What is known about immigrants’ integration into European society? To answer this question, governments need reliable and timely data, but most importantly, a representative immigrant sample.
Dr. Amparo González and Dr. Inmaculada Serrano, of the Spanish National Research Council (CSIC), explain how population registers can help and why better cross-country research is needed.
A hot political and social debate around immigration has taken centre stage during the last years in Europe (and elsewhere). Numbers are important in this debate: how many immigrants are there, how many are arriving? But questions about the incorporation of immigrants into European societies are even more relevant: what are they like? In what conditions do they arrive, live and settle down (if they do)? How do they (and their descendants) contribute to the economy? How do they think and behave? Answering such questions requires accurate, reliable and comparable data.
The UN’s
2030 Agenda for Sustainable Development (SDGs) underlines the contributions of well-managed migration for inclusive growth and sustainable development, but the Global Compact for Migration (GCM) reiterates the crucial importance of data for well-informed policy-making. In fact,
Objective 1 of the GCM calls on signatory countries to improve the collection, analysis and dissemination of migration data.
Understanding how integration unfolds requires data of different types, including behavioural and attitudinal data. Moreover, we need a cross-country and across time (longitudinal) perspective in order to assess how policies affect integration. Surveys are the most fitting instrument to produce such data, but their quality depends on two crucial aspects: achieving a representative (unbiased) sample of immigrant populations, and succeeding in contacting and interviewing the selected individuals. Population registers, in countries where they exist, are key for both goals.
The role of population registers
The gold standard for immigrant sampling is using population registers as a sample frame to conduct random sampling. The reason is that population registers are the most updated and comprehensive dataset on a country’s population. However, and despite recent efforts to improve and homogenise migration statistics, population registers sometimes suffer limitations that hinder or make drawing representative samples of the immigrant population impossible. In other cases, population registers simply do not exist, but the issues we raise in this post are also largely applicable to contexts in which the Population Census is the best potential sampling frame at hand.
Can we correctly identify immigrant populations in the register?
Immigrant populations can be identified using two main criteria: country of birth (foreign-born vs natives) or citizenship.Another criterion, more rarely used in Europe, is ethnicity. Additional pieces of information that are frequently relevant for sample selection are age and time of arrival.The availability of both variables is particularly desirable because the increasing numbers of naturalised residents, a growing second-generation population or the return of past and recent emigrants often blurs the frontier between immigrants/natives and foreigners/nationals. Furthermore, in order to capture 2nd generation immigrants, registers should also include and make available information about parents (at least their country of birth).
Figure 1. Variables available for the identification of immigrant origin populations in different European Population Registers
Good enough |
Limited |
Very limited |
NL: country of birth + citizenship + "migration background" with country of birth of his/her parents (identify 2nd generation)
DK: SW: country of birth + citizenship + same information for parents and grandparents (identify 2nd and 3rd generation)
|
ES, DE: country of birth + citizenship |
IT: citizenship only |
Legend: NL = Netherlands, DK = Denmark, SW = Sweden, ES = Spain, DE = Germany, IT = Italy
Source: Authors’ own elaboration.
|
Sometimes the legal framework hinders the possibility of identifying part of the population of immigrant origin in the registers. For instance, German data protection law restricts data collection to data that are necessary for public administration. For this reason, the inclusion of the variable “migration background” (including parental information) into the microcensus 2004 0required a specific exceptional regulation.
Is the whole immigrant population registered?
Samples become biased and not representative when the population registers do not include the entire immigrant population. In most cases, population registers exclude irregular immigrants, a segment of the immigrant population of crucial importance for the integration process and the political debate surrounding immigration. Short-term migrants (stays shorter than 12 months) are often also excluded.
Leaving aside these exclusions, a good coverage of the immigrant population also requires that incentives (and no barriers) for registration are in place; this is frequently not the case, particularly for EU-citizens. Spain is a worldwide exception in this regard. Registration is not only mandatory but also incentivized for all types of population, including irregular immigrants: it is needed to access basic public services (including health and education for irregular migrants) and it is accepted as proof of residence for regularization purposes.Moreover, since municipal budget depends on their total population, municipalities also take an interest (and actively campaigned) for registration of immigrants.
Figure 2. Inclusion/exclusion of irregular migrants and incentives for registration in Population Registers
|
Good enough |
Limited and very limited |
Exclusion |
ES: irregular immigrants included and strong incentives (and guarantees): access to basic services, proof of residence for regularization.
DE: almost the entire population living within its territory, regardless of citizenship and right of residence but excluding some categories of paperless migrants.
|
IT, DK, SW: only regular immigrants (with residence permit).
NL: only regular immigrants (with residence permit). Asylum seekers are registered when they get a residence permit or if they have stayed for 6 months in an Asylum Seeker Centre.
|
Incentives |
ES: mandatory for all residents, and strong incentives for registration for all (access to basic services, proof of residence for regularization).
DE: registration is compulsory no later than 2 weeks after moving in a municipality. Fines can be imposed otherwise, and access to social services is impossible without registration.
|
IT: mandatory (for all migrants with residence permits and after 3 months for EU citizens), but no incentives (estimated 15% under-coverage of regular and irregular migrant population).
NL: mandatory for stays of 4 months or more, but no incentives (in particular for EU migrants).
DK: mandatory for stays longer than 3 or 6 months but no particular incentives.
SW: mandatory for stays of 1 year or more but no particular incentives.
|
Legend: NL = Netherlands, DK = Denmark, SW = Sweden, ES = Spain, DE = Germany, IT = Italy
Source: Authors’ own elaboration.
|
Currently, sampling from the population register leads in most cases to selecting only regular and resident migrants, which are the most stable segment of the immigrant population. If access and incentives for registration are provided for irregular and short-term migrants, our capacity to understand migration and integration processes across different contexts would significantly improve.
Are these data accessible for researchers?
Data protection laws and administrative procedures often restrict access to and use of data collected by governments. At one end of the spectrum we have the Nordic countries, like Denmark and Sweden, where population registers are fully available to researchers. Also, researchers can request statistical institutes to link data from other specialized registers and large surveys (such as tax income, labour force survey), which allows refining sample selection and cutting down questionnaires (which improves survey data in a number of ways).
At the other end of the spectrum, we have countries like Italy and Germany, where no central access to the register is possible, and each municipality must be approached individually, making procurement complex, time-consuming and uncertain. This also hinders significantly the sampling methodology, where selection of municipalities is frequently a first needed step.
Figure 3. Access to data in Population Registers to construct sampling frames
Good enough |
Limited and very limited |
DK, SW: fully available to researchers affiliated with national research institutions, and possibility to link with specialized registers (longitudinal information).
NL: direct access to types of migrants defined by legal status or migration motive, and possibility to link with census, specialized registers and large surveys.
ES: available to researchers with official support from some Spanish Public Administration (other than universities). Public universities can benefit from a 15% discount in the final price for the sample. Possibility to link with specialized registers virtually unused so far.
|
IT: no central register, municipalities have to be approached individually, at the discretion of the public officer
DE: no central register, district or towns (several thousand) have to be approached individually to negotiate data release and costs for complex sample selections. This means that geographically extensive sampling is extremely costly and risky. The legal framework does not allow linkage of public registers.
|
Legend: NL = Netherlands, DK = Denmark, SW = Sweden, ES = Spain, DE = Germany, IT = Italy
Source: Authors’ own elaboration.
|
Is it feasible (and affordable) to locate and interview the selected migrants?
Migrants are a relatively small-size population (even more so by specific countries of origin) and also more mobile than the general population. Finding them is particularly costly, since it requires wider geographical coverage for relatively small numbers (especially if the survey focuses on some particular migrant groups). Their higher mobility often results in lower contact rates, worsened by inaccuracies in the registers, for which de-registration is not mandatory or it has few incentives.
Figure 4. De-registration in Population Registers
Good enough |
Limited |
Very limited |
ES: mandatory renewal of registration every 2 (non-EU) or 5 years (EU and permanent residents) since 2006 and 2011. No strong incentives, but systematic cleaning-up operations on an on-going basis.
DK: mandatory re-registration within 5 days of residence change within country. Failure to communicate this may result in lack of access to public services and benefits in new municipality of residence. Mandatory re-registration if moving abroad for 6 months but no incentives. Over-coverage is corrected when absence is identified. (Estimate at 0.97% of foreign-born population).
|
SW: Mandatory de-registration if moving abroad for 6 months but no incentives. Over-coverage (estimated at 4-8% of foreign-born population) is corrected when absence is identified (failure to renew stamp with Migration Agency)
NL: Mandatory de-registration if moving abroad for 8 months but no incentives: Over-coverage is assessed with the help of the local authorities (the data are published separately, and added to the number of registered emigrants).
|
IT: non-mandatory and no incentive for de-registration. Over-coverage is corrected only after every census. |
Legend: NL = Netherlands, DK = Denmark, SW = Sweden, ES = Spain, DE = Germany, IT = Italy
Source: Authors’ own elaboration.
|
Migrants also have, on average, higher vulnerability levels, making it harder to gain their trust to participate (leading to higher refusal rates). Distrust can be aggravated if it becomes clear that data facilitated by the administration are being used to single them out.
Improve the registers and meanwhile, let's collaborate!
Improving sample frames (registers), and data-collection strategies, are a first step to increase the quality and reliability of migration surveys, which constitute a key instrument to understanding integration processes in Europe. Achieving a harmonised data collection on immigrants remains a major challenge, halted by national-level legal frameworks and administrative structures. Given these constraints, researchers should foster cross-country research teams and implement the best possible sampling strategy in each country. Researchers and analysts should also increase reporting and awareness about: caveats for comparability and limits in sample frames and sampling strategies, including contact rates and refusal rates.
Figure 5. Suggested best sampling strategies for individual countries (general immigrant population)
Good enough |
Limited and very limited |
ES: register-based sampling adjusting sampling strategy to diminish costs (e.g. using household as a sampling unit; clustering by neighborhoods; or introducing flexibility in contracts replacement and number of trips allowed.
NL: register-based sampling using two-stage stratified sampling, excluding or merging municipalities with low numbers of immigrants
|
DE: register-based sampling using two-stage approach (but costly due to decentralization) and alternative methods to include also irregular migrants
IT: no registered-based sampling; use Centre Sampling Technique to include irregular migrants and obtain unbiased estimates
|
Legend: NL = Netherlands, DK = Denmark, SW = Sweden, ES = Spain, DE = Germany, IT = Italy
Source: Authors’ own elaboration.
|
Disclaimer: The opinions expressed in this blog are those of the authors and do not necessarily reflect the views of the International Organization for Migration (IOM) or the United Nations. Any designations employed and the presentation of material throughout the blog do not imply the expression of any opinion whatsoever on the part of IOM concerning the legal status of any country, territory, city or area, or of its authorities, or concerning its frontiers and boundaries.