysis found that lowerlevels of labour market participation for non-EUbornwomen and men are associated with greaterpublic awareness of discrimination againstforeigners. However, the direction of causationis not clear. Foreign-born immigrants may havelower outcomes because they are being discriminatedagainst or, they are being discriminatedagainst because they have lower outcomes. Interestingly,the countries with higher awarenessof discrimination are often the countries withmore inclusive integration policies, in particularstrong anti-discrimination laws and extensive labourmarket integration policies. We also foundthat countries with high support for equal rightsfor legal immigrants also adopt more inclusiveintegration policies. This indicates that countrieswith greater challenges to integrate immigrants,higher perceived levels of discrimination andmore favourable attitudes towards immigrantsare also the countries that adopt more ambitiousanti-discrimination and labour market policies.Some have argued that public opinion, indeed,can be relevant for employment outcomes ofimmigrants beyond discrimination on the labourmarket. Anti-immigrant perceptions restrict government’sability to promote high-skilled labourmigration which European companies are inneed of.Discrimination awareness training for employers,anonymous job applications, strong implementationof ant-discrimination legislation andpublic anti-discrimination campaigns have thepotential to reduce the impact of discriminationand negative public opinion on employment ofimmigrants.2.2.2 Which context factors influenceeducation outcomes?The type of education system matters. Thelevel of underachievement among immigrantstudents and the general population are linked.As a general trend, the share of underachieversamong foreign-born immigrant students is higherin countries with more underachievers withinthe general population. We have also found thatthe share of the foreign-born with a universitydegree is higher in countries with more universitygraduates within the general population. Moremigrants leave school early in countries with alarger share of early school leavers within thegeneral population. This performance correlationacross all four education indicators impliesthat the general educational system is a majorfactor for the general population, including formigrants. Where the general population faresbetter, migrants generally also do better.Overall performance of education systemThe project’s bivariate analysis has found thathigher levels of spending on education as percentageof GDP is associated with higher readingscores of 15 year old immigrants. While ratesfor foreign-born immigrants might generally behigher in countries that spend more on education,the differences between foreign-born immigrantsand non-immigrants are also larger.The difference (gap) between foreign-born immigrantsand the general population is greaterin countries with greater levels of wealth andequality within the general population. There areoften more underachievers among foreign-bornimmigrant students than among the generalstudent body in countries where natives havea relatively high socio-economic position. Thisfinding suggests that a country with a wealthier,equal, and educated general population will bemore likely to have greater student achievementgaps between the general population andforeign-born immigrants. In poorer and more unequalsocieties, native students are often just asaffected by underachievement as migrant students.Even though migrants generally do betterin countries where everybody does better, thedifferences between migrants and natives appearto be larger in countries where the generalpopulation has better conditions for high performance.Generally, this analysis finds differentsituations in many North and Northwest Europeancountries in comparison to many Centraland Southern European countries.School segregationSchool segregation can be considered one relevantfactor influencing the education of immigrants.Several studies show that students withan immigrant background tend to face the doublechallenge of coming from a disadvantaged backgroundthemselves and going to a school with amore disadvantaged profile (measured by the averagesocio-economic background of a school’sinfluence) - both of which are negatively relatedwith student performance. While there may be anegative effect of ethnic segregation, it appearsthat its impact is considerably smaller than the effectof the socio-economic position of the parents.This means that the issue is not ethnic segregationof schools but socio-economic segregation.There is evidence that the outcomes of immigrant

children can be improved by decreasing socio-economicsegregation in schools, increasing the hoursspent in school, improving the quality of teaching,delaying the age of tracking and supporting studentsbefore and during the transition into highereducation. Smaller classes and parental involvementprojects have also proven to be effective inmany cases for improving immigrant children’soutcomes.TrackingThe impact of tracking - where students aregrouped in different school tracks at different agesaccording to their abilities - is very much debatedin research circles. Generally, there is a large bodyof evidence suggesting that both native and migrantstudents have, on average, higher scores incomprehensive educational systems comparedwith similar students in highly stratified educationalsystems. According to the OECD, almostall of the countries with large performance gapstend to have greater differentiation in their schoolsystems. Many studies have found evidence thatearly division of students into tracks increasesoutcome gaps over time. Some researchers havefound that students in schools with generallypoorer students do better in comprehensive systems(one-track) than in multi-tracked systems. Inshort, poorer students in schools with on averagepoorer classmates benefit most from comprehensiveschools systems. However, others contendthat some effects of the tracked education systemmight actually be due to differences in thecomposition of the migrant population.DiscriminationInstitutional discrimination is a possible explanationfor lower migrant achievement, especiallywhen other factors are insufficient to explainpersistent differences between immigrantsand non-immigrants. Discrimination may occurin terms of teacher’s decision on grade repetition,tracking and referral to special educationprogrammes. In addition, textbooks and teachingmaterials may not reflect the diversity ofinfluence’ cultural and language backgrounds.Many studies across Europe have found that immigrantsare more likely to go to a lower trackschool even when they have similar grades thantheir non-immigrants peers. This can be due todiscrimination by teachers but also the choicesof parents. Discrimination could be reducedthrough discrimination awareness training, supportfor teachers teaching second languagestudents, more objective decision making proceduresat transition periods in school careers, andmore intensive guidance.2.2.3 Which context factors influencesocial inclusion outcomes?Some evidence for most high-income EU countriesshows that generous countries with strongredistributive welfare states also have strongantipoverty policies that help alleviate materialdeprivation for both immigrants as well as nonimmigrantswithin each country. Studies haveshown that tax-benefit programs reduce childpoverty of immigrants significantly. Anotherstudy finds that family benefits have a positiveeffect on immigrants in some countries whenthey are designed to accommodate migrantfamilies, which are usually bigger than nativebornfamilies.Benefits and social spendingThe foreign-born population are more likely toreceive benefits than natives in countries wherethey are relatively poorer than natives. Countrieswith higher income gaps and higher poverty riskamong foreign-born immigrants also have higherlevels of social spending. In countries whereforeign-born immigrants are at greater risk ofpoverty and earn less income, the foreign-bornpopulation is also more likely to receive unemploymentbenefits compared to non-immigrants.These countries also spend more on social benefits.This suggests that countries that spendmore on social benefits may protect immigrantsfrom the worst; however, they do not significantlyreduce income and poverty gaps between immigrantsand non-immigrants.HousingPoor housing is often related to poverty. Theproject’s bivariate analysis shows that ‘overcrowding’is related to both income levels andpoverty risk of foreign-born immigrants. Thismeans that the foreign-born have less incomeand a higher risk of poverty in countries wherethey live in worse housing conditions comparedto the total population. The importance of housingis reflected in many of EU countries’ migrantintegration monitoring. Nevertheless, the link betweenhousing and migrant integration remainsunder-researched.DiscriminationSimilar to labour market outcomes, social inclusionoutcomes of immigrants are associatedwith awareness of discrimination. Based onthe project’s bivariate analysis, we see that thepublic perceives higher levels of discriminationagainst foreigners in countries with lower rela-21

22tive income of the foreign-born population andhigher relative poverty risk. This suggests thathigher levels of perceived discrimination againstforeigners are associated with larger incomegaps between foreign-born immigrants and native-born.Contrary to a common belief that immigrantsreceive disproportional amounts of social benefits,the outlined factors may cause an ‘underuse’of benefits for immigrants. The EU fundedstudy ‘active inclusion of immigrants’ found thatimmigrants take up less welfare than non-immigrantsin most European countries if we considersimilar social backgrounds and differentkinds of benefits. Studies in Germany have alsoshown that foreign-born immigrants benefit lessin terms of poverty reduction from the systemof redistribution (for example through taxes, andsocial insurance contributions) than natives. Therelative under-use of welfare benefits of immigrantsmight fit the hypothesis that they havelimited access to services due to linguistic, cultural,social or discrimination issues. In addition,legal barriers that influence welfare eligibilitycould limit access to poverty reducing servicesfor immigrants. More research is needed to analyseaccess, uptake and impact of social serviceson the situation of immigrants.2.2.4 Which context factors influenceactive citizenship outcomes?Generally, immigrants, regardless of their ethnicorigin, are more politically active in the countrieswhere natives are most politically active (Northand Northwest Europe). It appears that the generalpolitical environment influences active citizenshipoutcomes just as the labour market andthe education system influence employment andeducation outcomes.The impact of citizenship policies on active citizenshipoutcomes will be covered in the followingsection on migration and integration policies.2.3 Migration and integrationpoliciesThe third and last set of factors that influencesintegration outcomes are specific migration andtargeted integration policies. Migration and integrationpolicies are difficult to use as explanatoryfactors for the EU migrant integration indicatorsdue to limited comparative data. The link betweenpolicies and outcomes is often not direct, difficultto prove and interpret. Research to evaluate thelink between policies and outcomes is complex,cumbersome and costly. It requires very gooddata and advanced econometric methods forcausal evaluations. Existing evaluations arelimited to specific policies, countries and targetgroups. Although increasing in number in the EU,there are generally few comparative and rigorousimpact evaluations of policies. The project’sanalysis and findings from other research providesome first viable insights for further discussion.Migration policies regulate the inflow of immigrantsand are concerned with the question ofhow many immigrants come into the countrythrough which channel (e.g. labour migrants,family migrants, influence). Generally speakingthe evidence suggests that the size of theforeign-born population has no impact on integrationoutcomes and the difference betweenimmigrants and non-immigrants. Across all indicatorareas, the project’s analysis could notestablish a significant relation between the sizeof the foreign-born population and integrationoutcomes. However, there is a relationship betweenoutcomes and the channel of migration.Labour migrants have on average higher qualificationsthan family or humanitarian migrants.Not surprisingly, some case studies showed thatrefugees require on average more social benefitsthan the general population. Some scholarshave argued that countries such as Australiaand Canada, which have had much more selectiveimmigration policies, tend to exhibit fewerand smaller ethnic penalties than do countriesthat had major guest-worker programmes. Dueto a lack of research, the following will focusonly on integration policies. Integration policiesaddress the situation of immigrants after theyhave settled in the country.2.3.1 Which policy factors influenceemployment outcomes?Targeted labour market policies are defined aspublic interventions, which are explicitly targetedat groups of persons with difficulties in thelabour market, including immigrants. Labourmarket policies include employment services,activation measures and unemployment benefits.In countries where immigrants have loweremployment rates such measures may influencetheir outcomes. So far, there is limited comparableresearch on the effectiveness of respectiveprogrammes.There are some evaluations of specific integrationprogrammes in several EU countries, especiallywith regard to integration courses. Evaluationsgenerally show positive results in terms of improvingemployment and further education. Becausethese evaluations vary greatly in terms

of methods and focus, this case study evidenceis difficult to compare on an international level.There is a body of evidence suggesting thatcitizenship legislation has an impact on employment.Naturalised immigrants generallyhave better labour market outcomes than foreignnationals, even after controlling for otherfactors such as education, country of origin andlength of stay. Citizenship is believed to open uppublic sector jobs, reduce discrimination by employersand provide incentives to invest more ineducation and training. Employers may also takenaturalisation as sign of a positive commitmentto integrate. There are a couple of case studies,especially in Germany, France, Sweden and theUnited States that find that naturalisation has animpact on labour market outcomes, in particularof lower skilled immigrants. The findings are,however, difficult to compare across countries.2.3.2 Which integration policy factorsinfluence education outcomes?The study of education policies has expandedconsiderably in recent years. There are numerousapproaches and policies. There are alsoan increasing number of impact evaluations ofeducation policies. According to OECD literaturereviews a number of general education policieshave affected education outcomes of immigrantsin some cases: expenditure per student,hours of language instruction per week, compulsoryschool years or the age when influenceare selected for different tracks of schooling,sustained language support across grade levels;centrally developed curriculum documents;trained teachers in second language teaching;individual assessment of student needs andprogress with adequate diagnostic materials;early language interventions and parental involvementin language instruction; a focus onacademic language; integration of languageand content learning; and the valuing of mothertongues.Based on rigorous impact evaluations, we canidentify certain ‘good policies or programmes’where a positive effect has been proven in certaincircumstances for certain target groups.Based on international literature reviews ofimpact evaluations, we can find large evidencefor a positive effect of early childhood education,parental involvement programmes andclass size reductions on the education outcomes.There is modest evidence for positive impact ofpostponing the age of ability grouping in schools(tracking) and increasing teacher quality. There ismixed evidence for reducing school compositionthrough allowing parental school choice, payingteachers higher salaries, hiring teachers with amigrant background and language support programmes.There has been very little evidence yeton reducing entire school tracks and the impactof intercultural education in terms of diversity incurricula and teaching materials.Teaching quality and class sizesSeveral national studies have shown that teachingquality is one of the most important schoollevelfactors influencing student outcomes, regardlessof socio-economic and demographicfactors. While the impact of smaller classes onmainstream influence seems to be modest, asubstantive body of literature shows that classsize reductions do have a large and significanteffect on disadvantaged students, including migrants,ethnic minorities and low-income childrenwith low-educated parents. Moreover, theeffect is greatest for younger children in earliergrades, particularly from kindergarten to thirdgrade.Access to early childhood educationAnalysis of 2003 PISA data shows that participationin pre-school is strongly associated withbetter education outcomes at age 15, even whensocio-economic background is considered. However,it is very difficult to establish whether otherfactors have contributed to better outcomes andwhether these policies also work in other contexts.Greater challenges, more ambitious policiesFor example, the Migration Integration PolicyIndex (MIPEX) compares EU government’s policiestowards access, support and monitoring ofimmigrants from pre-primary to higher educationalong 22 sub-indicators. Narrower gaps,smaller immigrant populations, and lower socioeconomiclevels are associated with less inclusivemigrant education policies. Countries withgreater resources, larger numbers of immigrantstudents, and wider achievement gaps tend toadopt more inclusive migrant education policies.Indeed, migrant achievement gaps are often ajustification for changes in policy. We have observedthe same trend in other areas. Generally,the countries that have adopted ambitious targetedpolicies also seem to be facing a relativelylarger integration challenge. In this regard, theEU migrant integration indicators help to identifypolicy trends across the EU. Currently, they cannotbe used to answer, however, whether morefavourable policies are effective in increasing23

24integration outcomes according to the outlinedEU indicators.2.3.3 Which integration policyfactors influence socialinclusion outcomes?It seems that there are hardly any social policiestargeted specifically at immigrants as most socialpolicies in the field of poverty reduction applyto the whole population. There are some studiesassessing whether social benefits reduce povertymore or less for immigrants and others that investigatewhether immigrants receive more benefitsthan non-immigrants. However, for the mostpart the link between social policies and migrantintegration remains largely under-researched.2.3.4 Which integration policyfactors influence activecitizenship outcomes?Naturalisation policies have a significant effect onimmigrants’ acquisition of citizenship. Researchershave used several citizenship policy indexesand come up with very similar results. UsingMIPEX, many researchers have found positivecorrelations between countries’ policies and naturalisationrates: the more restrictive the policy, thelower the overall rate. Eurostat’s bivariate analysisfinds that naturalisation policies explain 50 %of the variation in Member States’ naturalisationrates. Other colleagues focused on one immigrantgroup (Turks) and observed that differences inpolicies explained 43 % of the variation in theirnaturalisation rates across 11 EU Member States.According to forthcoming EU funded project ACIT,immigrants from developed countries are morelikely to naturalise in countries that facilitatenaturalisation, while immigrants from developingcountries are twice as likely. The dual citizenshippolicy of countries of origin also impacts naturalisationrates. According to the ACIT analysis, immigrantswho come from countries allowing dualnationality are 88 percent more likely to naturalisein their new country of residence. Immigrantsbecome citizens and long-term residence moreoften in countries where the process is more inclusiveand where dual citizenship is accepted inboth the country of origin and destination. Shorterresidence requirements, acceptance of dual citizenship,some forms of birth-right citizenship,and support to pay naturalisation requirementsare examples than can boost active citizenship ofimmigrants.The interaction between citizenship acquisitionand integration is a complex one. The acquisitionof citizenship is not only a result of immigrants’integration, but also a status that furtherimproves their social, economic, and politicalintegration. Evidence of citizenship’s impact onintegration has been collected by researchers,including the OECD and the ACIT project. Somelongitudinal studies find that naturalised immigrantstend to obtain better-paid and higherskilledjobs, especially in the public sector. Themost vulnerable immigrant groups are mostlikely to see their economic integration improvethrough naturalisation: first generation fromlow-income countries, the second generation,and in some cases groups with low employmentrates. Citizenship can reduce real (or perceived)legal barriers and administrative costs for hiringimmigrants. It is also a signal that somebodywants to settle long-term which increases incentivesfor employers to invest in an employee.Long-term residenceFew quantitative research exists on the factorsinfluencing long-term residence, partly due tothe inavailability of data. The project’s bivariateanalysis used recently made available data onall national and EU permits. A slightly positiverelationship emerges between long-term residencepolicies and the share of long-term residents.The more inclusive the policy, the morethird-country nationals are long-term residents.The project’s bivariate analysis also found aslightly negative relationship between naturalisationpolicies and the share of long-termresidents. The more restrictive the naturalisationpolicies, the more third-country nationals arelong-term residents.Citizenship and long-term residence are only twoelements of active citizenship. Other forms ofpolitical participation of migrants such as voting,membership and/or participation in organisations,running for or holding a political office, volunteeringor participating in social movementsand protests are to be further explored to captureimmigrant’s political and civic involvement.Chapter 3:The Relevance of EUmigrant integrationindicatorsThis section of the report discusses how alreadyidentified and possible additional EU indicatorsare relevant for integration. All existing and anyproposed EU integration indicators are assessedin terms of their relevance for EU objectives,

national policies, and the research-evidencebase. National and international studies haveidentified the relationships between differentintegration outcomes and policies. This project’sresearch looked at the scientific relevance of theZaragoza indicators. These and other integrationindicators are ideally modelled on existingor proposed EU indicators and targets in variousareas of European cooperation and on existingways of calculating and collecting data.In this regard, relevant indicators are either importantfor EU objectives, measured in national monitoringreports or associated by research evidence tohave an association with migrant integration outcomesand policies. Most proposed additional indicatorscomply with the relevance criteria outlined inthe 2009 Swedish EU Presidency’s Conference Conclusions.15 Taking into consideration the discussionsat the seminars on existing examples of national integrationmonitoring and bearing in mind the definitionof integration as a two way process, this projectproposes to add a new area, namely Indicators of a‘Welcoming Society’.Regarding all indicators, improving the data tobetter measure outcomes for immigrants canand is already partly be done within the existingcooperation and resources shared by Eurostatand the National Statistical Institutes. Furtherboosting can be done through the voluntary actionof National Statistical Institutes.List of Zaragoza indicators and proposed additional indicatorsEmployment Education Social Inclusion Active Citizenship Welcoming SocietyZaragoza indicatorsProposed New IndicatorsEmployment rateHighest educationalattainmentAt-risk-of-poverty (andsocial exclusion)Naturalisation rateUnemployment rate Tertiary attainment Income Share of long-termresidenceActivity rate Early school leaving Self-reported health status(controlling for age)Share of elected representatives(research)*Self-employment Low-achievers (PISA) Property ownership Voter turnout (research)*Over-qualificationPublic sector employmentTemporary employmentPart-time employmentLong-termunemploymentShare of foreigndiplomasrecognised(survey)**Retention of internationalstudents (research)*Language skills of nonnativespeakers (LFSmodule)**Early childhood educationand care (SILC/PISA)**Participation in lifelonglearning (LFS, AES)Not in education,employment or training(LFS)Resilient students(PISA)**Concentration in lowperformingschools(PISA)**Child poverty (SILC)Self-reported unmetneed for medical care(SILC)Life expectancy (SILC)Healthy life years (SILC)Housing cost overburden(SILC)**Overcrowding (SILC)**In-work poverty-risk(SILC)Persistent poverty-risk(SILC)Participation involuntary organisations(survey)*Membership in tradeunions (survey)*Membership in politicalparties (survey)*Political activity(survey)*Perceived experience ofdiscrimination (survey)*Trust in public institutions(survey)*Sense of belonging(survey)*Public perception ofracial/ethnic discrimination(Eurobarometer)Public attitudes to politicalleader with ethnicminority background(Eurobarometer)Note: One star (*) marks indicators for which data needs to be collected or migrant sample sizes boosted. Two stars (**) markindicators for which data is not available every year (ad hoc basis). The authors of this study propose a new category ofindicators of the ‘welcoming society’. It includes the already proposed Zaragoza indicators ‘perceived discrimination’, ‘trustin public institutions’ and ‘sense of belonging’.(15) Relevance criteria: Within one of four agreed areas; long and stable for data collection; existing and comparablefor most Member States; limited in number; comparable in time; productive and cost-effective; simple to understandand easy to communicate; focused on outcome; subjective and objective25

3.1The relevance ofEU migrant employmentindicatorsThe existing migrant employment indicators arerelevant for the general EU employment targets aswell as integration policymakers and researchers.The first three core indicators – activity, unemployment,and employment rates – are stronglycorrelated to one another and measure the sameaspects of labour market participation. One ofthe EUROPE 2020 Headline targets for inclusivegrowth—the 75 % employment rate for womenand men aged 20-64—can be achieved by gettingmore people into work, including through theintegration of migrants. The unemployment rateclearly demonstrates labour market disadvantagesof the first and second generation.Two additional indicators measure relevant aspectsof the labour market: the ‘quality’ of employment(over-qualification) and type of employment(self-employment). Over-qualificationrates and gaps are relatively easy-to-interpretand relevant in nearly all EU countries, whereemployers may waste the skills and qualificationsof foreign/foreign-born men and women,especially non-EU migrants. 16 This project proposesone additional core employment indicator:public sector employment. Public sector employmentis not just an indicator of the qualityof employment. For the public sector to reflectthe public that it serves, this indicator serves asa benchmark for long-term integration, particularlyfor naturalised immigrants and the secondgeneration.LFS includes information on occupationalsector (e.g. education, health care). Aswas done by the OECD study ‘settling in’ (2012),public sector employment could be defined asPoints of discussion: Despite its overall usefulness, the activity rateis slightly harder to interpret as an integration indicator, since immigrantsare on average younger than the general population and thusmore likely to be younger studying or raising a family. This is why werecommend using this indicator with specific age groups (e.g. 25-54)to ensure better comparability. The interpretation of self-employmentis also subject to debate, as migrants may turn to self-employmentas an escape from long-term unemployment, discrimination, languagebarriers, or labour market restrictions. If any new indicator is notbased on LFS or SILC, then data must be collected, e.g. on applicationor recognition rates for foreign qualifications (LFS).people working in occupations most likely to befunded by the public sector, such as public administration,human health and social work activitiesor education.On an ad hoc basis, EU integration monitoringcould report on other employment indicators,such as those raised by international research,seminar participants and the European Commission’sDirectorate General for Employment, SocialAffairs and Inclusion. For example, part-timeand temporary employment rates can be monitoredfor immigrants as key indicators of underemployment.In addition, further dis-aggregationsby gender, education level and age groupwould demonstrate some of key determinantsof migrant employment outcomes. The retentionof international students was also raised by afew participants as a possible indicator of a favourablesituation for labour market integration,even if this statistic is used traditionally to capture‘brain drain’ and recently in some countriesto indicate highly-skilled immigration.3.2 The relevance ofEU migrant educationindicatorsThe Zaragoza migrant education indicatorscover most of the relevant Europe 2020 andET2020 benchmarks. For the second generationand for those who immigrated as children(the ‘1.5 generation’), the indicators on ‘tertiary’and ‘highest’ educational attainment show notonly their achievements, but also the areas forimprovement in the educational system. 17 ‘Earlyschool leavers’ are those people who have onlyachieved pre-primary, primary or lower secondaryeducation. The ‘low-achieving 15-year-oldsin reading, mathematics and sciences,’ an establishedET2020 benchmark, shows the share ofstudents who do not reach baseline proficiency.Within these indicators, the outcomes of thesecond generation can be a long-term ‘benchmark’for integration. A good education bringsbenefits throughout a person’s life, regardless oftheir employment status. Both first-generationimmigrants and the welcoming society want tosee improvements for the second generation asa sign of social mobility.26(16) There is more information on how to measure and interpret the over-qualification rate in the discussion paper forthe Berlin expert seminar on Employment in the context of this project. For the definition of overqualification, seeglossary attached to this document.(17) So far, data on second generation are not available from the core LFS but only in the ad-hoc modules. The nextad-hoc module is planned in 2014 (the previous one was carried out in 2008).

This project proposes that the two remainingET2020 benchmarks are relevant core indicatorsfor migrant education: participation in earlychildhood education and care and participationin lifelong learning (i.e. adult participationin any forms of education or training). As notedby the EU Council in 2011, the ET2020 benchmarkon adult learning is relevant for newcomers,who are under-represented in lifelong learning.Training helps them develop their potential,adapt to the local labour market, and improvetheir social participation. Increasing access tohigh quality early childhood education and careis also an integration priority raised by the EUcouncil in 2009 and in their national integrationpolicies.Other migrant education indicators may interestthe Commission and Member States. The share ofpeople currently “not in education, employment,or training” (NEETs) captures the opportunities fortraining for unemployed or inactive people, eitherfor young people aged 18-24 or more broadlyfor immigrant adults. Additional indicators havelimited availability of migrant-specific variables.Disaggregation by age at migration, languagespoken at home, and parents’ socio-economicstatus (including education level) are worthwhile.PISA data can be used every three years to monitordisaggregations and indicators proven to influenceoutcomes, such as the concentration ofimmigrant pupils in schools with above-averageshares of economically disadvantaged pupils and‘resilient students’ - those coming from a disadvantagedsocio-economic background but attaininghigh scores by international standards.Points of discussion: For people who immigratedas adults, the ‘highest’ and ‘tertiary’educational attainment indicatorsinclude both domestically-trained and foreign-trainedpeople. Disaggregation by ageat immigration would distinguish betweenthese two groups.For the indicator ‘participation in early childhoodeducation and care,’ a difference of afew percentage points in participation ratesbetween immigrant and native children canhave a major impact on the educational careersof those children affected. To capturethese small differences, countries wouldneed to improve EU-SILC migrant samplesizes (see this project’s data assessmentreport). Measuring language skills of nonnativespeakers would require several indicatorsin an EU-funded targeted survey orad hoc module.3.3 The relevance of EUmigrants’ social inclusionindicatorsImmigrants are a critical target group for theEU’s overall strategy on social inclusion andfighting poverty. As an overarching concept, socialinclusion can be a priority for national integrationpolicies and research. A basic income,housing, and good health are related to otherareas of integration and may be pre-conditionsfor immigrants’ participation in society.In most EU countries, the general population hasgenerally higher incomes and a lower risk-of poverty-or-social-exclusionthan the foreign-born, especiallynon-EU newcomers. Income is measuredby the median annual equalised disposable income.The EU’s overall target in this area is therisk of poverty or social exclusion. This compositeindicator combines the share of those at riskof poverty, severely materially deprived or livingin households with very low work intensity.The EU integration indicators could include morekey EU indicators on social inclusion and socialprotection, which are relatively easy-to-calculateusing EU-SILC. Europe 2020 has a major focus onchild poverty—across the EU, the risk-of-povertyfor the children of foreign-born parent(s) is twoto-fivetimes greater than for the children of thenative-born. Additional indicators for discussion arein-work poverty-risk and persistent poverty-risk.‘In-work at-risk-of-poverty’ monitors whether employmentis a sufficient protection against poverty,since indeed immigrants are more likely to be inlow-paying, temporary, or part-time jobs. Persistentat-risk-of-poverty provides insights intowhether labour market activation policies are appropriatefor long-term unemployed immigrants.On health, self-reported health status can bereported at aggregate level and after controllingfor age and gender. Life expectancy, healthylife years, and self-reported unmet need formedical care (see annex) are key indicators forthe EU’s OMC on Social Inclusion and Social Protection.They can be calculated based on availabledemographic data primarily through SILC,subject to reliability tests. Using self-reportedhealth status and needs as well as life expectancyand healthy life years provides a usefulcombination of commonly used subjective andobjective measures for further investigation ofthe health situation of immigrants in Europe.On housing, foreigners and foreign-born people oftenlive in insecure and overcrowded housing andface greater housing costs. Property ownership is27

an indicator of immigrants’ long-term settlementin the country as well as a protection from discriminationon the rental market. In addition, overcrowdingand housing cost overburden are twokey general EU social inclusion indicators relevantfor immigrants across Europe. Nearly 1 in 4 peoplein deprived or overcrowded housing in OECD countrieslive in an immigrant household. The housingcost overburden rate allows policymakers to assesshow housing costs affects immigrants’ poverty andquality of life.Points of discussion: Further disaggregation of the social inclusionindicators would be useful to identify more vulnerable groups, such ashouseholds with children, the elderly or long-term unemployed. Still,users should keep in mind that social exclusion cannot be fully capturedthrough statistics, especially for the hardest-to-reach.may naturalisation help immigrants become moreintegrated. But also more integrated immigrantsmay be more likely to naturalise. Active citizenshipindicators like naturalisation can therefore be seenas both a final step in a process and as a tool tofurther improve integration in several areas of life.The share of immigrants who acquired permanentor long-term residence is a relevant outcomemeasure for long-term residence. Long-termresidence can now be measured as both the EUlong-term residence permit (2003/109) and anyother type of national long-term residence permit.The resulting indicator describes how common oruncommon it is for non-EU residents to have longtermresidence and, by extension, the same socioeconomicrights and responsibilities as nationals.283.4 The relevance of EUmigrants’ active citizenshipindicatorsThe EU has a broad policy agenda on active citizenshipinvolving many institutions. Active citizenshipis about the acquisition and the exercise ofequal rights and responsibilities for immigrantsand citizens. When immigrants take up and useequal rights and responsibilities, they send a strongsignal to themselves and others about their senseof belonging in the country. Beyond this symbolicvalue, this process can improve immigrants’ social,economic, and political participation, the public’sperceptions of immigrants, and the democraticlegitimacy of the state. On several occasions, ministersresponsible for integration have agreed thatimmigrants should have the opportunity to naturalise,become long-term residents, and participatein the democratic process because these achievementssupport their integration and enhance theirsense of belonging. Each of the EU’s migrant activecitizenship indicators is also supported by specificstandards on active citizenship that Member Stateshave agreed together at EU level.Indicators on naturalisation, long-term residence,and civic participation are just as relevant as theother integration indicators because they also capturethe different national contexts. For example,naturalisation is not simply the results of citizenshippolicies, but also the differences in immigrantpopulations and other policies in the country ofresidence and origin. The country of origin, durationof residence, and socio-economic participation areall related to naturalisation. Futhermore, naturalisedimmigrants generally have better integrationoutcomes than non-naturalised immigrants, ofteneven after controlling for other factors. Not onlyThe share of immigrants who acquired citizenshipis a long-standing indicator in nationaland international research on immigrant integration.Naturalisation is a reliable and meaningfulmeasure of the outcomes of policies and of otherkey contextual factors, such as immigrants’motivation to naturalise, duration of residence,and settlement in the country. This indicatoropens an important debate about the importanceof these policies and other factors. Multiplemeasures of naturalisation complete thepicture of citizenship acquisition. ‘The share ofnaturalised immigrants’ can measure on an adhoc basis how many immigrants have becomecitizens over time (See Annex). A third measurecould be an estimate of the ‘naturalisation ofeligible immigrants’, calculated as either a shareor rate and based on countries’ ordinary requirementfor years of residence.Points of discussion: The 2010 ZaragozaDeclaration observed that ‘there is currentlyno unified view among Member States onindicators in the area of active citizenship’.It explained that governments have differentinterpretations of active citizenship,depending on their political views, goals,and regulatory frameworks for integrationpolicies. This project shows how active citizenshipindicators can be used to measurethe acquisition and use of rights, as both ameans and an ends for successful integration.Analysis of these indicators thus providesgovernments and stakeholders an evidence-baseto debate their different views,goals, and policies. This debate reveals thegeneral need for greater research on boththe effects of integration outcomes, includingactive citizenship, on other areas of integrationas well as the links between theirpolicy objectives and their policy outcomes.

Civic and political participationThis project proposes that the active citizenshipindicators measure not only the acquisition butalso the exercise of rights and responsibilities,namely civic and political participation. Beyondthe existing additional indicators on the shareof immigrants among elected representativesand voter turnout among eligible immigrants,the standard indicators of civic and political participationare membership in voluntary organisations,membership in trade unions, membershipin political parties, and political activities(e.g. contacting politicians, petitions, boycotts,demonstrations). These indicators would illustratethe process of civic and political participationbefore and after naturalisation. In particular,disaggregation for naturalised citizens and foreignerswould capture the levels of participationfor immigrants who cannot or do not wantto naturalise, since citizenship is not a requirementfor all forms of democratic participation.Immigrants’ participation can currently only bemeasured internationally by pooling data overlong periods of time. To capture this data, theEU can invest in a targeted survey of immigrantsor boosting immigrant samples in the EuropeanSocial Survey or EU-SILC’s ad hoc module on socialparticipation. Methodologies are also availableto measure immigrants’ sense of belongingand representation in political and decisionmakingbodies. Both aspects were highlighted asa potentially complex to capture but significantcomplementary element to analyse integrationin this area.3.5 The relevance of EUindicators of a welcomingsocietyThis project proposes for discussion indicators tomeasure how the receiving society plays a rolefor migrant integration. The way that the ‘receivingsociety’ perceives integration can have a seriousimpact on the how immigrants integrate intosociety. These more ‘subjective’ measurementscan be used complementary to more ‘objective’integration outcome indicators. In every expertseminar, participants asked for indicators thatmeasure integration as the two-way process ofmutual accommodation, as stated by the first EUCommon Basic Principle. Many national monitorsalready use such indicators. Alternatively, theseindicators could be part of the existing four areas.These indicators capture the cross-cutting issuesof discrimination and the subjective attitudesof the general public and of immigrants themselves.Similar indicators crop up in integrationmonitoring at local, national, and internationallevel (e.g. OECD, ILO). The 2010 Zaragoza Declarationhas already named additional indicatorsthat are relevant as indicators of a welcomingsociety: experiences of discrimination; trust inpublic institutions; and sense of belonging. Usingthese ‘subjective’ indicators, initial Europeanresearch has found that the ‘sense of belonging’and ‘trust in political institutions’ amongthe foreign-born is related to other integrationoutcomes and tends to converge with those ofnatives over time. Moreover, a significant bodyof national and international research exists onmeasuring discrimination, whether through theILO’s situation testing or minorities’ own perceptions(e.g. EU-MIDIS study). This subjective datacan be collected through a targeted immigrantsurvey or boosting immigrant samples in existingEuropean surveys. It is less reliable to poolexisting survey data over a period of, for example,more than three years; however, this can bea cost-effective, short-term option.So far, none of the existing EU indicators directlyaddress the general public. The Eurobarometerasks the public about their awareness of discriminationtowards immigrants. The same publicquestions on discrimination are conductedannually by Eurostat’s Eurobarometer service.This project proposes to use two key indicators– public perception of ethnic discriminationin the country and public attitudes towards apolitical leader with an ethnic minority background.Around half of EU citizens (56 %) stillthink that ethnic discrimination is widespread intheir country – and more so than other groundsof discrimination. Beneath this EU average lays awide gulf in public perceptions of ethnic discriminationin different EU Member States. In addition,Europeans are still only slightly comfortable withthe idea that someday their country could be ledby a person with an ethnic minority background.On a scale from 0-10, the average European puttheir comfort level with a female president orprime minister at a 8,6. For an ethnic minoritycandidate, they gave a 6,5. The average Europeanwould only feel more uncomfortable with someoneunder 30, over 75, or transgendered or transsexual.While ethnic discrimination and ethnic minoritiescover wider groups than persons with animmigrant background, these public attitudes area helpful barometer for specific attitudes towardsdiverse immigrant groups, particularly those withdifferent racial and ethnic backgrounds. As notedearlier in this report, the public’s openness andawareness of discrimination tends to be greaterin countries with significant gaps in integrationoutcomes and more inclusive integration policies.29

Indicators of the welcoming society could be usedto facilitate how public opinion and perception ofdiscrimination influence integration outcomes - alink which receives growing attention in research.Chapter 4:Migrant integrationand the Europe 2020The Europe 2020 Strategy is the European Union’sten-year growth strategy. It sets out overall targetsin the area of employment, education, environment,poverty and social exclusion. All MemberStates have committed to achieving Europe 2020targets and have translated them into nationaltargets and policies. The targets are coordinatedthrough monitoring, coordination and reporting inthe framework of the European Semester, a yearlycycle of economic policy coordination. The Europe2020 targets are relevant for immigrant integration,namely employment, early school leaving, tertiaryeducation, poverty or social exclusion.However, migrant integration has so far not sufficientlybeen mainstreamed into main EU policyareas despite the fact that there are wellfunctioningpolicy mechanisms in place. Only afew indicators used in these mechanisms havebeen disaggregated for immigrants even whendata allows for such break-downs. The outcomesof the Analysis Report can be used to incorporateimmigrant integration into the monitoring andtarget setting of established policy frameworks inemployment, education and social inclusion.Closing the gapThis report outlines relevant factors that influencemigrant integration to inform integrationpolicies at various levels of governance. Integrationpolicy often aims at closing the gap betweenimmigrants and non-immigrants in theEU. Providing equal opportunities for immigrantsand delivering specific support are importantbecause immigrants are a particular vulnerablegroup. Furthermore, society as a whole benefitsfrom closing the gaps between people with andwithout an immigrant background, in particularwhere and when immigrants are or become alarge part of the population.Between 10-15 % of the total population in EU-15 countries were foreign-born in 2010. Immigrantsare particularly overrepresented amongyounger age groups in many countries. Accordingto a Eurostat working paper on demographicprojections of the foreign-born population in EUcountries from 2010, the share of foreign-bornis likely to more than double by 2061. The mostconservative projection estimates that 26.5 % ofthe EU population would have a ‘foreign background’by 2061. By 2061, at least every thirdperson in Austria, Belgium, Cyprus, Germany,Greece, Luxemburg, Portugal, Spain, Sweden,and the United Kingdom is estimated to have aforeign background. As the relative importanceof immigrants in our society increases, the outcomesof migrant integration become more relevantfor where the general society is heading.If policymakers in a Member State want to reducethe number of students leaving schoolswithout a degree, they have to take into accountthe particular challenges of immigrants as theyrepresent a large share of new children in schoolin many countries. They then may wish to comparethis group with their peers without an immigrantbackground and control for gender andsocio-economic background.To illustrate the impact of effectively integratingimmigrants into the EU agenda, the projecthas calculated a ‘closing the gap-scenario’ usingseveral indicators as examples. The ‘closing thegap- scenario’ assumes equal outcomes of themigrant population in comparison with the totalpopulation. Based on this hypothetical scenario,we show the potential impact of complete convergenceof outcomes on overall improvementand on meeting the respective Europe 2020 targets(see annex for full list). 18Currently, the total employment rate in the EU is69 %. The employment rate for the foreign-bornis 64 %. The Europe 2020 target is to increase theoverall rate to 75 %. Closing the employment gapfor foreign-born immigrants accounts for 10.7 %of meeting the Europe 2020 target across all EUcountries for which targets and data are available.Given the ‘no gap scenario’, Austria, Germany, theNetherlands, and Sweden would half-way meettheir national Europe 2020 target (see annex).Member States could prevent half a million peoplefrom leaving school early, if they could close30(18) Our calculations are based on Eurostat data which is available online. We take the rates and population sizesfor 2010. The Europe 2020 targets are based on the National Reform Programmes of April 2011. Data was notavailable for Romania and Slovakia in most cases. These calculations could be done for a comparison betweenthe native born and foreign –born population.

the gap for migrants. This accounts for 8.7 % of allearly school leavers in the EU. The EU as a wholewould be 30 % closer its headline target of reducingthe early school leaver rate from 14 to 10 %.The ‘no gap scenario’ accounts for more than 50 %of reaching the target in Belgium, Cyprus, Denmark,Germany, Greece, and Italy. In fact, Sweden wouldexceed its national education target (see Annex).23 % of the EU population is at risk of povertyor social exclusion. The rate is 9 % higher for theforeign-born population (32 %). If this gap wereclosed, the EU could lift 3.3 million immigrantsout of poverty or social exclusion. This number accountsfor 5 % of all people at risk of poverty orsocial exclusion in the EU. This stands for 17 % ofall people at risk of poverty or social exclusion inAustria, 19 % in Belgium, and almost 19 % in Sweden.Closing the gap for immigrants would bringthe whole of the EU 16.2 % closer to reaching itsheadline poverty target. The migrant gap representsmore than 50 % of the national targets inAustria, Belgium, Greece, and the Netherlands.Of course, the ‘no gap scenario’ is unlikely in theshort run. It is based on targets and populationstatistics that are subject to constant changes.However, this exercise is useful to emphasise thatimmigrants play a significant role for Europe inreaching its overall targets. Mainstreaming migrantintegration into established monitoring andtarget setting mechanism at EU level is crucial toaccount for the (increasing) relative importance ofthe migrant population in many EU countries.The ‘closing the gap – scenario’ for the Europe 2020 headlineRate of the totalpopulation, %2010Rates of theforeign-born, %2010Europe 2020Target (2011)The number ofpeople lifted outof poverty riskor social exclusiongiven the'no migrant gapscenario'% of all peoplethat would belifted out of povertyrisk or socialexclusion giventhe ‘no migrantgap scenario’Share of 'nomigrant gap'of reaching theEurope 2020targets, in %EU25 (w/o RO, SL) 23 32 20000000 3249117 49 16.2Belgium 20 40 380000 225665 19.6 59.4Bulgaria 37 45 260000 3654 1.9 1.4Czech Republic 14 24 Remain the same 34488 3.7Denmark 19 39 22000 low work 76707 12.2intensityGermany 21 28 33000 long-term 598311 5.8unemployedEstonia 22 26 Only risk of poverty 57493.2Ireland 29 31 186000 9213 1.2 5.0Greece 28 51 450000 230768 12.0 51.3Spain 25 36 1450000 564851 7.7 39.0France 20 32 Only risk of poverty 629175 8.5Italy 24 34 2200000 391679 4.4 17.8Cyprus 20 31 27000 11718 11.4 43.4Latvia 37 40 121000 7972 1.5 6.6Lithuania 34 37 170000 3710 0.5 2.2Luxembourg 18 22 x 5601 10.2Hungary 30 26 450000 -13844 -0.7 -3.1Malta 19 23 6560 928 1.9 14.2Netherlands 16 28 100000 181759 11.2 181.8Austria 16 30 235000 139409 17.0 59.3Poland 28 27 1500000 -409 0.0 0.0Portugal 24 24 200000 -2638 -0.2 -1.3Romania 40 : 580000Slovenia 18 27 40000 17916 7.5 44.8Slovakia 20 27 170000 0Finland 17 40 150000 42716 7.9 28.5Sweden 14 28 x 142747 18.7United Kingdom 21 28 x 379597 4.931

PART 2: Data assessment 19Chapter 5:Availability of indicatorsAvailability of indicators primarily depends onthe sample sizes (i.e. number of observations) ofimmigrants in the data sources, because the majorityof indicators are based on sample surveys.The quality criteria for publication depend on thedata source of each of the indicators. 20The data source for migrant integration indicatorsin the area of employment is the EuropeanUnion Labour Force Survey (LFS). For the annualdataset of the EU LFS publishing guidelines areset by Eurostat, which depend on the estimatedfigures based on the sample (i.e. the weightedfigures). The thresholds were set in order toavoid publication of figures based on unreliabledata (i.e. too few observations) and to maintainconfidentiality. The reliability limits are consideredhigh enough given that the limits werechosen based on sample sizes (for each countryseparately). 21The data source for the social inclusion area isthe European Union Statistics on Income andLiving Conditions (SILC). Data from the EU-SILCare published when an estimate is based on50 or more sample observations and the itemnon-response does not exceed 20 percent. Estimatesbased on 20 to 49 sample observationsor non-response between 20 and 50 percent areflagged as unreliable. Estimates based on fewerthan 20 sample observations or non-responseexceeding 50 percent are not published.The data source for one of the education indicatorsis the Programme for International StudentAssessment (PISA). 22 PISA data are not publishedif the estimates are based on fewer than 30 studentsor less than five schools.Due to the limitations of smaller sample sizesfor immigrants, it is mostly countries with lownumbers of immigrants where data are missingaccording to Eurostat’s publishing guidelines.Major problems with availability of the majorityof indicators concern the countries Bulgaria,Hungary, Malta, Lithuania, Poland, Romania andSlovakia. For illustration of availability of indicatorsfor third country nationals aged 25 to 54,see Figure 5 in the annex.The only indicator defined in the Zaragoza Declarationthat could not be produced due to missingdata sources is the indicator on political representationof immigrants (‘immigrants amongelected representatives’). No international datacollection on this indicator is available at themoment.A potential source for data on political participationof immigrants is the European Social Survey(ESS), which however does not qualify for annualmonitoring sample sizes of immigrants remaintoo small. Several waves of the survey need tobe pooled in order to obtain large enough samples(see discussion below). Pooling can be acost-effective short-term solution to make availablemore data to measure immigrant integrationwhile it has to be kept in mind that somequality of data is lost in the process. The bestavailable indicators for all groups are the activityand employment rates as well as the ratesof persons holding long term residence status(third country nationals only) and acquisitionsof citizenship. All other indicators are in principleavailable for all groups (foreign population orforeign-born) most countries, but availability isreduced when further breakdowns are made (e.g.by gender, age groups, born in the EU vs. bornoutside of the EU, EU-citizens vs. third-countrynationals). The best availability of indicators isachieved for the foreign-born population as therelevant sample size are higher than for foreignersin most cases.32(19) This section is drafted by Albert Kraler and David Reichel (of the ICMPD). The authors of the report are responsiblefor the final version.(20) Cf. Eurostat 2011: Indicators of Immigrant Integration. A Pilot Study. 2011 edition. Eurostat Methodologies andWorking papers. Luxembourg, pp. 18-22.(21) Contrary to the EU SILC, there is not the same minimum number of observations set, which is needed for publishingresults. But the underlying minimum sample sizes appear similar to those of the EU-SILC, as described inthe following.(22) Other potential data sources for education indicators are the TIMMS and PIRLS survey. However, these were notconsidered in this study because too few EU countries have participated in these surveys and the migrant sampleis generally too small for comparative analysis. PIRLS and TIMMS also focus on students in the same graderather than the same age which makes the comparison across countries more difficult.

Chapter 6:Quality of data sourcesThe data sources used for the production of theindicators include the EU-Labour Force Survey(EU LFS), the EU Statistics on Income and LivingConditions (EU-SILC), OECD’s Programme for InternationalStudents Assessment (PISA) as well as Eurostat’smigration statistics. This section highlightsthe main issues related to the quality of the data,including representativeness of data, reliabilityof data and comparability across countries. For amore detailed description and analysis of the data,reference is made to the comprehensive assessmentreport (Data Assessment Background Report).6.1 The LFS and EU-SILCThe main surveys, EU LFS and EU-SILC, are EUwidesurveys, which are implemented by nationalstatistical institutes (NSIs). EU LFS is the largestand the EU-SILC the second largest populationsurvey in Europe. Because of their large samplesthe surveys can be used for monitoring issuesrelated to immigrants. However, since the samplesrepresent the total population, the numbersof immigrants included in the samples are lower.Particularly in countries with low numbers/ percentagesof immigrants, sample sizes of immigrantsare naturally lower as well. The qualitycriteria employed by Eurostat are based on goodexperience with data collection. The quality criteriadefine minimum numbers of observations (andresponse rates), which need to be met before dataare published. Data based on observations belowcertain thresholds are either flagged as being oflimited reliability or not published. The reliabilitylimits are deemed sufficient. The same applies toPISA which is an internationally organised surveyand therefore highly comparable across countries.The LFS as well as the other surveys were not designedas a tool to monitor migrant integration andthus immigrants are captured to a varying extentin EU Member States. Some countries have madeincreasing efforts to capture immigrants by usingmulti-lingual questionnaires and interviewers. Somecountries also include the number of immigrants (ormostly non-nationals) in the total population whenproducing weights accounting for non-response.The overall sample sizes in the LFS yearly dataset2009 range from 12,370 in Island and just below20,000 in Luxembourg and Estonia to a maximumof almost 660,000 in Italy. The average sample sizeprovided is 150,000. The sample numbers of foreign-bornpersons range from 94 in Bulgaria to over42,000 in Italy and 44,260 in France. The averagenumber of foreign-born interviewed in the yearlydataset is 11,100. When restricting the samplesto foreign-born persons aged 25 to 54 the samplesizes decrease to an average of below 6,500 (rangingfrom 36 to 28,000). Figure 1, below, shows thesample sizes of the annual LFS dataset 2009.Figure 1: Sample sizes of different migrant population sub-groups in LFS2009 yearly datasetSource: EU-LFS, own calculations, Germany missing due to denied access to micro-data33

EU SILC is an international comparative surveywhich has been methodologically well developedinvolving much expertise and methodological research.It is a good source for providing indicatorson immigrant integration.The sample sizes required for each participatingcountry are defined by EU regulations. The totalsamples of the EU SILC 2009 cross-sectionaldata range from a minimum of 6,500 in Islandor 7,500 in Cyprus to a maximum of 43,000 inItaly. Naturally, the numbers of foreign-born ismuch smaller in the samples ranging from 11in Romania to 4,233 in Luxembourg (mean of1,200). Focusing only on the age group 25 to54 the samples of foreign-born range from 4 to3040 (mean of 715).Figure 2: Sample sizes of foreign-born in EU SILC 2009 cross-sectional datafile, total and age group 25 to 54Source: Own calculation, EU SILC 2009, Germany missing due denied access to micro-dataThus, most countries show modest sample sizesfor the foreign-born population. Romania andBulgaria as well as Poland clearly do not includeenough immigrants to present reliable estimateson certain key indicators when data is brokendown by gender.6.2 PISA 23PISA achieves a good coverage of its target population(15 year-old students) and no major impactof non-response is assumed. Countries withinsufficient sample sizes for immigrants are excluded.Very recent immigrants without sufficientlanguage skills can be excluded, but immigrantswho are taught in the assessment language justlittle longer than a year should be kept in thesample. It is thus not surprising that there areimmigrants included which have more difficultiesin solving the tasks requested by PISA dueto lack of language skills.Figure 3 shows sample sizes in the PISA 2009 dataof foreign-born students. There are five countrieswith samples below 60, which leads to unavailabilityof estimates differentiated by gender.34(23) Other potential data sources for education indicators are the TIMMS and PIRLS survey. However, these were notconsidered in this study because too few EU countries have participated in these surveys and the migrant sampleis generally too small for comparative analysis. PIRLS and TIMMS also focus on students in the same graderather than the same age which makes the comparison across countries more difficult.

Figure 3: Sample sizes of foreign-born in PISA 2009 in EU-27, Accessionand Candidate CountriesSource: PISA 2009For future use of the LFS and EU-SILC as a sourcefor the European indicators of migrant integration,efforts should continue to increasingly captureimmigrants in all EU countries. The discussionof harmonised ways to produce weights (basedon country of birth, citizenship or not) should becontinued as well. Clearly, more methodologicalresearch is needed on this issue in the future.Problems with data quality – especially samplesizes – can be further addressed by presentinguncertainty in the results (i.e. confidence intervals).Estimating confidence intervals for immigrantshas to be made in a consistent manner byNational Statistical Institutes (NSIs).All data sources are good representations of thetotal population. There are some known problemswith coverage of immigrants. Immigrants sometimesshow higher non-response rates and aresometimes not well covered in the samples (especiallyrecent immigrants and immigrants with lowersocial background or poor language knowledge).Countries have started addressing these problemswith improvements expected in the future.Non-response limits the quality of sample surveyssince they bias the sample for certain characteristicsand they increase the variance of the resultsof the survey. The weighting of results reducesbias, but only for those characteristics which canexplain non-response. All countries are requiredto weight the sample of the LFS based on certaincharacteristics known for the total population.Most countries weight the LFS by regional distribution,age and sex. Additionally, eight countries(DK, DE, EE, ES, IT, LU, AT, CH) used informationon nationality for weighting the results of the LFS2009. The Netherlands used information on ethnicbackground for weighting .24Variations in the degree of comparability areunavoidable in international settings due to differencesin data collection methods, definitionsof the target population or specific quality issues.However, in terms of definitions used forconcepts, the LFS and EU-SILC can be consideredthe best available sources providing harmoniseddefinitions based on international standards definedby international organisations and regulatedby EU legislation. Nevertheless, there isstill room for improvement as national questionnairesstill differ, for example, in regard to keyvariables such as the labour status (cf. Eurostat2011c: 27-28). Lack of comparability over timeis another issue. Due to the recent introductionof the indicators of migrant integration, changesover time do not yet play a major role for thereliability of the results. However, comparabilityover time should be looked at in the future givenchanges of data collection in the 1980s and, especiallyin recent years in some Member States.Finally, there is a lot of potential for more researchand analysis on migrant integration byusing data from the EU LFS and EU-SILC (alsoPISA), which should be exhausted in order tofurther examine the meaning of indicators and(24) Eurostat (2011): Labour force survey in the EU, candidate and EFTA countries. Main characteristics of nationalsurveys, 2009. Eurostat Methodologies and Working papers.35

further the knowledge on issues related to immigrants’integration in Europe. For instance, thelongitudinal component of EU-SILC could be usedfor measuring integration processes over fouryears in countries with large enough samples.Contrary to the data in the areas of employment,education and social inclusion, data on citizenshipacquisitions and long term residence status(area active citizenship) are based on nationaladministrative data. Therefore, the data aremuch more reliable and generally good coverageis achieved. Differences in comparability stemfrom differences in national legislations (whichare sometimes difficult to harmonise acrosscountries) and differences in capturing the foreignpopulation by migration statistics. Eurostathas considerably improved data collection on theforeign population through the implementationof the EC Regulation 862/2007 on Communitystatistics on migration and international protectionin the past years.Chapter 7:Robustness of indicatorsRobustness refers to the question whether or notdata are prone to outliers (e.g. extreme cases)and therefore unreliable. Analysing the datapublished by Eurostat shows that there are a fewspecial cases of outliers. For instance, the gap inthe employment rates between the foreign-bornpopulation and the total population in Romaniais much higher than the gaps in other countries,which might be explained by limited reliabilityof the data (as indicated by Eurostat). The gapsin the percentages of persons holding a certainlevel of education differ significantly acrosscountries. This is presumably due to the differentcharacteristics of immigrant groups in the countriesand is therefore not a problem of unreliabilityof the estimates. Here it is important tohighlight the difference between data reliabilityand proper interpretation. The consistency of theresults can be studied in Figure 4, below.Figure 4: Results of indicators on migrant integration foreign-born,aged 25 to 54, 200936Notes: Median Income was subtracted by 100. Source: ICMPD calculations based on data from EurostatPilot Study 2011

Within the four areas, a high degree of internalconsistency is observed, which means that withineach of the four areas, all indicators are statisticallyrelated to some extent. Among the employmentindicators for men only, there is someinconsistency observable. Some indicators showa few outliers which partly stem from data withlimited reliability. Data are more consistent whenbroken down by gender. The indicators on healthstatus, gaps in educational levels, the gap inself-employed persons as well as the long-termresident status rate show strong variations. Thereasons for the variations can be also explainedby differences in the situations across countriesand must not be a result of unreliability in thedata. Explanations for such strong variationacross countries need further investigation.In spite of missing data from some countries andmissing data for some indicators the overall situationrelated to data availability and quality is positively.There are enough reliable and robust datafor making international comparisons across theMember States of the European Union (EU). TheLFS and PISA data have been successfully exploredfor the use of migrant integration. Current changesto SILC data expect to provide new opportunitiesfor using SILC data to measure migrant integration.Chapter 8:Alternative datasourcesThe data sources used for the indicators onmigrant integration are without doubt the bestavailable sources for the purpose of monitoringimmigrant integration throughout the EU and forproducing the indicators defined in the ZaragozaDeclaration. There are no other internationallycomparable data sources as good as the EU LFSand EU SILC for providing annual data coveringmigrants in comparison to the total population.Other international – generally high quality surveys– are, among others, the European SocialSurvey (ESS), the Eurobarometer, the World ValuesSurvey, the European Values Survey, the EuropeanElection Study and the International SocialSurvey Programme. Those surveys – in spite oftheir generally good quality – do not provide sufficientlylarge samples of immigrants for annualor biennial monitoring of migrant integration.It is generally not recommended to use alternativesources from national data sources, as forinstance from existing register data or nationalspecific surveys with larger samples. The use ofnational data sources is not recommended becausethe data are mostly not comparable withthe data from other countries due to deviatingdefinitions and coverage of the target population.Since there is no tool to systematicallycheck comparability, there is no better solutionthan to stick to internationally organised andcross-checked data collections (like the onesused). Data from national data sources, oftendrawn from national registers, are usually ofhigher quality than from internationally harmoniseddata sources. Although it is obvious thatnational data sources allow for more detailedanalysis, the advantage of comparability of internationalsurveys is necessary for the purposeof EU-wide integration monitoring.Chapter 9:Availability of datasources for additionalindicatorsThere are various potential additional indicatorsfor monitoring the integration of immigrants inthe European Union. Eurostat already producedresults for additional indicators on overqualification,self-employment and risk-of-povertyand social inclusion. According to the assessmentof the project coordinator (Migration PolicyGroup), further potential additional indicators inthe area of employment were identified: temporaryemployment, part-time employment, longtermunemployment and public sector employment.These indicators are available from theEU Labour Force Survey and are thus regularlyproduced in all EU Member States. Potentialadditional indicators in the area of social inclusioninclude in-work at-risk-of-poverty, at-riskof-povertyrate before social transfers (exceptpensions), persistent at-risk-of-poverty, housingcost overburden rate, as well as other indicatorsrelated to health, such as unmet health needs.All these indicators are available for most of EUcountries from EU-SILC data.The potential additional indicators in the areaactive citizenship are related to civic and politicalparticipation, such as membership in organisations,membership in trade unions, membershipin political parties, volunteering or contact withdecision-makers. These indicators are measuredin the European Social Survey; however, samplesizes are too small for annual use. Even pooleddata for two or three years may not be enough37

to compare immigrants across many EU countries,with regard to one specific indicator. Moreover,the importance of measuring discriminationin the framework of integration monitoringsystems was highlighted by participants at theexpert seminars organised in the framework ofthis project and the project coordinators. Besidesexperiences of discrimination, data on awarenessof discrimination in society and awarenessof rights of discrimination legislation / equalitybodies are deemed useful to monitor discriminationagainst immigrants. 25In addition, other indicators identified in theZaragoza Declaration are language skills, experiencesof discrimination, trust in public institutions,voter turnout among the population entitledto vote, and sense of belonging. All of thesepotential additional indicators are not availablefor immigrants from the main international surveys.Major European opinion surveys collectdata on these indicators (except language skills),but the sample sizes are not large enough tomonitor integration of immigrants on an annualor biennial basis.However, that is not to say that monitoring suchindicators is not possible, but currently only to alimited extent. General improvements and availabilityof additional indicators can be achievedeither via increasing samples of existing surveysor via targeted data collections, such as the inclusionof targeted modules in existing surveys.The most promising source for political participationand discrimination is the European SocialSurvey (ESS) 26 which has been conductedfive times since 2002. The survey allows oneoff evaluations on the additional indicators bymerging data from several waves and, thus,making statements on longer periods of time. 27If the ESS or other opinion surveys are to be usedfor monitoring immigrant integration, a solutionwould be to boost samples of immigrants in orderto obtain more reliable results. Current samplesizes of immigrants in the ESS range fromaround 20 up to over 300 per country.38(25) For a discussion on measuring discrimination against immigrants see: OECD (2012): Settling In. OECD Indicators ofImmigrant Integration. http://www.oecd.org/migration/integrationindicators/ or Fundamental Rights Agency (2011):Migrants, Minorities and Employment – Exclusion and discrimination in the 27 Member States of the EuropeanUnion (Update 2003-2008). Section 4.(26) Another source for political participation could be the European Election Study (ESS, see http://www.ees-homepage.net/ and http://www.piredeu.eu/) and for discrimination the European Union minorities and discrimination survey(EU-MIDIS, see http://fra.europa.eu/en/project/2011/eu-midis-european-union-minorities-and-discrimination-survey).(27) See for instance the OECD indicators on immigrant integration, which also use ESS data from all five waves OECD(2012): Settling In. OECD Indicators of Immigrant Integration. http://www.oecd.org/migration/integrationindicators/.

PART 3: Using EU indicatorsIntroductionMonitoring integration has become a reality atlocal, regional, national, and international level.At the 13-14 December 2011 Justice and HomeAffairs Council, Member States proposed to workwith the Commission to further use the commonintegration indicators, as part of a non-bindingcoordination mechanism to improve existingstructures and tool for EU knowledge exchange.This report outlines how policy actors can furtheruse indicators to improve integration.The first section of the report summarises experienceswith integration indicators in the MemberStates and with indicators for the EU’s generalpolicy goals, which are also relevant for integration.Drawing on these lessons learned, the mainsection of this report proposes an incrementalapproach to using the EU integration indicatorsat EU level and within the Member States.Throughout the development of the EU integrationindicators pilot project, three key purposeshave been discussed for integration indicators:understanding the national contexts for integrationin a comparative way, evaluating the resultsof policies, and using targets to mainstreamand improve integration. The EU has alreadyachieved first steps on all three policy purposesthrough already existing EU indicators, this project’sanalysis, and existing EU-funded sources.These current achievements should be continuedand used as the starting point for further analysis.Building on this baseline, the Commissionand Member States can obtain better tools forevidence-based policymaking. A detailed pictureof integration, robust policy evaluations, andgreater mainstreaming can make policies moreeffective for immigrants and better accepted bythe public.This main section proposes that the integrationindicators build on existing European cooperationwith integration stakeholders. The reportconsiders what roles EU integration stakeholdersand networks can play in developing and usingthe indicators. Eurostat and national statisticalinstitutes are working to supply and improvedata. Integration information and indicators areregularly gathered through Eurostat, the EuropeanWebsite on Integration, the European MigrationNetwork, and EU-funded research. TheNCPIs and European Integration Forum can beconsulted for their insights of policymakers,practitioners, and immigrant representatives.The report highlights the added value of EU integrationindicators for both Member States withand without existing national integration indicators.Using these indicators involves enhancedcooperation of Eurostat, the various CommissionDirectorate-Generals, the Member States, andstakeholders.Chapter 10:Background on usingintegration indicators10.1 Key issues forintegration indicatorsWhat do integration indicators tell us about integration?Are high employment rates for immigrantsa measure for ‘good integration’ policy?Are small differences in educational attainmentbetween immigrants and non-immigrants agood benchmark for ‘integration’? This projecthighlighted a number of issues to keep in mindwhen designing or using indicators, whatever thepolicy field.Who is an immigrant?The definition of immigrants in internationalstatistics depends on the available variablesthat can identify immigrants in the dataset. ‘Immigrant’is a catch-all category which can referto ‘country of birth’, ‘citizenship’, ‘country ofbirth of parents’. Based on these variables, immigrantsare referred to as ‘foreign-born’, ‘foreigners’,‘second generation.’ Immigrants bornto at least one foreign-born parent are referredto ‘mixed background’. Age groups are also importantfor analysing indicators. Most indicatorsfocus on a specific sub-group of the population.For example, employment rates look at immigrantsbetween 20 and 64, while early schoolleaver rates look at immigrants aged 18 to 24.The population of one age group may have differentcharacteristics than the majority of immigrantsof all ages.39

40Who is missing from statistics?The most socially excluded people are often under-representedin the surveys upon which manynational and EU indicators are built. They may beunder-sampled due to their higher non-responserates. In many countries, people are excludedfrom SILC if they are living in collective housingor homeless. Official statistics will never captureall hard-to-reach groups.Measuring rates vs. gapsIt is important to consider the type of comparisonrelevant for integration. Rates are shares or levelsmeasured as percentages. Gaps are the differencebetween the shares of immigrants andthe total or non-immigrant population. The twomeasurements are two different ways of lookingat society. For example, smaller gaps are generallyobserved for Central European countries,mainly because the rates for the total populationtend to be lower in comparison to the total populationin Western European countries. One mightalso think that rates and gaps are negativelycorrelated; when shares of immigrants increase,the gaps between immigrants and the totalpopulation would decrease. However, this is notalways the case because the share for the totalpopulation may have increased to an even largerextent than the share for immigrants. As a result,both rates and gaps are relevant to compare thesituation of immigrants across countries.Reference groupWho do we compare immigrants to? Eurostathas selected the ‘total population’ as a referencegroup. This includes immigrants. Including immigrantsin the reference group may have some effecton the results, depending on the proportionof immigrants within the total population and thesize of the difference in shares/levels between thetwo groups. Comparing immigrants to non-immigrantsis more precise for measuring the situationof immigrants. However, the total population maybe a relevant reference group to measure in howfar immigrants are converging towards the wholeof society, of which they are part.Areas of integration: employment, education, socialinclusion, and active citizenshipThis project’s analysis showed that the indicatorsresults within one area of integration areoften correlated to one another. For example,countries’ with high outcomes on one educationindicator (e.g. reading skills of 15 year olds)usually also have high outcomes on other educationindicators (e.g. early school leaving). Thedifferent areas of integration are not always correlatedto one another. Looking at only one ortwo areas of integration, such as employment oreducation, provides a misleading picture. Mostcountries have high outcomes in one area andlow outcomes in another.How does society affect integration?The characteristics of the receiving society area key but often neglected factor that influencesthe integration situation and the effectivenessof integration policies. Integration ‘outcomes’may be less the result of targeted integrationpolicies. The structure of the school system, thelabour market, or the political system, may havea greater effect on immigrants’ educational attainment,employment situation, or political participation.The situation of immigrants in manyof these areas is also related to the situation ofnatives in the country. It is advisable to compareimmigrants and non-immigrants with similarbackgrounds and analyse the effect of the generalpolicies and context.Correlation vs. causalityA common debate in science is the question ofcorrelation and causation. Two variables that arerelated to each other at the same time (e.g. correlationsbetween contextual factors and integrationoutcomes) are not necessarily taken to havea cause-and-effect relationship, meaning that onecauses the other. We do not know which one is thecause and which the effect (a.k.a. ‘reverse causality’).In addition, the two variables may only berelated to each other because both are causesor effects of a third ‘unobserved variable.’ Integrationindicators, which use aggregate data toassess individual behaviour, have their limits forexplaining the integration process. A correlationobserved at the group level cannot be assumed toapply at the individual level (usually referred to asan ‘ecological fallacy’). What is true at the countrylevel is not necessary true at the individual level.Further analysis (e.g. multivariate, longitudinal,econometric methods) helps clarify the full meaningof aggregate-level data.10.2 Integration indicatorsin the Member StatesStarting in the 1990s, a few governments havesupported projects to measure integration invarious forms, often as part of multi-annualstrategic plans on integration. Several others arecurrently debating whether to establish integrationindicators. Ambitious local and regional governmentsintroduced their own indicators along-

side those of national governments (e.g. Germanstates and cities, Catalonia, Flanders, Copenhagen,and Vienna). A few Member State governmentshave established national integration indicatorsand monitoring systems. Most of these studiesare prepared by the national statistical institute(e.g. Austria, France, Netherlands, Norway), researchers(the Czech Republic, Germany). Monitoringcan also be undertaken by the government(e.g. Denmark) or civil society, for example the IntegrationCentre and Economic and Social ResearchInstitute in Ireland. Seminar participants alsomentioned examples of civil society monitoringin other countries, such as SVR’s ‘IntegrationBarometer’ in Germany, ISMU’s Integration monitoringin Italy, and the work of the ImmigrationObservatory in Portugal. The type of systemmainly depends on political definition and goals ofintegration as well as the availability of officialnational statistics and alternative (quantitative orqualitative) survey data.The analysis of integration indicators varies significantlyacross Europe (see Annex). This projectreviewed the international literature on integrationindicators as well as local and nationalexamples available in Dutch, English, French, orGerman. So far, most Member States with integrationindicators have commissioned one-offreports on different areas and indicators (e.g. Estonia,France, Italy, and Sweden). Through morecomprehensive monitoring systems, Austria, theCzech Republic, Germany, Ireland, the Netherlands,and Norway have a regular monitor of thesituation of immigrants. A standardised reportuses the same set of indicators in many areasrelevant to integration. The Netherlands Institutefor Social Research (SCP) also publishes an accompanyingad hoc thematic report that delvesinto one thematic area. Denmark uses benchmarkingto compare the situation of immigrantsto the governments’ policy goals. Rather thanmonitor integration for its own sake, a benchmarkingsystem attempts to evaluate governmentperformance in relation to intended policyoutcomes.These reports’ target groups include broadgroups, such as people with a migrant background,and sometimes more specific groups,such as third-country nationals, refugees, certainnationalities and ethnic groups. To a limitedextent, national integration indicators capturethe differences in the situation at local and regionallevel: the Länder in Austria, municipalitiesin Denmark, and the regions and départementsin France. The most common type of analysis isbasic descriptive analysis of the situation of immigrants(e.g. the employment rate) or the comparison(e.g. gaps in employment rates) with thetotal population or general categories of people(e.g. nationals, native-born, people withoutany immigrant parents). Multivariate analysis,a key statistical method explained in this report,is used by researchers in just a few indicatorsystems (e.g. France, Germany, and the Netherlands).The relationship between integrationoutcomes and policies is only analysed in the reportin Ireland. The effects of specific policies aremainly analysed in separate policy evaluations,where they exist.Member States with national indicators canuse the EU indicators to better understand theintegration context in other Member States andacross the EU, and place their own nationalsituation in context. International comparisonsidentify the differences in national contexts aswell as common trends and challenges. Comparingcountries does not ignore the national context.On the contrary, comparative research canmeasure contextual factors across countries anddemonstrate which matter for integration outcomes.This analysis clarifies how each country’ssituation is the result of the mix of these differentcontextual factors. The future developmentand use of the EU integration indicatorsmay give them new ideas to develop their ownapproach to monitoring. Moreover, replicatinganalysis with EU indicators at internationallevel increases the robustness of the findingsfrom their national indicator studies. MemberStates currently without national indicatorscan decide how to use the EU indicators at variouslevels of governance in their own country.They could also complement the EU indicatorswith additional national indicators.10.3 Beyond integration:Using indicators in otherareas of EU cooperationThe development of EU integration indicatorscan draw many lessons from the use of indicatorsin other areas of cooperation within the EU,especially the Europe 2020 strategy. For example,EU cooperation in the areas of employmentand social inclusion began nearly ten years agowith the long, incremental process of identifyingand using indicators. The relevant ministriescompiled EU portfolios of agreed indicators asan underlying data source to accurately comparetheir situations. EU Working Groups on Indicatorsnarrowed down the number of outcome indicatorsto a limited number of core indicators andbenchmarks. As these expert committees adaptthe indicators, Member States improve dataavailability and further discuss the interpretation41

42of the results. For example, a newly-proposedmethodology, known as the ‘Joint AssessmentFramework’, is available to interpret indicatorresults in EU cooperation on social inclusion,employment, and, most recently, education. Thisanalytical tool allows for a joint product andjoint ownership by the Commission and MemberStates. Together, they identify key challenges,potential risks, or best practices that make a differencein achieving better labour market andsocial outcomes.Given this policy context, the report argues thatthe existing EU integration indicators help policyactors to mainstream integration into countries’strategies on employment, education, social inclusion,and active citizenship. Since the existingEU integration indicators are often based onthe EU’s general targets and indicators, they caneasily be used in the development of generalpolicies, if monitoring migrants and other vulnerablegroups is made a priority.Chapter 11:How are integrationindicators used forpolicymaking?Purposes & methodsBehind the idea of ‘indicators’ lie various policypurposes and various research methods. Measuringsocietal outcomes is a starting point for evidence-basedpolicymaking, whatever the policyfield. These outcome indicators can then be analysedin different ways to answer specific policyquestions. These different types of analysis areneeded for policy debates, planning, implementation,and evaluation. Each type of analysis requiresthe proper quantitative and qualitative methods.Throughout the background documents on theEU integration indicators, the Commission and theresponsible national ministers have outlined threekey policy purposes for using integration indicators:understanding integration contexts, evaluatingthe results of policies, and mainstreamingintegration into general policies. These purposesare not one in the same. Measuring the situationof immigrants is not in itself an evaluation of theresults of integration policies. The results of thesepolicies cannot all be measured in terms of immigrants’outcomes on integration indicators. Forintegration stakeholders, not all Europe 2020 targetsare relevant for integration. For policymakersworking in other fields, many of their policies arenot significantly affected by the situation of immigrantsor the results of integration policies. Whileall of these purposes are important for policymaking,policy actors may choose to go further and developmore tools for one purpose than the others,depending on the context of each country and whatinformation they need for policymaking.Policy actors have various options for using integrationindicators for each of these purposes. Alloptions can be implemented and combined in differentways and at different times. The main onesare listed in the chart below and fully explained inthis section of the report. The first column detailsthe current achievements for all three policy purposes.At this point in the process, the Commissionand Member States will consider what furtheranalysis of the integration indicators wouldimprove policymaking at national and EU level.Maintaining and building on the current baseline,they can obtain better tools for evidence-basedpolicymaking. A detailed picture of integration, robustpolicy evaluations, and greater mainstreamingcan be used to make policies more effectivefor immigrants and better accepted by the public.11.1. First purpose:Understanding the contextUnderstanding the situation is the main purposeof any statistic. This situation is captured througha baseline of indicators, each of which capturesone key element or outcome of that situation.Numbers are produced for each indicator on thebasis of an established definition of the indicator,the use of harmonised data, and the propermethod for calculation. For comparing the situation(e.g. comparing over time, or between cities,regions, or countries, or against groups withsimilar or different characteristics, etc.), indicatorsmust be produced based on harmonised definitions,data, and calculations. National statisticalinstitutes often harmonise indicators between citiesand regions within the country. Internationally,national statistical institutes harmonise indicatorsthrough collaboration with the EU (through Eurostat),the OECD, and UN agencies.Understanding and comparing the situation ofimmigrants is the basic purposes of any integrationindicators, whether at international, national,regional, or local level. This basic purpose wasintroduced in the first paragraph of 2009 Conclusionsof the Malmo expert meeting, the documentcontaining the original list of indicators:

Proposal: An incremental approach to using integration indicatorsUnderstanding the contextof integration (statisticians,researchers)Evaluating the resultsof policies (researchers)Using targets to improve& mainstream; integration(policy actors)Work in Progress:Maintain the baselineImproved annual publication ofEU indicatorsWhat: Indicator results on Eurostatwebsite & basic descriptiveanalysis in a ‘Statistics in Focus’reportHow: Commission & MemberStates consider additionalindicators; Eurostat updatesthe results; NSIs improve dataavailability and quality (see assessmentreport)Existing resources: Eurostat &NSIs on-going work to mainstreammigrants in statistics,boost migrant samples, &develop ad hoc modules forsurveysCompare data on policies andoutcomesWhat: Analysis of the complexrelationships between integrationpolicies & outcomesHow: Define policies in terms ofintended outcomes on integrationindicators; conduct bivariateanalysis & mutual learningabout how policies influenceoutcomes and/or how outcomesare influencing policiesExisting resources: Up-to-date& comparable summaries ofpolicies through EU-fundedresearch, EMN, & EWSIKeep integration indicatorsrelevant for EU targetsWhat: Europe 2020 targets arebasis for integration indicators &thus can now be monitored forimmigrantsHow: EU integration indicatorscan be changed to reflectchanges in Europe 2020 targetsand statisticsExisting resources: Coordinationamong DGs & EurostatSecond option:Build on the baselineMultiannual ‘integration report’based on indicatorsWhat: In-depth descriptiveanalysis of the integration situation& different national contextsHow: Data disaggregated forspecific groups, monitored overtime, & compared between immigrants& non-immigrants withsame demographic characteristics(i.e. comparing ‘like withlike’ through statistical controls).Report can also include ad hocthematic chapters.Existing resources: Examplesinclude 2011 Eurostat ‘StatisticalPortrait’, 2012 OECD ‘SettlingIn’, as well as national and localreportsGather contextual data on whobenefits from policiesWhat: In-depth descriptiveanalysis of the implementation ofpolicies in the four indicator areasHow: Gather and share statisticson policy implementation, the potentialand current beneficiariesExisting resources: Administrativeand official statistics, EU-fundedresearch, EMN, & EWSICalculate how integrationimproves EU targetsWhat: Calculations identify areas& countries where immigrantsare a major target group forgeneral policiesHow: Statisticians calculate how‘closing the gap’ for immigrantshelps EU countries meet theirtargets for smart, sustainable,and inclusive growth.Existing resources: EU OMCs &EurostatThird option: EvidencebasedpolicymakingMultivariate & longitudinalanalysisWhat: Analysis determines whichcontextual factors have most/least influence on integrationoutcomesHow: Research analyses therelationship between EU integrationindicators and three setsof factors (differences in theimmigrant populations, generalpolicies & contexts, immigration& integration policies). Longitudinaldata can also be analysedwhere possible.Existing resources: Existinginternational datasets on thesecontextual factors, nationallongitudinal datasets, nationalmultivariate analysisEconometric causalevaluations of policy impactWhat: Assesses prospective orretrospective impact of specificnational policies on integrationoutcomesHow: Evaluation studies are doneat national level, sub-nationallevel, or between countries,depending on the availability ofdata. The EU can provide a reviewof such studies, exchange onmethods, & quality standardsExisting resources: Examples ofcausal evaluations in EU and traditionalcountries of immigration.Set specific national goals& targets for integrationWhat: Member States assessindicator results & set their ownspecific national goals & voluntarytargetsHow: Results assessed basedon best available multivariateanalysis & policy impact evaluations;Member States learn fromcountries using targetsExisting resources: EU OMCs,cooperation structures betweenrelevant ministries, and with civilsociety43

44“The function of the indicator is to give an overviewof, and the possibility to monitor the situation,and to indicate if something essential is happening,considering that integration is an on-goingprocess irrespective of different interventions.”The EU integration indicators make it easier to understandthe integration context in the EU MemberStates so that policy actors can better learnfrom one other. The results show the similaritiesand differences in national contexts, while furtheranalysis reveals what factors explain these similaritiesand differences. Integration outcomes indifferent countries are often related to the samekey aspects of the immigrant population, thegeneral context, and national policies. The morethese factors are present in a country, the morelikely are certain integration outcomes in thatcountry. For example, the more newcomers thereare in a country, often the greater are the integrationchallenges. This analysis helps policy actorsto appreciate the unique combination of factorsinfluencing integration in their country and othercountries. Moreover, the use of indicators overtime gives policy actors a new long-term perspectivefor policy planning. The availability of theseindicators is therefore a starting-point for moreinformed mutual learning across the EU.11.1.1. First option: Eurostat continuesto publish the EU indicators annuallyonline and to work with NSIs toimprove data availability and qualityon immigrants.An improved annual production of the EU indicatorsmaintains the robust comparative frameworkfor assessing the integration context in theEU Member States and over time. Eurostat hasprovided the public with updated indicator resultsever since the start of this pilot project. Eurostatmade the results more accessible through theirpublication in English, French, and German withinone of its most consulted sections on its website.Eurostat submitted the data to the same reliabilitytests as any other EU statistics and only publishedreliable results. This project’s data assessment reportreconfirmed that this harmonised data is thebest available and highest quality. Given the reliabilityof the data, the indicators are valuable forassessing integration outcomes.Future updates of the results can be publishedin an easy-to-read visual format. The results arereported both as shares (e.g. foreign-born employmentrate) and as gaps (e.g. foreign-bornemployment rate compared to total population).The results are broken down by the existing categories:broad group of country of citizenship andbirth, gender, and age group. The publication ofan accompanying ‘Statistics in Focus’ would makethe indicator results more accessible and easierto-interpret.These short and standardised summarisesprovide easy-to-interpret visuals andbasic descriptive analysis. For example, ‘Statisticsin Focus’ are annually published on demographicgrowth, migrant flows and populations, asylum,residence permits, and the acquisition of citizenship.These results are published as part of the Europeanmutual learning process involving the MemberStates. Eurostat can regularly discuss the indicatorsduring regular meetings with the national statisticalinstitutes (NSIs). The EU integration indicatorsreinforce the political relevance of the existingexercises undertaken by Eurostat and the NSIs toimprove the availability and quality of data on immigrants,which includes the data behind the EU integrationindicators. Specifically, this project’s dataassessment chapter recommends that:n Eurostat continue its work with NSIs on mainstreamingmigration into official statisticsn Eurostat and NSIs discuss common methodsfor weighting, boosting migrant sample sizes,decreasing non-response rates, and capturingthe second generation in annual EU surveysn NSIs calculate confidence intervals for the EUintegration indicatorsn EU collect the missing data for core and additionalindicators through migrant-specificsurveys or ad hoc modules in LFS, SILC, and/orthe European Social SurveyThe National Contact Points on Integration (NCPIs)are also actively involved in this process. NCPIsprovide feedback and discuss the lessons learnedfrom the results. They can also support and expandthe work of the NSIs in order to address themajor information needs of policy actors. Beyondthe existing EU indicators, the ministers responsiblefor integration may decide on additionalindicators, in particular those identified throughthis project’s expert seminars and review of theresearch and indicators at national and EU level.11.1.2. Second option: Regularpublication of an integration reporton Eruostat data with in-depthdescriptive analysis, includingspecific groups and themes.Eurostat or an independent research institutecould provide every three or four years a morerefined descriptive analysis through an ‘integrationreport.’ Previous examples of such reports

are the 2011 Eurostat ‘Migrants in Europe: AStatistical portrait,’ the 2012 OECD ‘Settling In:OECD Indicators of Immigrant Integration,’ andindicator systems at national or local level. Also,European Commission services in other policyareas, such as employment and education,sometimes include indicators on immigrants.Drawing from these examples, a central reportcould monitor changes over time, further disaggregatethe data for specific groups, controlfor demographic factors, compare immigrantswith similar groups in the population, and exploitunique datasets to investigate specific themes.Trends over time are a main focus of any multiannualpublication through the construction of‘cohorts’ (people arriving in same years/period)or descriptive analysis over several years (includingpooling data over three-year-periods,where possible). In-depth descriptive analysiscan disaggregate the indicators for more specificimmigrant groups beyond the standard breakdownsby country of origin and citizenship, age,and gender. This project proposed ‘demographicindicators,’ which can also be used for this disaggregation.These factors are often related tointegration outcomes:n Socio-economic status (education level, employmentstatus, income)n Duration of residence (e.g. 0-5, 6-10, 10-15,15-20, 20+ years)n Reason of migration: migrant workers, reunitedfamilies, humanitarian, and studentsn Origin (by country, world region, countries’ levelof development, or countries with differentofficial language(s) than that of the countryof residence).n Composition of the household (disaggregatedfor men and for women)The use of statistical controls shows the importanceof key factors influencing integration.Controls adjust for the differences in the immigrantpopulation across countries based onkey demographic characteristics (for example,see the OECD’s 2012 ‘Settling In’ report). Usingthese dis-aggregations and statistical controls,immigrants can then be compared to a ‘controlgroup’ of non-immigrants in similar situations.Immigrants may share the same outcomes andproblems as non-immigrants of the same age,gender, or socio-economic status. In such cases,overall gaps between immigrants and the totalpopulation actually have nothing to do with integration.Comparing a similar group of immigrantsand non-immigrants (a.k.a. comparing‘like with like’) is crucial for integration policymakingand mainstreaming. Without this, generalsocial problems risk being misdiagnosedas specific integration problems. In these cases,general policies may be more effective for immigrantsthan targeted integration policies.Lastly, ad hoc thematic chapters would make amultiannual report more relevant for integrationactors. The theme could follow the prioritiesfor integration policy and EUROPE 2020 orthe changing social situation (e.g. effects of thecrisis and austerity measures, neighbourhoodand school segregation, family composition). Additionalindicators can be piloted whenever thereare new data sources, including one-off surveysand datasets.11.1.3. Third option: The EU usesmultivariate and longitudinal analysisto determine what contextual factorsinfluence integration outcomes.Multivariate and longitudinal analyses determinewhich contextual factors have the mostand least influence on integration outcomes.Similar to statistical controls, multivariate analysistries to control for all demographic andcontextual factors influencing outcomes. Thesestudies also often use comparable control groupof non-immigrants. In terms of relevant factors,research need to look beyond the demographiccharacteristics of the immigrant population inorder to determine the major factors influencingintegration. Future projects can use the wealthof existing reliable international data on variouscontextual factors, which are collected acrossmany EU countries, regions, and cities. This pilotprojects’ analysis used many of these sourcesto identify significant correlations between theindicator outcomes and three types of contextualfactors: the characteristics of the immigrantpopulation, a country’s general context and policies,and migration and integration policies. Thisanalysis of contextual factors gives policy actorsa more accurate and systematic way to interpretthe results of integration indicators when comparingtheir national contexts to each other.Independent research institutes are best suitedto undertake this analysis through large-scaleresearch projects. This project’s analysis teamrecommends the use of individual data stemmingfrom cross-national surveys, followed byOLS and logistic regression in order to link individuallevel explanatory factors and country(and/or region) proxies to the outcome indicatorson the individual level. Analysis of the individuallevel data can help rule out possible misinterpre-45

46tations of the data that can emerge in the interpretationof patterns on the aggregate level. Thisis a common strategy in analysing cross-nationalpatterns with European surveys like LFS and EU-SILC. The use of more sophisticated statisticalmethods like multilevel analysis is generally notadvisable, given the limited number of countries(27 EU Member states) and the absence of randomsampling at country level.During these projects, the results of internationalanalysis can be checked against existingnational multivariate analysis or replicated onnational datasets in order to enhance the robustnessof the results from EU and national integrationindicators. The descriptive and multivariateanalysis described in this section is thereforecomplementary, since international findings onindicators can be corroborated and further developedwith more in-depth national, regional, orlocal studies.Longitudinal data and analysis, mostly availableat national level, present the best opportunitiesfor monitoring integration progress over time.Representative longitudinal datasets track individualsover several points in time. The EuropeanMigration Network’s 2010 Annual Conferenceconcluded that integration monitoring wouldbenefit from more immigrant longitudinal surveys,survey modules, and administrative datasets.Whatever the method used, longitudinaldatasets take significant time and money to obtainthe necessary sample sizes and level of sophisticationin methodology. The manifold usesof longitudinal data make it worth the effort. Researcherscan identify the major facilitators andobstacles influencing outcomes. Through regularupdates about newcomers’ integration experiences,policymakers can identify their needs andact on them immediately. Countries will also getthe most out of longitudinal surveys if they aredesigned to link up with the EU integration indicators,EU data-sources (e.g. LFS, SILC), and longitudinalsurveys in other Member States (as wasdone with these surveys in Canada, Australia, andNew Zealand).11.2. A second purpose:Evaluating the resultsof policiesEvaluating the results of policies remainsa challenging priority for integration policy. Ahandful of countries have developed indicatorsto monitor the situation of immigrants and usedthe results to change their policies. But to date,the effectiveness of integration policy has beenanalysed by very few policy evaluations. Theseevaluations require not only the political will, butalso sophisticated sets of data covering the periodsbefore and after the adoption of the policyor programme.Integration outcomes are sometimes misinterpretedas the outcomes of integration policies.However, the causal links between policy and integrationoutcomes are difficult to prove. Outcomeindicators are never the ‘pure’ results of policies.The integration situation reflects a broadersocio-economic and policy context. These otherneglected factors may have a greater effect onintegration than immigration or integration policiesthemselves. An exclusive focus on outcometargets reduces the potential for mutual learningas it is not always clear whether and which policiescontributed to changes in the situation.Ministers responsible for home affairs invited theNCPIs and Commission in June 2007 to promotethe development of common indicators and indexesto assess integration policy outcomes.In the 2010-2014 Stockholm Programme, theEuropean Council used similar terms: core indicatorsfor evaluating the results of integrationpolicies. The conclusions of the Malmo expertseminar, where the EU indicators first appeared,clarified the link between these indicators andpolicy evaluations:For the purpose of monitoring the outcome ofintegration policies, outcome indicators will beused…Core indicators will provide a basis formonitoring the situation of immigrants and theoutcome of integration policies. They will be acomplement to national indicators and reinforcethe basis for evaluations at national level.These outcome indicators can be used in evaluationsof integration policies. The idea is thatnational evaluations will assess how integrationpolicies affect employment rates, early schoolleaving, educational performance and attainment,at-risk-of-poverty-and-social-exclusion,health, housing, naturalisation, long-term residence,and political participation.Accordingly, this project conducted desk researchand expert seminars to ascertain what roles integrationindicators do and do not play in policyevaluations. Integration indicators can be used inthree ways to make to policy evaluations easierand better. The first two options provide usefuldata for policy evaluations. Policy actors can betterlink changes in outcomes to their policies ifthey provide data on policy changes, the policy’sintended outcomes, the beneficiaries, and theimplementation. Econometric causal evaluationscombine the EU indicators with this and other

data with the aim to assess policy effectiveness.These studies identify a policy’s specific effecton integration outcomes, controlling for all otherpossible explanatory factors.11.2.1. First option: Comparing bothoutcomes and policies provides a newevidence-base for policy planning anddebating policy effectiveness.As Eurostat publishes annual indicators resultson the situation of immigrants, policy actors cancompare it to the latest information on policiesand policy changes. During this project, the seminardiscussion papers included aggregate-levelbivariate (correlation) analysis demonstratedthat countries’ integration outcomes, as measuredby the EU indicators, are in part related totheir national contexts, including their nationalpolicies, for instance: their immigration channels,labour market integration programmes, generaleconomic structures, general and targeted educationpolicies, social benefits, political rights,residence and citizenship rules, and anti-discriminationlaws. This bivariate analysis provided abasic method to identify significant relationshipsbetween the situation of immigrants and differentnational policies. The relationships betweenoutcomes and policies were significant and, insome cases, surprising. During the seminars,experts began interpreting the results and welcomedthe chance to exchange about the complexrelationships between integration outcomesand policies in their countries.This baseline analysis is easy to replicate overtime as the latest data is made available on theEU indicators and these contextual factors. Howeverthis type of analysis is not possible withoutup-to-date comparable data on targeted andgeneral policies. Existing EU-funded researchand mechanisms have been keeping track of themany changes in the targeted and, to a lesserextent, general policies across the EU. Thanks tothis investment in up-to-date information, policyactors can conduct further research and mutuallearning to investigate if and how the situationof immigrants is related to the various generaland targeted policies across Europe.Comparisons of policies and outcomes are commonlyused by integration policymakers as thecontext or starting point for public debates andpolicy planning. This information can lead to morenuanced views of integration in public debates.Policymakers have used these indicators to betterreflect on their long-term goals for migrant integration.Notwithstanding these benefits, indicatorson their own only have an indirect value for policyplanning. The policy implications of indicators resultsare rarely clear. The results of outcome indicatorsare not the result of any one actor’s actions.Outcome indicators alone do not provide enoughinformation to justify statements or actions penalisingcertain groups. The road from indicators topolicy changes can be very long.As part of policy planning, Member States canredefine their policies in terms of their intendedmeasurable outcomes. Which integration indicatorsare relevant for the outcomes of which policies?Integration policies are focused on specificareas of life and specific beneficiaries and targetgroups. In contrast, most outcome indicatorstend to have very broad target groups (e.g. foreign-born),many of whom are not affected by oreligible for integration policies or programmes.For example, it would be inappropriate to definethe foreign-born employment rate as theoutcome of targeted integration programmes,which tend to focus on certain categories ofnon-EU newcomers. For other types of integrationpolicies, policymakers tend to describe thegoals of their policy in very broad terms. Forexample, anti-discrimination laws aim to reducediscrimination in society and guarantee equalopportunities, but how exactly? These objectivesare not only difficult to measure as outcomes,but also tend to create unrealistic expectationsabout the effectiveness of these policies. Selectingoutcome indicators for specific policies helpspolicymakers to design realistic and targetedpolicies, assess the prospective impact of theirpolicy, and respond to changes in the situation.11.2.2. Second option: Gathering contextualdata on policy implementationprovide the missing links betweenintegration policies and outcomes.Building on this baseline, data on the implementationof policies provides an evidence-base forevaluations of policy effectiveness. The results ofbivariate analysis reveal statistically significantrelationships behind integration outcomes andpolicies across countries, findings which meritfurther investigation. However, the absence of astrong statistical relationship is not the final wordin policy evaluation. Just because no significantrelationship emerges between the two, it doesnot follow that the specific policy has no effect onthe situation of immigrants. Instead, this findingmeans that the policy does not have the samesystematic effect on the situation of immigrantsin all countries. In some countries, the policy maybe new and thus far affect few immigrants. Thepolicy is well implemented and funded in somecountries, but not in others. Some countries mayhave an immigrant population that needs or benefitsmore from this policy than in other countries.47

48A country’s policy may still affect the situationof immigrants even if no general relationshipemerges between outcomes and policies acrosscountries. Simply monitoring indicators will noton its own say much about the effectiveness of apolicy in a specific national context.Data on the implementation and uptake of policiesare very useful for policy evaluation. Implementationstatistics would help policy actorsto assess the relevance of their policy for theintegration outcomes that policymakers want tochange through the policy. Statistical analysiscan identify and quantify the potential beneficiariesof a given policy, or, in other words, thenumber of immigrants eligible for a given policy.After adoption of the policy, implementing agenciescan collect statistics on their administrativeand financial inputs as well as the policy outputs,measured in terms of its uptake by beneficiaries.Data can also be collected for disadvantagedgroups that are less likely to access these services,for example women, the elderly, children andunaccompanied minors, young adults, family migrants,beneficiaries of international protection,and the lower-educated.11.2.3. Third option: Independentinstitutes conduct econometric causalevaluations of the impact of policieson integration outcomes, as measuredby EU indicators.Policy actors who want to know the results oftheir policy and its impact on integration outcomesneed dedicated evaluation studies of theirpolicy’s impact. Bivariate or multivariate analysisdoes not prove causality in the purest sense.There can always be more factors that influencean outcome which the multivariate analysis hasnot taken into account. Policy evaluations assesshow the policy and other factors are affectingspecific outcomes, such as those measured byintegration indicators.The feasibility of these evaluations largely dependson the availability and the quality of thedata on policy implementation and its beneficiaries.Depending on the policy and availability ofthe data, these evaluations can be done at local,regional, and national level or comparativelybetween cities, regions, and countries. AcrossEurope, there are a small but growing numberof evaluations making the link between policiesand outcomes. These policy evaluations are oftenlittle known internationally or even within thesame country because they are state-of-the-art,covering many different policy areas, and difficultto understand for the statistical layman. Giventhe few examples of these evaluations, MemberStates could benefit from a literature review ofthese studies in the EU and traditional countriesof immigration.The best impact evaluations rely on sophisticatedeconometric methods of causal inference.Rigorous impact evaluations are able to takeinto account any observable and non-observablefactor that could determine the outcomes. EUexchanges on the best methods for policy evaluationwould raise awareness among policymakersand provide valuable training to practitioners.Quality standards for evaluating integration policieswould help practitioners to put these methodsinto practice across Europe. Policymakerswould be better able to contribute to the designof the evaluation and assess the quality of themethods and the results.11.3. Third purpose: Usetargets to mainstream andimprove integrationSetting measurable objectives and targets helpspolicymakers to decide on, communicate, andthen evaluate their goals and actions. The terms‘target’, ‘goal’ and ‘objective’ are often used interchangeablyand the differences between themare subtle. Targets are a) specifying a level ofgovernment action, b) time-bound and c) measureableby indicators. When policymakers committo targets by which they will judge their policies,their proposals are more credible, clear, and accountablefor the public. Targets give indicatorsa policy purpose. With targets, indicators checkwhether immigrants’ situation is improving asintended. Impact assessments analyse how policiesaffected or could affect changes in the situation.Targets also work as a mainstreaming toolto improve integration through general policies.Targets draw attention to the importance of generalpolicies which may have a greater impact onreducing inequalities than targeted policies.Setting targets not only aim to ‘close the gaps’between immigrants and the total population,but also to improve the general conditions insociety. If policymakers want to reduce unemploymentand poverty in their country, create aknowledge-based society, and encourage activecitizens, then immigrants are a major targetgroup for general policies in many of the EU’scountries of immigration. This project’s analysisreport estimated the impact on society of improvingintegration (by ‘closing the gaps’). Howeverin many social reforms, immigrants areleft behind. Ignoring the situation of immigrantsand ethnic minorities is one of the reasons whycountries fail to meet their socio-economic

goals. Immigrants are often addressed as ‘vulnerablegroups.’ This label risks to overlook whatimmigrants add to society, for example theirentrepreneurship, work filling labour shortages,university qualifications, role as internationalstudents, foreign language skills, new families,active citizenship, diverse backgrounds, and contributionto greater openness in society. Immigrantsbring valuable resources for society andeconomic growth. Governments have the greatestimpact on their socio-economic goals byeffectively targeting the most relevant groups,including immigrants.Policymakers who attended the project’s seminarsshared the experience of countries and citiesthat have set integration targets. The Germanfederal government used the results of itsnational integration indicators to set measurablepolicy goals and targets, agree on new policies,and monitor progress. In addition, Denmark usesmeasurable goals and targets not only to monitorprogress through its Integration Barometer,but even to benchmark and reward municipalitiesfor their performance in contributing to theseoverall goals. Cities like Berlin and Vienna workwith Diversity Benchmarks to improve diversitywithin the city administration. A few MemberStates have also set integration-specific targetsas part of their National Reform Programmes. AtEU level, the use of targets for integration couldbe foreseen in existing policy frameworks suchas the Open Method of Coordination in the areaof employment, education, social inclusion andprotection, and culture. This would involve Eurostat,various national ministries, and various EUstructures11.3.1. First option: Looking at EUintegration indicators within theframework of key EU targets.The Europe 2020 Strategy aims for higherlevels of employment, productivity, and socialcohesion in the EU. The EU agreed five ambitiousHeadline Targets to be reached by 2020,towards which Member States set their ownnational targets. Migration and integration havesteadily received greater attention in existingEU OMCs. An important aspect of EU integrationindicators is that the Europe 2020 and theET 2020 have now been disaggregated andprovided for all EU Member States, accordingto harmonised definitions of country of birthand citizenship. The integration indicators cantherefore easily be included in these processes,if monitoring migrants is made a priority. TheEuropean Commission can use these indicators’results in order to mainstream migrants into EUrelevant policies.The existing list of EU integration indicatorslargely reflects the EU 2020 targets, agreed atthe highest political level. Since these generalpolicies and goals are relevant for integration,the EU integration indicators generally use thesame indicators, disaggregated for immigrantgroups. The Commission will have to adapt theEU integration indicators to keep up with changesin the EU’s main targets. For example, sincethe conclusions on the EU integration indicatorsat Zaragoza, the European Council in 2010adopted Europe 2020 targets that expandedthe definition of poverty from at-risk-of-povertyto at-risk-of-poverty-and-social-exclusion. Respondingto the crisis’ impact on youth unemployment,the European Council in May 2012added a new ET2020 benchmark on the share ofemployed graduates (20-34-year-olds) in theirfirst three years after graduation. New indicatorsmay be proposed on social inclusion, social protection,and health. Aligning the EU integrationindicators requires basic coordination betweenEU Commission services and the relevant indicatorsub-groups.11.3.2. Second option: Calculate howimproving integration helps MemberStates to meet their general goalsand targets for society.The EU integration indicators can be used toestimate the impact of improving integrationon the general conditions in society. Closingthe gap for immigrants will bring EU MemberStates much closer to their targets for smart,sustainable, and inclusive growth. Dependingon the characteristics and size of the population,the first and second generation are amajor target group for improving outcomes insociety, especially in the EU’s well-establishedcountries of immigration. The project’s analysisreport calculated these estimates for the foreign-bornfor Europe 2020 headline targets onemployment, early school leaving, and at-riskof-poverty-and-social-exclusion.Closing theemployment gap could account for around 10 %of the Europe 2020 targets across EU MemberStates. Half-a-million young people could beprevented from leaving school early (or 8.7 %of all early school leavers in the EU). 3.3 millionpeople could be lifted out of poverty or socialexclusion (or 5 % of all those currently at risk inthe EU). These calculations may be repeated fordifferent immigrant groups, such as foreigners,third-country nationals, the first and secondgeneration,women, and children. The calculationsidentify the areas and the countries whereimmigrants are a major target group for generalpolicies. These estimates initiate a datadrivenprocess of mainstreaming with the aim49

to create general policies that improve the situationof immigrants and other groups in society.11.3.3. Third option: Member Statesset their own specific national integrationgoals and voluntary targets.Ultimately, policymakers can use indicators toset targets for themselves. The first two optionsbuild the case that various general policies arerelevant for integration and affecting immigrantsas a major target group. Policymakers may cometo the conclusion that they want new policies andestimates of their potential effects. How muchwill a new reform raise immigrants’ employmentrate or reduce early-school leaving among immigrantpupils? Through targets, Member Statesdefine their policies in concrete objectives andmake the links between integration and generalpolicies. Target-setting is a demanding step inevidence-based policymaking. A well-researchedand collaborative process of target-setting canmotivate stakeholders and implementing agenciesto take part in the new policy and maximiseits effectiveness. In contrast, a poorly-plannedprocess may result in no targets or in self-defeatingtargets that undermine public confidencein integration policy. For example, criteria for selectingtarget areas and levels were summarisedin MPG’s 2010 study for the European NetworkAgainst Racism, entitled Target-setting for improvingthe socio-economic situation of migrantsand ethnic minorities in Europe.Two types of targets are designed to addressinequalities. In the first option, the target establishesa minimum standard for the total populationto ensure that no one is left behind. Forexample, the EU employment rate target setthe same level (75 %) for men and for womenaged 20-64 and uses this level as a referencefor youth, older workers, low-skilled workers, andimmigrants. The second option, ‘closing the gap,’aim to narrow the performance gap betweena group in the population (e.g. immigrants) andthe total population. Targets to reduce inequalitiessend a strong message about the governments’commitment to improve integrationthrough a convergence of societal outcomesover time. Reducing inequalities is a challengingtarget, since the situation of the disadvantagedgroup will have to improve faster than that ofthe rest of the population. In both cases, targetscan control for the key demographic differencesbetween the immigrant and non-immigrant population.These targets aim to reduce the differencesbetween immigrants and non-immigrantswith similar characteristics, usually age, gender,and possibly household composition and socioeconomicstatus. Statistical controls ensure thatintegration targets are realistic.The best available analysis and policy evaluationsare useful for policymakers to set their ownspecific national integration goals and voluntarytargets. Migrant-specific targets are harder to setwithout this evidence-base. When setting theirown targets, Member States can learn from othercountries, regions, and cities that have experiencewith integration targets. EU mutual learningmechanisms provide a space for Member Statesto learn from each other and receive financialand technical support. This project’s seminarshave developed a constituency of stakeholdersinterested in the EU integration indicators. Thesestakeholders can meet to discuss areas for mainstreamingand target-setting. These debates canoccur within Member States or between countrieswith similar migration histories and nationalcontexts on integration. At EU level, this processwould require greater cooperation between differentministries, EU Commission services, and civilsociety. The ‘Joint Assessment Framework’ mentionedearlier in this report shows how EU mutuallearning supports Member States in using indicatorresults and prioritising certain areas for futureaction and analysis.50

Annex1. List of the most relevant factors for immigrant integrationMigrant population (individual/group level)General context and policies(macro/ society level)Migration and integration policies(policy level)Categories Variables Categories Variables Categories VariablesSocio-demographic Labour market structures Migration policiesn Agen Gendern Country of birth(Level of development)n householdn compositionn Citizenshipn Year of residencesince immigrationn Family status(married, children)n Share of servicesn Share of producingsectorn Public sectoremploymentn Public spendingn Trade uniondensityn Minimum wagen Channel ofmigrationn Size of the migrantpopulationSocio-economic Education system Integration policiesn Educationalattainmentn Employmentn Occupationn Incomen Socio-economicstatus of parentsn Trackingn Expendituren Number of schoolformsn Early childhoodeducation andcaren Socio-economicsegregation inschoolsn Labour marketpoliciesn Migrant educationpoliciesn Naturalisationpoliciesn Anti-discriminationpoliciesSocio-culturaln Mother tonguen Language skillsn Language spokenat homen Subjectivefactors 28Social welfare systemn Social benefitsn Welfare generosityn Social spendingEqualityDiscriminationn Income inequality(Gini Coefficient)n Gender equalityn Awareness ofdiscriminationn Experience ofdiscriminationNote: This non-exhaustive list is an overview of relevant and measurable factors that can influence immigrant integration outcomes.It is based on the research reviewed and conducted in the context of this project.(28) Such as ‘sense of belonging’ and ‘institutional trust’.51

2. Closing the gap - scenario’ for the Europe 2020 headline indicator ‘early school leaving’Total Population% 2010foreign-born %rates 2010Europe 2020Target in %(2011)The numberof peoplethat would beprevented fromleaving schoolearly given the‘no migrant gapscenario’% of allpeople thatwould beprevented fromleaving schoolearly given the‘no migrant gapscenario’EU2514 26 10 501161 8.7 30.6(w/o RO, SL)Belgium (b) 12 22 9.5 12525 11.2 53.9Bulgaria 14 : 11 0.0 0.0Czech Republic 5 13 3851 8.0 8.0Denmark 11 17 9.9 3832 7.5 75.2Germany 12 24 9.9 94337 11.6 66.1Estonia 12 : 9.5 0.0 0.0Ireland 10 12 8 1292 3.3 16.5Greece 14 44 9.7 36138 30.1 97.9Spain 28 43 15 101943 10.3 22.1France 13 25 9.5 55567 7.4 27.6Italy 19 41 15.5 100874 12.2 66.5Cyprus 13 26 10 2550 22.9 99.4Latvia 13 : 0.0 0.0Lithuania 8 : 0.0 0.0Luxembourg 7 10 379 13.1 13.1Hungary 10 18 10 3413 3.8Malta 37 : 29 0.0 0.0Netherlands 10 12 7.9 3237 2.3 10.8Austria 8 21 15609 27.0 27.0Poland 5 : 4.5 0.0 0.0Portugal 29 27 10 -1522 -0.6 -0.9Romania 18 : 11.3Slovenia 5 20 5 2141 24.2Slovakia 5 :Finland 10 21 8 3093 6.7 33.7Sweden 10 12 9.9 2390 2.7 273.9United Kingdom 15 10 x -36390 -4.1Share of ‘nomigrant gap’of reaching theEU2020 targets,in %52

3. ‘No gap scenario’ for the Europe 2020 headline indicator ‘employment’Total Population% 2010foreign-born %rates 2010Europe 2020Target (2011)The numberof peoplethat would beemployed giventhe ‘no migrantgap scenario’% of allpeople thatwould beemployed giventhe ‘no migrantgap scenario’EU2569 64 75 1825347 0.9 10.7(w/o RO, SL)Belgium (b) 68 55 73.2 146682 3.7 48.6Bulgaria 65 54 76 5025 1.5 8.7Czech Republic 70 70 75 0 0.0Denmark 76 65 80 43494 1.7 33.2Germany 75 66 77 717973 1.9 72.4Estonia 67 60 74 9359 1.7 16.1Ireland 65 63 70 8774 0.5 6.4Greece 64 67 70 -29586 -0.7 -7.1Spain 63 60 74 152663 0.8 4.8France 69 60 75 467981 1.8 20.6Italy 61 65 68 -156672 -0.7 -6.1Cyprus 75 74 76 1149 0.3 22.7Latvia 65 61 73 8392 0.9 7.5Lithuania 64 63 72.8 1374 0.1 0.8Luxembourg 71 74 73 -3819 -1.7 -61.0Hungary 60 68 75 -24611 -0.7 -2.6Malta 60 62 62.9 -432 -0.3 -5.7Netherlands 77 65 80 175896 2.3 58.0Austria 75 68 77.5 68242 1.8 53.0Poland 65 54 71 11260 0.1 0.8Portugal 70 73 75 -19788 -0.4 -6.0Romania 63 74 70 0Slovenia 70 66 75 7963 0.9 12.1Slovakia 65 57 72 0Finland 73 64 78 16429 0.7 10.2Sweden 79 64 82 148695 3.4 90.7United Kingdom 74 69 x 267322 1.0Share of ‘nomigrant gap’of reaching theEU2020 targets,in %53

4. Correlations across all EU migrant integration indicators by genderNote: Orange circles indicate where strong relationships exist between migrant integration indicators from different areas. Forexample, we see here a strong positive correlation between the gap in poverty risk (after social transfers) between immigrantsand non-immigrants and the gap in primary education. This means that countries with immigrants are poorer compared to nonimmigrants,they also more often fail to complete primary education.The pies show the correlations across the indicators, measured as gaps of the foreign-born population compared to the totalpopulation. Red indicates negative correlations (the higher one indicators, the lower the other) and blue positive correlations(the higher one indicators, the higher the other). Higher correlations across areas are circled in orange. Correlations below0.5 (e.g. ) can be ignored as well as higher correlations between indicators within the same areas. This calculation is providedby David Reichel from the International Centre for Migration Policy Development.54

5. Overqualification rate of persons aged 25–54 by groups of country of birth and gender, 2008Source: Eurostat 2011, LFS 20086. People at risk of poverty or social exclusion aged 25–54 by groups of country of birthand selected household type, EU-27, 2008Source: Eurostat 2011, LFS 20087. Net median income ratio of foreign-born women and the share of foreign-born womenin 2-person households with children in 2009Source: Bivariate analysis for the purpose of this project55

8. Native-born offspring of immigrants employed in the public sector by level of education,persons aged 15 to 34, 2008 Difference with the offspring of native-born in percentage pointsSource: OECD 20129. Overlaps between at risk of poverty, material deprivation and low work intensity, by countryof birth, in the EU (thousands of individuals), 2007 income yearSource: Lelkes/ Zólyomi 2011, European Centre, calculations based on SILC 2008 data56

10. Reading performance by immigrant status, before and after accounting for socio-economicbackground:11. Proportion all students in PISA-samples not reaching the minimal level two for readingand the Hutchens index for social segregation (calculated on PISA 2009 data)Source: Bivariate analysis for the purpose of this project57

12. Gap in socio-economic position between foreign-born and natives and gap betweenforeign-born low achievers and general population for readingSource: Bivariate analysis for the purpose of this project13. Share of the overall population holding a tertiary degree (LFS, 2009) and public expenditureon education in percentage points of the GDP (2009)Source: Bivariate analysis for the purpose of this project58

14. Gaps by part of services in total female employment (2009).Source: Bivariate analysis for the purpose of this project15. Gaps female activity rates (2010) by level of awareness of discriminationSource: Bivariate analysis for the purpose of this project59

16. Female gap in poverty risk rates of the foreign-born and the female gap in socialbenefit recipients, 2009Source: Bivariate analysis for the purpose of this project17. Female gap in poverty risk rates (2009) and the discrimination awareness index (2006)Source: Bivariate analysis for the purpose of this project60

18. Female gap in poverty risk rates and the level of social spending in % of GDP in 2009Source: Bivariate analysis for the purpose of this project19. Gap between % low achieving foreign students and national students in reading (PISA, 2009)and MIPEX III score for education:Source: Bivariate analysis for the purpose of this project61

20. Eurostat bivariate correlation between naturalisation rates and policies21. Percentage of representatives with a migrant backgroundSource: Kirchberger et al. (2011), Becoming a Party of Choice: a Tool for Mainstreaming Diversity, Migration Policy Group.62

22. Availability of indicators for third country nationals aged 25 to 54 from Eurostat Pilot Study63

23. Different elements of national systems of integration indicatorsAustria (‘Migration und Integration 2011’): Austria presents basic descriptive analysis of a wide range of additionalfactors including ‘perception of immigrants and the host-society’, ‘life expectancy’, ‘sense of belonging’, ‘perceptionof discrimination’, ‘inter-ethnic contacts’, ‘inter-ethnic marriages’, ‘share of foreign students’ and ‘ crime rates’. Thereport breaks down results by the dominant immigrant groups (EU/EEA, Turkey, Former Yugoslavia). Employmentoutcomes are presented by occupational sector. The report includes profiles of the share of foreigners in each land.Denmark (Effectiveness measurement of Danish municipalities’ integration policies & ‘Integration Barometer’):The first report is a benchmarking system comparing municipalities’ success with newcomers (e.g. refugees andreuniting families). Taking migration and labour market factors into account, this system allows monitoring in distinctregions and an evaluation of a specific state integration policy (3-year integration programme). The integrationbarometer consists of a national and 98 local barometers – one for each municipality. The barometer measuresnational developments towards the government’s nine aims for integration.Estonia (Estonian Integration Monitoring 2011, summary in English): The report was based on an opinion polland qualitative focus groups with five different immigrant and non-immigrant target groups. The report develops anintegration index and organises results based on immigrants’ ‘degree’ of integration.France (Immigrés et descendants d’immigrés en France 2012): Since the publication of the EU integration indicators,the National Institute of Statistics and Economic Studies (INSEE) published a national report on these and otherindicators. In the first chapter, researchers present additional longitudinal and multivariate analysis in four areas.The main datasheets cover a wide range of areas, some regional differences, and different groups by gender, age,generation, country of origin, duration of residence, age at arrival, and household composition. Comparisons aremade with non-immigrant men and women.Germany (‘Integrationsindikatorenbericht’ 2012): The report examines broad groups, such as foreigners, personswith a migrant background and persons with a migration experience. Results are further dis-aggregated by age,gender, education, EU born foreigners. Beyond descriptive analysis, further multivariate analysis examines the influenceof factors such as age, gender, marital status, educational background (what and where), regional labourmarket situation (such as average income levels), years of experience in work, length of residence in the country(proxy for language skills) and social contacts/neighbourhood.Ireland (‘Annual Monitoring Report on Integration 2011’): This civil society report provides a descriptive analysis ofIreland’s results on the EU indicators, among other indicators. It also uses other research to help interpret the results.It discusses policy issues emerging from the results and provides the legislative context.Netherlands (‘Annual Report on Integration 2010’): The Dutch report includes results for the traditionally fourlargest non-western groups in the Netherlands: Turks, Moroccans, Surinamese and Antilleans and mainly comparesthe situation of the second generation in comparison to the first and the non-immigrant population. The Dutch report isa summary of several in-depth thematic studies. It captures additional indicators such as ‘crime rates’, ‘neighbourhoodsegregations’, ‘uptake of income support’, ‘sense of belonging’, ‘social contacts’, ‘involvement in community organisations’,‘self-perceived health’ and ‘use of health facilities’. Further statistical analysis is featured in thematic reports.Norway (‘Immigration and Immigrants 2010’): The report breaks down descriptive results by age, duration of residence,gender, region of birth, channel of migration (e.g. refugees), nationality and employment sector. Each sectionends with certain policy challenges and recommendations. In 2005 the Norwegian parliament introduced a set ofgoals with progress indicators, “Goals for Social Inclusion of the immigrant population”, as a means of actively monitoringand steering its integration policies. There are 17 goals for social inclusion, with a total of 8 ministries being responsiblefor achieving them. In order to follow developments and identify the results of policy, the sector ministries mustreport annually on the achievement of the goals in accordance with the sector responsibility principle. These reports arethen combined into a single report by the coordinating ministry in the yearly budget proposal presented to Parliament.Sweden (‘Pocket Facts 2010: Statistics on Integration’): The report presents, inter alia, participation rates overtime (1987-2009), outcomes by gender, age, residence duration, employment sector, region of birth, refugees.64

24. Different types of national integration indicatorsEMPLOYMENT AT DK FR DE IE NL NO SE EU OECDEmployment Rate √ √ √ √ √ √ √ √ √ √Unemployment rate √ √ √ √ √ √ √ √ √ √Activity Rate √ √ √ √ √ √ ●Overqualification √ √ √ √ √ √Self-employed √ √ √ √ √ √ √ √Youth Unemployment √ √ √ √ √ √Long-term unemployment rate √ √ √ √ √Training measures √ √ √Public sector employment √ √ √ √ √Sector/occupation √ √ √ √Temporary/ Part-time work √ √ √Promotion rateUnsocial working hours√EDUCATION AT DK FR DE IE NL NO SE EU OECDPre-Primary Education √ √ √ √ √ √Primary Education √ √ √ √ √ √Grade Repetition Rate √ √Early Leavers/ Drop Out-Rate √ √ √ √ √ √ √ √ √Low-achieving 15 year olds in reading,√ √ √ √ √ √maths and scienceHighest Educational Attainment √ √ √ √ √ √ √ √ √ √Tertiary Education √ √ √ √ √ √ √ √ √ √Not in employment, education,√ √ √ √or trainingTransition rate into training or jobs √ √ √ √ √School segregation√Language skills √ √ √Language programmes √ √ √ √ √School expectations of parentsParents' involvement in schools √ √√Social inclusion AT DK FR DE IE NL NO SE EU OECDMedian Income √ √ √ √ √ √ √ √At risk of poverty √ √ √ √ √ √Child povertyMaterial deprivation √ √ √ √Financial inclusion indicatorsLow-work-intensity households √ √ √ √in-work at risk of povertySocial benefit uptake √ √ √ √ √ √Poverty reduction rate √ √Criminal convictions √ √ √√√√√65

Social inclusion AT DK FR DE IE NL NO SE EU OECDHealthSelf-perceived health √ √ √ √ √ √Unmet health needs √ √Life expectancy√Obesity √ √ √Health consultation rate √ √Health insurance√Preventative care uptake √ √Self reported health issues √ √ √HousingProperty owners √ √ √ √ √ √ √Rent √ √Overcrowding √ √ √ √spatial segregation √ √ √ √ √Housing cost overburden √ √Social housing √ √ √ACTIVE CITIZENSHIP AT DK FR DE IE NL NO SE EU OECDElected representatives √ √ √ √Naturalisation √ √ √ √ √ √ √Permanent residence √ √ √Election turn out √ √ √ √Volunteering √ √ √Eligible voters √ √ √Membership √ √ √ √ √Leadership in civil society√Diversity in public institutions√Marriage patterns √ √ √ √Sense of belonging √ √ √Institutional trust √ √ √Social contacts √ √ √Cultural values √ √Perceived discrimination √ √ √ √ √Effective equality measures√Note: This is an indicative list of the types of indicators covered in country but not exhaustive or representative list of all indicatorscurrently covered in the countries below. Any updates or corrections are welcome.66

Eichhorst, Werner et al. (2011): The Integration of Migrants and its effects on the labour market. IZAResearch Report No. 40., study conducted for the European Parliament.European Commission (2008): Employment in Europe 2008. Directorate-General for Employment,Social Affairs and Equal Opportunities.Eurostat (2010a): Report on the quality of the LFS ad hoc module 2008 on the labour market situationof migrants and their descendants. Draft report. (Doc. Eurostat/F2/EMPL/22/10).Eurostat (2010b): Indicators of immigrant integration. A pilot study.Eurostat (2011a): Migrant in Europe. A statistical portrait of the first and second generation.Eurostat (2011b): Labour force survey in the EU, candidate and EFTA countries. Main characteristicsof national surveys, 2009. Eurostat Methodologies and Working papers.Eurostat (2011c): Quality report of the European Union Labour Force Survey 2009. EurostatMethodologies and Working papers.Eurostat (2011d): 2009 Comparative EU Intermediate Quality Report. Version 3 – July 2011.Fleischmann, Fenella et al. (2007): The effects of social and labour market policies of EU-countrieson the socio-economic integration of first and second generation immigrants from different countriesof origin, EUI working paper RSCAS 2007/19.Goedemé Tim (2010): The standard error of estimates based on EU-SILC. An exploration through theEurope 2020 poverty indicators. CSB Working Paper No 10 / 09Heath, Anthony (2009): The labour market integration of the children of immigrants. Main determinantsof educational and labour market outcomes. Conference paper. The European Commission andthe OECD, 1st October 2009, Brussels.Huddleston, Thomas et al (2011): Migrant Integration Policy Index.Kirchberger et al. (2011): Becoming a Party of Choice: a Tool for Mainstreaming Diversity.Kraler, Albert; Reichel, David (2010): Statistics on Migration, Integration and Discrimination inEurope. PROMINSTAT Final Report. Vienna: ICMPD.Lanzieri, Giampaolo (2010): Fewer, older and multicultural? Projections of the EU populations byforeign/national background, Eurostat, Methodologies and working papers.Lelkes, Orsolya (2007): Poverty Among Migrants in Europe, European Centre, Policy Brief April 2007.Liebig, T. and S. Widmaier (2009): “Children of Immigrants in the Labour Markets of EU and OECDCountries: An Overview”, OECD Social, Employment and Migration. Working Papers, No. 97, OECDPublishing.OECD (2007): The Programme for International Student Assessment (PISA). PISA 2006: ScienceCompetencies for Tomorrow’s World Executive Summary.OECD (2010): Equal Opportunities?: The Labour Market Integration of the Children of Immigrants,OECD Publishing.OECD (2010): Closing the Gap for Immigrant Students: Policies, Practice and Performance, OECDReviews of Migrant Education, OECD Publishing.68OECD (2011), Naturalisation: A Passport for the Better Integration of Immigrants?, OECD Publishing.

OECD (2012): Settling in. OECD indicators of migrant integration. OECD Publishing.OECD (2012): PISA 2009. Technical Report. OECD Publishing.Marilyn, Thomas (2008): Improving Migrant Participation in the Labour Force Survey: Non-Responseand Attitudes of Non-English Speaking Migrants to Participation. In: Survey Methodology Bulletin, No. 63,September 2008, Office for National Statistics, pp. 39-51.Morales, Laura/ Giugni, Marco (ed.) (2011): Social Capital, Political Participation and Migration inEurope. Migration, Minorities and Citizenship. Basingstoke: Palgrave, 2011.Münnich, Ralf et al. (2011): Policy Recommendations and Methodological Report. Deliverables 10.1and 10.2. AMELI Project.Niessen, Jan / Huddleston, Thomas (2009): Legal Frameworks for the Integration of Third-CountryNationalsNiessen, Jan/ Huddleston, Thomas (2007): Setting up a system of benchmarking to measure thesuccess of integration policies in Europe. Report for the European Parliament, Citizens Rights andConstitutional Affairs, Directorate-General Internal Policies, IP/C/LIBE/ST/2005-93Nusche, Deborah (2009): What Works in Migrant Education? A Review of Evidence and Policy Options.OECD Education Working Paper No. 22.Reichel, David (2011): Do legal regulations hinder naturalisation? Citizenship policies and naturalisationrates in Europe. EUI Working Paper RSCAS 2011/51. Robert Schuman Centre for Advanced Studies,EUDO Citizenship Observatory.Schlotter, Martin et al. (2010): Econometric Methods for Causal Evaluation of Education Policies andPractices: A Non-Technical Guide, Ifo Institute and IZA. Discussion Paper No. 4725Smeeding, Timothy M et al. (2009): Income Poverty and Income Support for Minority and ImmigrantChildren in Rich Countries. Institute for Research on Poverty. Discussion Paper no. 1371-09.Swedish Presidency of the European Union (2009): Malmö Conference Conclusions, 14-16December 2009,http://www.se2009.eu/polopoly_fs/1.28600!menu/standard/file/Indicators %20and %20monitoring%20of %20outcome %20of %20integration %20policies.pdfVink, Maarten Peter (2013): Immigrant naturalization in the context of institutional diversity: policymatters, but to whom? European Institute/ University of Maastricht, Working Paper.Zimmermann, Klaus et al. (2011): Study on Active Inclusion of Migrants, commissioned by the EuropeanUnion Programme for Employment and Social Solidarity - PROGRESS (2007-2013).69

Glossary 29Active Citizenship: Active citizenship is an umbrella term for the acquisition and exercise of rightsfor political participation. As such, it includes citizenship and residence, membership in (political) organisations,voting, running for office, volunteering or participation in political protest. In some cases,it can include subjective indicators such as ‘sense of belonging’, ‘institutional trust’, ‘awareness ofdiscrimination’ and ‘public attitudes towards immigration’.Activity rate: The share of unemployed or employed persons as a percentage of the total populationof the same age group.At risk of poverty or social exclusion rate: The number of persons who are at risk of poverty orsocial exclusion according to at least one of the three following dimensions: at risk of poverty aftersocial transfers; severe material deprivation; living in a very low work-intensity household.At risk of poverty rate: The share of population with net disposable income of less than 60 per centof national median.Citizenship: The particular legal bond between an individual and his or her State, acquired by birth ornaturalisation, whether by declaration, choice, marriage or other means according to national legislation.Country of origin: The country (or countries) which are a source of migratory flows and of which amigrant may have citizenship. In refugee context, this means the country (or countries) of nationalityor, for stateless persons, of former habitual residence.Discrimination: There are different forms of discrimination: 1. Direct: Where one person is treated lessfavourably than another is, has been or would be treated in a comparable situation on grounds ofracial or ethnic origin. 2. Indicrect: Where an apparently neutral provision, criterion or practice wouldput persons of a racial or ethnic origin at a particular disadvantage compared with other persons, unlessthat provision, criterion or practice is objectively justified by a legitimate aim and the means ofachieving that aim are appropriate and necessary.Early leavers from education and training: The share of persons between the ages of 18-24 underthe following two conditions: the highest level of education or training attained is ISCED 0, 1, 2 or3c short and respondent declared not having received any education or training in the four weekspreceding the survey.Elected representatives: The share of immigrants among elected representatives.Employment rate: The share of persons between 20-64 who, during the reference week, performedwork, even for just one hour, for pay, profit or family gain or were not at work but had a job or businessfrom which they were temporarily absent because of, e.g. illness, holidays, industrial dispute,education or training.EU-15/ EU 12: EU-15 countries are the EU Member States before enlargement in 2004. EU-12 includesEU Member States that entered the Union in 2004 and 2007.Foreign (or immigrant background) background: A person with parents where at least one parentwas born outside of the country of residence. This includes the first and second generation as well asthe naturalised.Foreign-born: A person whose place of birth (or usual residence of the mother at the time of thebirth), is outside the country of his/her usual residence.70(29) Where possible, official definitions were taken from Eurostat and the European Migration Network.

Foreigner: (non-nationals) A person who are not citizens of the country in which they reside, includingpersons of unknown citizenship and stateless persons.Health status: The share of population perceiving their health status as good or poor.Highest educational attainment: The level of education is defined in accordance with the 1997International Standard Classification of Education (ISCED 1997) and is often aggregated into threelevels. Low: below the second cycle of secondary education (up to ISCED level 3c short). Medium:second cycle of secondary education (ISCED levels 3–4 other than 3c short). High: tertiary education(ISCED levels 5–6).Immigrant (or ethnic) penalty: The disadvantage of an immigrant or ethnic groups that can be exclusivelyattributed to the immigrant status rather than any other characteristics.Immigrant: See foreign background. In the EU context, a person undertaking immigration. Generalterm that includes persons with a foreign background (at least one parent born outside the countryof residence).Long-term residence: The share of immigrants holding permanent or long-term residence permits.Low-achieving 15-year-olds in reading, mathematics and science: The number of influence whoproficient below Level 2 of the assessment. PISA covers influence who are aged between 15 years3 months and 16 years 2 months at the time of assessment and who have completed at least 6 yearsof formal schooling, regardless of the type of institution in which they are enrolled and of whetherthey are in full-time or part-time education, of whether they attend academic or vocational programmes,and of whether they attend public or private schools or foreign schools within the country.Median net income: Income from work, comprising employee income and self-employment income;(b) property income, including interest, dividends, profits from capital investment in an unincorporatedbusiness; (c) income from rental of a property or land; (d) pensions from individual private plans;(e) income from social benefits, taking into account unemployment benefits, old-age benefits, survivorbenefits, sickness and disability benefits, education-related allowances, family- or child-relatedbenefits, social exclusion allowances and housing allowances; and (f) regular inter-household cashtransfers received.Migrant integration: A dynamic, two-way process of mutual accommodation by all immigrants andresidents of Member States’. The integration process involves immigrants, who exercise their rights andresponsibilities in relation to their new country of residence, and the receiving society, which should createthe opportunities for immigrants’ full economic, social, cultural, and political participation. Migrantintegration can be measured by the long-term convergence across a wide range of common socialindicators taking into account different background characteristics.Naturalisation rate: The share of immigrants that have acquired citizenship. This can be measuredin annual naturalisations rates or as the rate of naturalised immigrants as a share of all immigrantsliving in the country.Naturalisation: Any mode of acquisition after birth of a nationality not previously held by the targetperson that requires an application by this person or his/her legal agent as well as an act of grantingnationality by a public authority.Non-EU immigrant: Any person not having the nationality of an EU Member State or a person thatwas born outside the EU.Non-EU national: Third-country nationals, persons who are usually resident in the EU-27 and who donot have the citizenship of an EU-27 Member State.Non-national EU citizens: Persons who have citizenship of an EU-27 Member State and who are usuallyresident in another EU-27 Member State.71

Outcome: In the context of migrant integration indicators, the term ‘outcome’ describes a (statistical)result of a certain indicator, usually measured in rates. Outcomes are compared between immigrantsand non-immigrants (gaps). In general, outcomes of indicators can describe the situation of integrationof immigrants in a certain area of society.Overcrowding rate: The number of rooms in the house to the number of people. A person is consideredas living in an overcrowded dwelling if the household does not have at its disposal a minimumnumber of rooms equal to: one room for the household; one room per couple in the household; oneroom for each single person aged 18 or more; one room per pair of single people of the same genderbetween 12 and 17 years of age; one room for each single person between 12 and 17 years of ageand not included in the previous category; one room per pair of children under 12 years of age.Overqualification rate: share of the population with a high educational level (i.e. having completedtertiary education, ISCED 5 or 6), and having low- or medium-skilled jobs (ISCO occupation levels 4 to 9)among employed persons having attained a high educational level.Property ownership: Ratio of property owners to non-property owners among immigrants and thetotal population.Share of 30-34-year-olds with tertiary educational attainment: The share of the immigrant populationthat has achieved ISCED level 5 or 6.Social exclusion: In the EU context, a situation whereby a person is prevented (or excluded) fromcontributing to and benefiting from economic and social progress.Social inclusion: In the EU context, a framework for national strategy development, as well as for policycoordination between the Member States, on issues relating to tackling poverty and social exclusion.Social inclusion is an umbrella term that covers the inclusion of groups in society in regards to income,poverty, health and housing.Socio-economic background/ socio-economic status: Economic and sociological combined totalmeasure of a person’s position in relation to others in society, usually measured by income, occupationand education.Unemployment rate: The share of persons between 20-64 who were without work during the referenceweek, were currently available for work and were either actively seeking work in the past fourweeks or had already found a job to start within the next three months.72