Skip to main content
European Commission logo print header

Migration between Africa and Europe

Final Report Summary - MAFE (Migration between Africa and Europe)

Executive Summary:

International migration from Sub-Saharan Africa to Europe has generated increasing public and policy attention. The flotilla of boats bringing would-be migrants to the Canary Islands, and attempts to reach Spanish territory in Ceuta and Mellila have drawn a rapid response from Europe in the form of new policy measures. Yet the scope, nature and likely development of Sub-Saharan African migration to Europe remains poorly understood, and, as a result, European polices may be ineffective. A major cause of this lack of understanding is the absence of comprehensive data on the causes of migration and circulation between Africa and Europe.

MAFE collected unique data on the characteristics and behaviour of migrants from Sub-Saharan countries to Europe. The key notion underpinning the project is that migration must not only be seen as a one-way flow from Africa to Europe. We argue that return migration, circulation and transnational practices are significant and must be understood in order to design better migration policy. The MAFE project focused on migration flows between Europe, namely Belgium, France, Italy, the Netherlands, Spain and the United Kingdom (UK) and Senegal, the Democratic Republic of Congo (DR Congo) and Ghana, which together account for over a quarter of all African migration to the EU. In each of these 'migration systems', our research seeks to address four key areas:

1. Patterns of migration: trends, migrants' characteristics, migratory routes
2. Determinants of Migration: poverty, education, gender, policies etc.
3. Migration and economic integration: remittances, investments, integration and reintegration of migrants
4. Migrations and families: family, structure and formation, families over time and space.

MAFE investigated three 'migration systems' built around the migration processes of three African populations. The MAFE project presents very original results that go way beyond results of conventional data on international migration. The MAFE survey was designed to allow for the measurement of various migration aspects that were un-measured so far. For instance, it brings answers to questions such as: Has Europe actually become the main destination of African migrants over the last decades? To what extent did irregular migration grow? Do all migrants choose to bring their relatives in Europe? How common are transnational families, i.e. families whose members live apart across borders? To what extent do migrants return in their home country? Do they successfully reintegrate into the labour market after return?

The project results provide a dynamic view of migration. In other terms, it does not simply provide a description of the populations of interest at the time of the survey. It rather takes advantage of the longitudinal nature of the MAFE data to present original results on trends and trajectories. Finally, and maybe more importantly, taking advantage of the transnational nature of the survey, we systematically approached migration with a double viewpoint, looking both at the destination and origin places. For instance, work dedicated to the patterns of migration display new results on trends of departure, but also on trends of return. The work dedicated to migrants' families look at how prevalent transnational families are (i.e. families living apart across borders) and how they function.

Project Context and Objectives:

The scientific objectives and the methodological choices of MAFE are characterised by a multi-sited perspective, in the sense that the project is interested both in sending and receiving countries and also in the various forms of migration (departure, return and circulation), rather than simply in migration from Africa to Europe. This approach rests on three recent theoretical perspectives (migration systems, transnationalism, and new economics of labour migration) that simultaneously emerged in the 1990s to overcome the limitations of the previous theories that were characterised by

1. a unidirectional approach (migration from developing countries to developed countries) and
2. a point of view centred on destination countries.

Changing patterns of African migration

Despite the lack of fully appropriate data, broad patterns of African migration to Europe can nevertheless be sketched:

1. a diversification of the migrants' destinations;
2. the non-marginal extent of return migration;
3. the complexity and changing nature of migration routes to Europe.

Changing destinations of African migrants. Even though the United States of America (USA) and Canada have attracted a growing number of migrants over the last decades (Zeleza 2002; Zlotnik 1993), Europe remains by far the major destination for sub-Saharan migrants (Lucas 2006). However, choices of destination within Europe have evolved. Whereas in the 1960s and 1970s the major migration flows went mainly from the former colonies to the former metropoles (France, UK, Portugal, Belgium), this symmetry has lost some of its relevance, with the growing scale of migration to such countries as Italy, Spain, the Netherlands or Germany (Black et al. 2003; Hamilton 1997; Zlotnik 1996). Spain and Italy have become increasingly popular destinations for Africans since the 1980s, and they are currently the major points of entry of illegal migrants reaching Europe by sea (Van Moppes 2006; ICMPD 2005). Former colonial nations such as France and the UK have also attracted growing numbers of migrants from countries with which they had no colonial links. For instance, France and the UK currently host the largest and the third largest Congolese community in Europe, respectively. The reasons for these changing destination patterns are not fully understood. Present research indicates that destination is determined by a varying balance of choice and constraints, involving 'historic ties with the destination, language, existing networks of migrants, ease of access and transportation and perceptions of economic and social conditions' (Hatton, 2004, p.17). National differences in migration policies are also mentioned to explain not only points of entry but also intra-European migration (Hatton, 2004). But the extent and the patterns of mobility within Europe have rarely been studied. This is a field on which MAFE can bring new evidence through the analysis of migrant trajectories.

Return migration and circulation. As stated above, there is no systematic registration of return migration and this type of flow has so far received little attention in the literature despite its considerable significance for policy (King 2000; Dustmann 1996). Scattered figures indicate nevertheless that return from Europe is not a marginal phenomenon. In Switzerland, 63% of the 1981 immigration cohort returned home between 1981 and 1989 (Dhima 1991). In Germany, 21% of the migrants included in the first wave of the German Socio-Economic Panel (GSOEP) had returned within 6 years (Schmidt 1994). More specifically, in Belgium, between 1970 and 1990, the mean annual number of entries of Congolese nationals was around 1,500 and the number of departures was around 1,200. Even though out-migration is not the same thing as return migration, these figures suggest that the Congolese migration is not simply a one-way flow towards Europe. It appears however that return migration tended to decrease during the 1990s, at least for some African groups such as Congolese or Ghanaians (Black et al. 2003). Several types of return must be distinguished:

1. assisted return (through national or multi-lateral programmes aimed at supporting migrants' return),
2. forced return (expulsion of undocumented migrants, including rejected asylum seekers), and
3. autonomous return.

Actually, the first and second categories represent very few migrants. Although rejected asylum seekers are supposed to leave the country where they made their application, it seems that a high proportion of them do not move (Hatton, 2004). Regarding assisted return programmes, which have been quite largely studied, they never have a massive impact (IOM 2004). The bulk of return migration is thus voluntary, but very little research has been devoted to this question. The MAFE project will shed light on this type of movement by providing new evidence on the scale and the drivers of this flow. Furthermore, it will document the question of migrant circulation on which qualitative research has given interesting insights (see, for instance, Macgaffey et al. 2000 on the circulatory strategies developed by the Congolese), but on which quantitative evidence is rare (Constant et al. 2003) despite growing interest among scientists and policy-makers (Adepoju 2006; Hugo 2003; Vertovec 2007).

The routes of African migrants to Europe. The prevention of illegal migration is a priority in the European agenda. Although cooperative management is sought with the origin countries, the enforcement of external border controls is the major tool of this policy (COM/2005/621 final). Research in migration management is scarce but some results suggest that this type of action is not really relevant. Clandestine entry is thought to be relatively small-scale among undocumented migrants. Significant numbers enter legally and overstay their visas, and rejected asylum-seekers who do not leave the country are another major category of undocumented migrants (Collyer 2006; Düvell 2006). Furthermore, interception statistics indicate that migration itineraries shift over time, in response notably to tightened controls and changing policies in transit countries (Düvell 2006; Väyrynen 2003, Alscher 2005). Itineraries may also change in response to visa policies in transit countries. For instance, the Turkish visa regime for sub-Saharan African countries was changed in 2005, making migration to Europe through Turkey more difficult (Brewer et al., 2006). Migrants' routes appear to be quite complex. Though trans-Saharan and Atlantic itineraries have caught the attention of policy makers and researchers (Collyer 2006; de Haas 2006; Hamood 2006; Van Moppes 2006), some migrants follow other complex routes (some Congolese, for instance, travel through South Africa or Angola to reach Europe by air. See: Sumata 2002). All in all, the current literature gives only an impressionistic view of the migration routes. Although some information on the way African migrants enter Europe and on the changes in their legal status over time was collected for some African countries in the Push-Pull project in the 1990s (Schoorl et al. 2000; Içduygu et al. 2001) more recent and detailed data are not available. MAFE will provide new quantitative evidence on the routes used by sub-Saharan migrants' to enter Europe.

Factors determining African migration.

As stated by Lucas (2006; p.358) 'the literature on the determinants of migration in Sub-Saharan Africa is very extensive', but 'most of the literature looks at internal migration'. By comparison with the theoretical advances and the volume of literature on other parts of the world, the quantitative analysis of the determinants of African international migration, especially to Europe, appears to be a relatively neglected field of research.

Advances in the theories of international migration. Analysis of the determinants of migration has long been dominated by a economic approach in which the migration decision is framed within a cost-benefit calculus. Various macro-economic indicators such as bi-national wage gap, unemployment rate, exchange rates, inflation and so on, have been mentioned as factors underlying varying propensities to migrate internationally. However, it has become quite clear that they cannot fully account for the circularity of migration nor its generalisation over time (Portes et al., 1989; Massey et al. 1997). Interestingly, research based on micro-data has demonstrated the multi-level character of the decision-making process and revealed how tightly the factors explaining migration are interwoven. For instance, human capital appears to be a major determinant of migration. But its effect on the propensity to migrate depends not only on the level of education but also on other individual factors, such as the migrant's legal status (Taylor 1987; Borjas 1993), his/her age, sex, family status at migration, the timing of migration, and the availability of social networks (Cerruti et al. 2001); and on contextual variables, such as the extent to which education is transferable across borders and rewarded at destination (Friedberg 2000), the characteristics of the labour market and income distribution in the sending and host countries (Portes et al. 1985; Zhou et al. 1989; Borjas et al., 1991, Chiswick, 1999), and the admission policy in the immigration country (Chiswick 1987; Borjas 1993; Reitz 1998). In sum, there is increasing recognition that international migration is not a single individual event but rather a dynamic social process that is guided by structural opportunities in sending and receiving societies, the extent of migration networks, and macroeconomic conditions (Massey et al. 1997). In this line, the new economics of labour migration (NELM) have analysed international migration as a household survival strategy, and have put to the forth the idea that an ideal trip to a foreign country is one that does not last very long and that allows the migrants to return to their home communities (Stark 1991).

Factors of migration out of Africa. In the existing reviews of literature, authors agree on four main drivers: demographic pressure, economic difficulties, ethno-political conflicts and ecological deterioration (Lucas 2006, Adepoju 2004). But actually, there are very few empirical analyses to support this broad framework. Hatton et al. (2003) provide one of the rare pieces of evidence on the factors of international migration out of Africa. Using aggregate data, they show that the slower the economic growth in the country of origin, the higher the gap between real wage at home and abroad, the larger the share of population aged 15 to 29, then the greater the rate of out-migration. Lucas (2006) underlines the numerous limitations of this kind of macro-analysis and advocates new analyses based on micro datasets, a requirement to understand the process of migration decision making. Van Dalen et al. (2005) provide a unique contribution on the drivers of African emigration at the micro-level, using the data of the 'Push-pull' project. But, as already mentioned (section 1.1.3) their results are about migration intentions and not actual migration. All in all, the existing literature on the determinants of African international migration is characterised by two gaps. Firstly, there is a missing link between micro and macro-level of analyses. While theories on international migration have highlighted the multi-level character of the migration decision process, studies on African migration remain either micro or macro, do not link these two levels and do not explore meso levels (household, community, region). Secondly, quantitative works on African international migration are focused on out-migration and never consider return migration, although international migration theories (especially the NELM) and empirical results in other parts of the world have shown the significance of temporary migration. As explained before, the aim of the MAFE project is to gather new data in order to document the multi-level nature of the migration decision process and to explore the multiple facets of African migration (departure, return, circulation).

Economic changes: Integration and re-integration of African migrants. The question of how migration trajectories and economic integration interact pertains to two distinct fields of study and theories, depending on whether migrants are observed from the receiving or the sending countries point of view. On the one hand, the question is covered by the literature on integration and social cohesion in developed nations. On the other hand, the issue of the re-integration of returnees concerns the research realm of migration and development.

Integration in destination countries. Looking first at integration and social cohesion, both traditional academic and popular conventional wisdom assumes that over time, immigrants will progressively integrate into their country of destination. Depending partly on ideological perspective, and partly on national context, this is expected to occur either through a process of 'assimilation', involving not only economic improvement but also cultural convergence with 'mainstream' society over time (Alba et al. 1997) or through a more limited economic 'integration' in which economic performance and opportunities begin to approximate those of the wider society, even if immigrants retain or even reinforce aspects of their own national culture or social organisation (Perlman et al. 1997) However, a wide literature now demonstrates that both 'assimilation' and 'integration' often remain an illusion.

It is by now well known that the foreign-born population is at an economic disadvantage compared to natives. Immigrants have higher unemployment levels, jobs of lower quality and receive lower earnings than the native-born population (Alba and Nee 2003). Within the European Union, African migrants clearly exhibit dramatically lower employment rates than European Union (EU) nationals or migrants from other origins (from Balkan countries, for example, whose employment rates are at or over EU nationals' levels) (Commission of the European Communities, 2004). Furthermore, studies have pointed to the 'brain wastage' of African migration, the employment of these skilled professionals being below their level of qualifications once they arrive in Europe (Stillwell et al. 2004). Researchers have argued that the economic difficulties immigrants experience are partly explained by their human capital at arrival in the host county (van Tubergen, 2006) A substantial part of the foreign-born population has little command of the official destination-language (Bean and Stevens 2003), and has obtained educational diplomas in the home country that are not equally valued in the host country. Some studies point also to the possibility that racism may be at play in receiving-country labour markets. More generally, it is hypothesised that institutional differences between receiving countries affect integration of immigrants as well, since large differences exist between destination countries (Inglessi et al. 2004). The difficulty with drawing definitive conclusions is that almost no data exist in which migrants' last employment before migrating is recorded, nor do we have data with accurate estimates of years of schooling completed in migrants' home countries. This is a gap that MAFE aims to fill by collecting migrants' trajectories and contextual data able to describe the context of reception in the various countries.

Re-integration in home countries. The last decade has seen an upsurge of literature on links between migration and development. This includes attention from academics (Van Hear et al. 2003; Lucas 2006), the EU (COM/2005/390) and the United Nations (UN) (GCIM 2005; UN 2006), although much of this remains focused on specific issues such as remittances and the brain drain. In particular, although policy-makers often assume that migrants will return home and contribute to their home economies, relatively little attention has been paid to the extent to which migrants really do gain improved skills and qualifications whilst abroad, nor to the challenges of re-integration (but see Ammassari 2004). In relation to the specific question of whether, or under what conditions, returning migrants might invest in entrepreneurial activity on their return, there is some consensus that important factors include education, work experience, and duration of time spent abroad (McCormick et al. 2001), all factors that are encompassed in the MAFE project, but which are rarely systematically analysed. In addition, some have called for researchers and policy-makers to view re-integration in much the same light as integration - in other words, it is often just as challenging for those returning to their home country to re-integrate as it is for migrants to integrate in host countries (Black et al., 2004).

The MAFE project reflects this literature by treating integration and reintegration as conceptually similar, consistent also with the view of some scholars that the path of migrants increasingly involves 'simultaneous incorporation' in both sending and receiving contexts (Levitt et al., 2004). In this context, social and economic mobility within the host society may translate into investments and projects enhancing both life standards and status in both sending and receiving environments.

Socio-demographic changes: Migration and family formation. Although a significant fraction of contemporary international migration takes place in the context of family chain migration, as suggested by the high number of migrants admitted for family reunification in Europe (especially in the old immigration countries), international migration has rarely been studied from a family perspective (Kofman 2004, Grillo et al. 2008). Various reasons have been put forward to account for such an omission in the academic literature: neglect of the role of the family in economic theory, the interpretation of migration as transactions between individuals and states (Vatz Laaorussi 2001; Zlotnik 1995), its association with female migration and dependency rather than with work and autonomy (Kofman 1999), and the lack of statistical data to trace the recomposition of family groups across time and space (Coleman, 2004). One of the main consequences of the absence of reliable statistical information on the family-related aspects of international immigration has been the unwarranted amalgamation of entries due to family reunification and entries due to family formation, which are distinct forms of family-related migration that usually take place at different stages of the migration process, affect different groups of migrants and, therefore, should be treated differently by legislation and public policy (González-Ferrer 2006). So far, the literature produced in the 'migration and family' field has largely consisted in a comparison of migrants and natives within receiving societies (integration approach). However, recent studies, in line with the transnational approach, help to shed light on the interactions between migration and family formation.

Integration approach. Marriage patterns and fertility levels have often been viewed as indicators of migrant integration. In most developed receiving countries, marriage and parenthood are frequently postponed and divorce is common (Dykstra et al. 2006), whereas migrants usually come from more traditional backgrounds, in which a young age at marriage and childbirth (especially for women) is preferred and divorce is less accepted (Valk, 2007). A key question in the literature is therefore: to what extent are migrants different from the natives in terms of family formation (marriage, divorce, childbearing)? It is generally acknowledged that African migrants have lower rates of fertility than their co-nationals in the country of origin, but higher rates than natives (Bledsoe et al., 2007). This convergence towards receiving society trends is explained by a set of well-established theories: disruption, adaptation, assimilation, selection (Kulu 2005). African migrants are also associated with the image of large (sometimes polygamous) 'traditional families', with a hierarchical organisation. However, various studies have shown that migrants' marriage patterns do not conform to these expectations. For instance, De Valk et al. (2001) show that marriage is less common for migrants from some parts of Africa, like Ghana; and that, if marriage occurs, it is at a relatively later age, and that divorce is more common among migrants than among the Dutch. In France, another example, it appears that single-parent households, headed by women, are quite frequent among the Senegalese (Genereux 1997). It is argued that this can be explained by the difficulty of the migration process, the different institutional context in the destination countries (where the welfare state allows people to be independent from their families) and the break-up of marriages originally contracted solely for the purpose of receiving a residence permit. This last hypothesis raises the issue of the role of family-related events in migration strategies, a question recently tackled in transnational studies of African migration.

Transnational studies. Recent research has shifted the approach of 'family and migration' studies from a destination-centred perspective to a bi-focal viewpoint in order to better understand, on the one hand, the interactions between family formation and migration strategies and, on the other hand, the functioning of transnational families.

1. The 'Transnational vital events' project at the Max Planck Institute for Demographic Research, asks 'how people [African migrants] distribute their temporalised vital events -especially birth and marriage- across international boundaries' with the hypothesis that the timing of these events is related to the migrants' quest for legitimacy in European destination countries (Bledsoe 2004). Although this research project initially aimed to use both qualitative and quantitative approaches, its main results rely on socio-anthropological investigations because of the lack of appropriate data.
2. Other recent research has looked at how migration impacts family structures and has demonstrated the existence of transnational families, of which some members are in the receiving country and some are in the sending country. Pessar and Mahler (2003) identify many of the ways families seek to 'be family' in the transnational context. Parrenas (2005) shows how mothers try to 'mother' via phone calls and gift packages. For some African migrants, polygamy raises specific problems (Lubkemann 2000) and calls for innovative arrangements (Bledsoe et al. 2006). Mazzucato (2005) highlights the double family commitment of some Ghanaian migrants regarding child rearing: while they may be concerned with their children's schooling in the receiving country, they may be responsible for funding the education of another child in their extended family in the sending country.

Finally, the study of the interactions between migration and family raises the question of the position of women within their families (Grillo et al. 2008, Genereux 2007). For some women, migration may mean an increase in social mobility, economic independence, and relative autonomy. This is especially true if women's moves are accompanied by increased participation in the labour market. However, participation in the labour force does not automatically improve equality between a migrant woman and her husband (Kibria, 1996; Darvishpour, 2002). Among other factors, the result depends on both the country of origin (since gender roles differ from one country to another) and the country of destination (since welfare systems are different and more or less supportive to families and especially to women) (Wall and Sao Jose, 2003).

Project Results:

No data was available at the beginning of the MAFE project to provide new scientific evidence on the various topics mentioned above. For this reason, the MAFE project first and foremost aimed at filling a gap in data availability on African international migration. As well as addressing substantive and policy issues, the MAFE project's ambition was also to contribute to advances in data-collection and compilation, and to testing developments in survey methodology on international migration.

Main results achieved with regard to objective 'overcoming lack of data'

The design of the MAFE survey builds on the following key studies on international migration in the world. First, the 'Mexican Migration Project' (MMP), which is a major longitudinal dataset that provided numerous insights into patterns and consequences of Mexican migration to the United States (Massey 1987). Based on the 'lessons learned' from this project, the MAFE project implemented major improvements to the 'ethnosurvey' methodology that was developed by that study, to ensure its applicability to the African and European contexts. Second, recent experience with biographic surveys in Europe and in Africa has provided inspiration for the design of MAFE project questionnaires (GRAB 1999; Poirier et al. 2001; Schoumaker 2006). Third, the project 'Push and Pull Factors of International Migration', a large Eurostat-funded project in the mid-1990's collecting data from selected countries in West Africa, the Mediterranean region and Europe, has provided inspiration in matters of research design and sampling strategy (Groenewold et al. 2004). The collected data provided new insights on various (new) factors explaining migration intentions in migrant-sending countries (Dalen et al. 2005). However, the data do not allow to address, in a satisfactory way, the factors explaining migration realisation (including return and circulation), nor can they be used to assess the potential effect of (migration) policies and of neighbourhood and community-level 'contextual' factors on migration decision making.

The creation of a new dataset on African migration is the foundation stone of the MAFE project. Comparable data have been collected in both sending and destination countries. Importantly, these data are longitudinal - including retrospective migration, education, work and family histories for individuals - and multi-level - linking individual histories to other contextual data in both origin and destination countries. The methodology used in the MAFE project is built upon existing experience. It replicates a survey applied to study Senegalese migration (MAFE-Senegal project, with data collected in Senegal, France, Italy and Spain) on new populations, i.e. Congolese and Ghanaians (with data collected respectively in DR Congo, Belgium, UK; and in Ghana, UK and Netherlands). For financial reasons, samples are not nationally representative, though the project is built upon a progressive strategy, which will make it possible to enlarge the samples in the future.

Questionnaires

The MAFE surveys rely on two different questionnaires that are almost entirely identical from one country to another. The few differences consist in:

1. cultural variables (religion, ethnic groups, matrimonial status etc.)
2. new questions introduced on fostered children in the MAFE Congo and MAFE Ghana biographic questionnaires (module on children)
3. the order of the questions relating to migration in the household questionnaire (module A).

The Household Questionnaire:

1. used only in African countries
2. contains information on the members of the household and also on people who live outside the household and who are related to it (head's children, partners living abroad, other relatives of the head or his/her spouse who live abroad and with whom the household has been in touch within the last 12 months)
3. topics: socio-demographic variables of each individual, short migration histories, remittances, household assets, housing history
4. available in French (MAFE-Senegal, MAFE-Congo) and English (MAFE-Ghana).

The Biographic Questionnaire:

1. used in all African and European countries
2. contains life histories of all the surveyed individuals, whatever their migratory status at the time of the survey (non-migrant, return migrant, current migrant)
3. a grid is used, jointly to the questionnaire, to help the interviewee to recall important dates of his/her history
4. topics: family formation, education and employment, housing, migration, investments (housing, business, community amenities or infrastructure) etc.
5. available in French (MAFE-Senegal, MAFE-Congo), English (MAFE-Ghana), Italian and Spanish (MAFE-Senegal).

Data collection

Data collection was carried out in both sending countries in Africa and destination countries in Europe, in order to constitute transnational samples. At the end of data collection, each team provided a synthetic data collection report. Data collection in the three African countries was coordinated by the African partners of the project: IPDSR in Senegal, the Department of population and development of the University of Kinshasa in DR Congo and the Centre for Migration Studies of University of Ghana. The surveys were conducted in 2008 in Dakar (Senegal), and in 2009 in Kinshasa (DR Congo) and in Accra and Kumasi (Ghana). Two questionnaires were used: A household questionnaire, and a biographic questionnaire. The number of households successfully interviewed varied between 1 143 in Senegal and 1 616 in DR Congo. The number of biographic questionnaires varied between 1 067 in Senegal and 1 666 in DR Congo.

Data collection in the six European countries was organised by the European partners of the project: INED in France, FIERI in Italy, University Pompeu Fabra in Spain, University of Sussex in the UK, University of Maastricht in the Netherlands, and the Catholic University of Louvain (UCL) in Belgium. Surveys were conducted in 2008 (France, Spain, Italy) and 2009 to 2010 (Belgium, the Netherlands, UK). In France, Spain and Italy, survey firms were hired to conduct the survey (CSA in France, DOXA in Italy and Metroscopia in Spain), in close collaboration with the researchers in charge of the project. In Belgium, UK and the Netherlands, the surveys were organised and supervised by the universities. Only biographic data were collected in Europe, using the same biographic questionnaire as the one used in Africa. Overall, approximately 1450 migrants were successfully interviewed (200 Senegalese in France, 198 Senegalese in Spain, and 202 Senegalese in Italy; 138 Ghanaians in the UK and 279 Ghanaians in the Netherlands; 278 Congolese in Belgium and 150 Congolese in the UK).

The main activities related to data collection consisted in organising a pilot survey (only in Africa), implementing the sampling strategy, recruiting and training the survey staff (interviewers, supervisors, editors, coders, data entry staff), carrying out the survey (interviews, editing) and data entry and cleaning. In total, more than 200 people were recruited and trained for data collection and data entry (approximate 120 in Africa and 110 in Europe). Approximately 5 500 biographies were collected (around 4 000 in Africa and 1 500 in Europe), and around 4 000 household questionnaires were successfully completed in Africa. Thanks to the dedication and hard work of the research teams and survey staff, no major difficulty was encountered, although data collection and data entry experienced some delays in several countries.

Due to data entry and cleaning delays, the computation of sampling weights in Ghana, Congo, the UK, Belgium, and the Netherlands was completed in September 2011 (weights had previously been computed for Senegal, France, Italy and Spain). In European countries, this consisted in computing weights by comparing the age-sex distribution of the sample of migrants to an estimated age-sex distribution of the corresponding population. In African countries, sampling weights were computed using information from the sampling design (sampling rates at each stage).

The specific objective during the second period of the project was to produce a workable micro data-base, well documented so that sound analysis could be performed. In fall 2011 (period three), the data-base was ready for use by the analysis teams of the MAFE project. A series of preliminary analysis was ran in fall 2011 and allowed to operationalise the MAFE data-base by cleaning inconsistencies and improving its quality.

Sampling and weighting procedure

An important work on the sampling strategy adopted by the teams was carried out from January to June 2012. It resulted in a methodological note made public on the methodology section of the MAFE website. This note describes the sampling strategy and the computation of weights in the MAFE surveys. MAFE combines random sampling of households and individuals in cities in origin countries, with quota sampling among migrant populations in destination countries. Overall, between early 2008 and early 2010, more than 4 000 household questionnaires and 4 000 individual biographies were successfully completed in three African countries (Senegal, Ghana, RD Congo), and around more than 1 400 biographies of African migrants were collected in six European countries (Belgium, France, Italy, The Netherlands, Spain, UK).

In the three African countries, stratified random samples of households and individuals in the target areas were selected. The target areas were the city of Kinshasa in DR Congo, the city of Dakar in Senegal, and two cities (Accra and Kumasi) in Ghana. In each of the cities, a sampling frame of primary sampling units was prepared, and primary sampling units randomly selected. A listing operation was carried out in each of the selected survey sites to prepare the sampling frame of households. The sampling frame at the second stage (households) was stratified, so that households with return migrants could be oversampled. The number of households successfully interviewed was 1 143 in Senegal (2008), 1 248 in Ghana (2009) and 1 616 in DR Congo (2009).

In each of the selected households, one or several respondents were selected (people aged between 25 and 75, and born in the origin country) for the biographic survey. In Ghana and DR Congo, all the return migrants and partners of migrants currently abroad were selected, and one other eligible member was randomly selected. In Senegal, up to two return migrants and partner of migrants were randomly selected, and another individual was randomly selected. The number of biographies successfully completed is close to the number of households (1 062 in Senegal, 1 243 in Ghana and 1 638 in DR Congo).

In five of the six European countries, no suitable sampling frame was available to select randomly individual respondents (Spain is the exception, with the Padrón). As a result, quota sampling was used. In all the countries, the quotas were set by age and gender at least. In France, the socio-professional category was also included as criteria in the quotas, while in Belgium and the UK, the place of residence was used in the quotas. In most countries sub-regions concentrating the majority of migrants were selected.

Randomness was included in the samples in different ways. For instance, in Belgium, a random sample of places was selected according to the number of people of Congolese origin living in these places. Respondents were selected in these places. The combination of different recruitment methods also ensured that different types of persons had a probability of being included in the sample. For instance, some respondents were recruited in public spaces (street, metro station, hairdresser etc.), others were randomly selected from list of volunteers identified in churches. In France, Italy and Spain, some of the respondents were also selected using the contacts obtained in the household survey in Senegal.

Overall, around 200 migrants were interviewed per origin country in a destination country. It is a little lower in the UK (around 150 per origin country) and higher in the Netherlands and Belgium (around 280). Currently, about 1 450 African migrants have been interviewed in Europe. A new wave of about 400 migrants in Spain was launched in 2010 and 2011 and will be added to the overall sample.

All this work sets the frame for the public opening of the data as of January 2014, providing standardised information to external researchers wishing to access the data. The work already performed on the data ensures a good level of quality. However, the MAFE data remains complex to handle. Plans towards the creation of simplified data-sets or data-sets for training purposes, as well as a series of training events possibly organised in parallel to the main international conferences in the field in 2013 and 2014, are under discussions within the teams (see dissemination plan in final report).

Main results achieved with regard to objective 'advancing knowledge on Afro-European migration'

We present below only summary results extracted from the MAFE policy briefs. More detailed results are available on the MAFE project website. In particular, full working papers on each of the four analytical work packages (WPs) are publically available. They are all constructed around the same chapters: introduction, synthesis analysis for the three migration flow, Congolese chapter, Ghanaian chapter and Senegalese chapter.

WP5: Leaving and returning - the changing dynamics of African migration

WP leader: Bruno Schoumaker, UCL

1. MAFE research suggests that despite restrictive immigration policies in Europe, migration rates to Europe have increased in two of the three African countries studied by the MAFE Project. These policies have also coincided with increasing irregularisation of migrants and - for two countries - decreasing rates of return. However, immigration policies are not the only factors that shape migration; conditions in the country of origin and opportunities in other African countries are also important.
2. Although the overall probability of Africans migrating internationally has increased over the past 35 years, there has not been a consistent rise in the likelihood of migration to Europe. In the case of Senegal and Ghana, the proportion of migrants moving to Europe has increased since the 1970s. However, migration to Europe from the DR Congo has stagnated since the 1990s.
3. Return migration from Europe to Africa appears to be on the decline. The probability of migrants returning from Europe has steeply declined in DR Congo and Senegal since 1975, whilst in Ghana return probability decreased drastically in the 1990s before increasing again between 2000 and 2008.
4. Although the majority of African migrants are legally resident in Europe, irregular migration appears to be on the increase. This is particularly so in the case of Senegal and DR Congo, less so in the case of Ghana and - importantly - also varies according to different European destination countries.
5. African migrants to 'new' European destination countries tend to have different profiles than those who move to more 'traditional' European destination countries. Specifically, they are more likely to be less educated and undocumented.
6. Transit migration is much more common for migration to 'new' European migration destinations, and a significant proportion of migrants appear to be moving within Europe itself, particularly from traditional receiving countries to 'new' destinations.

WP6 : Investigating the role of individual, household and contextual factors

WP leader: Amparo Gonzalez, SCIC

1. MAFE research shows that Africans with at least some tertiary education are more likely to migrate to Europe. However, education levels apparently have little impact on the probability of return.
2. Migration from Africa is selective on the basis of age and sex, but not consistently across all countries. In Senegal and Ghana, those aged under 35 have a higher likelihood of migrating to Europe, but this is not true in the DR Congo. In Senegal, men are more likely to migrate to Europe than women, but there is no significant difference between men and women in Ghana or DR Congo.
3. Initial migration is often affected by the location of family members, friends and acquaintances, re-affirming previous research findings about the importance of social networks in facilitating mobility. Having an adult relative - especially a partner - in Europe was found to substantially increase the probability of migration. However, the significance of nuclear family links should not be overstated as one-quarter of Africans who moved to Europe were single at the time of their migration.
4. The likelihood of return migration among African migrants in Europe is often linked to reasons for initial migration. For example, migrants from Ghana and DR Congo who went to Europe to study were over five times more likely to return than migrants who left for other reasons, whereas those who left DR Congo for political reasons were extremely unlikely to return.
5. In instances where migrants retain strong links with their country of origin, this is not a guarantee of imminent return migration. Indeed, Congolese and Senegalese migrants who had sent remittances to or visited their country of origin were found, paradoxically, to be more likely to delay return.

WP7 : Labour outcomes for migrants and returnees

WP leader: Eleonora Castagnone, FIERI

1. The employment profiles of migrants upon arrival in Europe vary considerably between flows from different origin countries, and between 'traditional' and 'new' destination countries, with the latter generally attracting more low-skilled migrants.
2. The employment trajectories of African migrants in Europe are also diverse, reflecting different educational profiles, ease of integration, and access to jobs that match migrants' skills.
3. There is a significant link between studying for a higher degree in Europe and access to highly skilled positions thereafter. Migrants were much more likely to enter skilled work if they studied after arrival in Europe than if they arrived from a skilled occupation in Africa.
4. Migrant women are less likely to be economically active than men, with the largest gap existing between male and female Senegalese migrants and the smallest between Ghanaian male and female migrants.
5. Economic contributions of migrants to their country of origin through remittances, contributions to hometown associations, and the purchase of property increase according to time spent in Europe.
6. The economic re-integration of migrants from all three countries is broadly positive after return. However, origin countries tend to experience a 'brain re-gain' rather than a 'brain gain', as the occupational level of returnees mainly reflects their initial position before migrating.

WP8: Family life between Africa and Europe

WP leader: Valentina Mazzucato, UM

1. Three-quarters of the migrants interviewed in the MAFE Project were part of a nuclear family, meaning that they had a spouse and/or children. For two out of five of these migrants, migration led to the creation of a transnational family structure, where at least one member of the nuclear family was living in another country. The proportion of transnational families was especially high for Senegalese migrants, but less so for Congolese and Ghanaian migrants.
2. Across all destinations, migrants with transnational family arrangements had typically been in Europe for less time than those in unified or re-unified families. Amongst Senegalese migrants in Europe, men were more likely to have transnational families than women, and for all three African groups undocumented migrants were more likely to have them than documented migrants.
3. Of those migrants with families, reunification in Europe was far from universal, with one quarter of Ghanaian and one third of Congolese migrants being reunified at the time of the survey. Reunification was even less prevalent amongst Senegalese families.
4. Contrary to policy assumptions, African migrant families did not always reunify in Europe; in fact, a significant number of all family reunification took place in the country of origin.
5. Households in Africa typically did not contribute to the financial costs of migration for their members with only one fifth of Ghanaian households, and one quarter of Congolese and Senegalese households making a contribution. In all cases, children of household heads were the most frequently supported.
6. A high proportion of households in Africa had access to international social networks and received remittances. Spouses, children and siblings of the household head were the most frequent remitters, but in some cases more extended kin also sent remittances.

Potential Impact:

MAFE is likely to have an important impact in the field of quantitative analysis of international migration. By collecting and preparing new quantitative data on Afro-European migration, the MAFE project achieved a strong impact on the scientific community. The data will be released in January 2014 and is already very much expected.

Besides, the analysis on the new data-set confirms, with quantitative evidence, several trends and behaviours that, so far, had only been describe by qualitative work. They also highlight interesting new results that have been communicating to end-users of the MAFE data. We believe African migration remains, however, an understudied field of research (the bulk of the academic literature is built upon Latin American migration, and is especially about the Mexican case) and policy making is handicapped by a lack of objective evidence. Thus, MAFE provided empirical results on African migration to nuance the distorted but prevalent images provided by media and some policy circles.

Scientific impact

Impact on training a new generation of young scientists

The MAFE partners bring together specialists of quantitative data analysis and partners that have developed more qualitative approaches to the study of migration. In addition, the specific nature of the data collected made it important to organise statistical analysis trainings to the team of young researchers involved in the data analysis throughout the project. With the support of Ined statistical methods department, a series of statistical analysis trainings were organised throughout the project.

Several research stays also took place at Ined in 2011 and 2012, in order for MAFE team members to meet and work with Ined statistical methods department and the MAFE data manager in particular.

The network of researchers created by the MAFE team members and the new MAFE data provided an excellent background for the supervision of students and the development of PhD projects. With additional resources levered to that end, the MAFE group welcomed 12 Phd students. From October 2008 to December 2012, five theses were defended with high success. All these new Doctors now hold post-doc or researcher positions in public or private research organisations.

The data collection necessitated a large pull of qualified interviewers which were mainly recruited among the students of the MAFE partners, at different level. In addition to the coordinators, four categories of staff have participated in the MAFE surveys, namely interviewers, fieldwork supervisors, editors and coders, and data entry agents.

The project also relied on the strong implications of experienced researchers in the context of post-doctoral positions. The MAFE project seems to have been a good reference for their career development.

The MAFE data will be made available through Ined website, according to regular data access procedures. Ined's data access page provides access via a Nesstar server to the complete list of surveys undertaken by Ined. As well as providing online access to the documentation linked to each survey (presentations, field of study, methodology, questionnaire, etc.), for certain surveys Ined's Nesstar server offers detailed documentation of data to the level of the variables used (questions, instructions to interviewers, filters, response categories and frequency distributions). A multi-criteria research tool is also available. As for all Ined data, access will be granted via the website of the Réseau Quêtelet, the French data archive in SHS. A reflexion is however going on to find the right dissemination channel to this international data-sets, as the French speaking interface of both Ined and Réseau Quêtelet are not totally suited to the MAFE data-set.

The dissemination strategy of the data also plans to develop a simplified data-set that would be made available to masters' degree in social sciences and population study. Inclusion into the curriculum of the European Doctoral School of Demography is also under preparation.

New collaborations

The MAFE data-set is highly expected in the research community. Already, several projects are underway to valorise the data taking into account both the content and methodological aspects of the data. A few examples are given below.

1. Spanish developments of MAFE data-sets: the MESE project (Migraciones entre España y Senegal' -CSO2009-12816). Principal investigator: Amparo Gonzalez-Ferrer (CSIC) and Pau Baizán (UPF) funded by the Spanish Ministry of Science and Technology. The project Migrations Between Senegal and Spain (MESE) aims to contribute to overcome the existing lack of understanding of the patterns and determinants of African migration to Europe by providing statistically representative data for the Senegalese population residing in Spain (both documented and undocumented). They will be analysed within the wider context of the data collected in the framework of the Seventh Framework Programme (FP7) Project Migration Between Africa and Europe (MAFE) Project.
2. MAFE was also a key data sources for two project proposals submitted under the latest FP7 call on temporary migration (Call SSH.2013.3.1-1 Addressing European governance of temporary migration and mobility to Europe). The TEMPER proposal consisted in providing a comprehensive assessment of the pros and cons of recent initiatives to promote circular migration as an alternative to both more traditional forms of temporary and permanent migration, not only among prospective migrants but also among current residents in the EU. A competitive project was proposed by IMI (Oxford University, UK). The MAFE data is mobilised to carry out original analysis on migrant circulation.
3. Proposals for a trilateral collaboration with the Max Planck Institute for Demographic Research (Germany) have emerged. The objective of this emerging project is to develop an agent-based model of international migration, using the MAFE data to inform the different stages of the emigration decision, but also including the contextual data into the model. The projection should give plausible migration numbers, plausible age profiles and a plausible composition of migrants (by sex, skill level, etc.).

Final project conference

The final MAFE conference, 'Comparative and multi-sited approaches to international migration' took place in Paris from 12 to 14 December 2012. The objective of this conference was to promote a multi-sited and comparative approach to international migration, explicitly bringing together researchers and research evidence from different parts of the world. The conference focused on quantitative approaches to international migration that deal simultaneously with processes in places of origin and destination. Papers tackled a number of areas and regions, and addressed significant policy concerns.

Following a call for paper launched in March 2012, the scientific committee of the final MAFE conference received more than 62 paper proposals. Ninety-two persons attended the conference, which gathered some high level scholars and data users from the USA and Europe. It opened the way for new initiatives to compare international migration systems in Europe and the USA.

Scientific Committee: Pau Baizan (Universitat Pompeu Fabra, Spain), Cris Beauchemin (Ined, France), Jørgen Carling (Peace Research Institue, Oslo), Eleonora Castagnone (Forum Internazionale ed Europeo di Ricerche sull'Immigrazione, Italy), Katharine Donato (Vanderbilt University, USA), Jean-Christophe Dumont (OECD), Hein de Haas (International Migration Institute, Oxford), Amparo Gonzalez (Consejo Superior de Investigaciones Científicas, Spain), Flore Gubert (DIAL, France), Lama Kabbanji (CEPED, France), José Mangalu (Université de Kinshasa, DR Congo), Douglas Massey (Princeton University, USA), Valentina Mazzucato (Maastricht University, Netherlands), Bruno Schoumaker (Université catholique de Louvain, Belgium).

Following the conference, a special issue of the International Migration Review, a leading journal in the field, is being prepared.

Publication of two MAFE books

Two books were planned to be published in a relatively short time after the end of the project: both were expected in 2013.

1. 'Migrations africaines: le co-développement en questions. Essai de démographie politique.' Ed. Cris Beauchemin (INED), Lama Kabbanji (IRD), Papa Sakho (IPDSR-UCAD), and Bruno Schoumaker (Université catholique de Louvain). This first book is published under the research collection of Armand Colin in French. One of the innovations of the work is to propose an analysis of 'political demography.' This does not constitute a sub-discipline as such, but rather a scientific bias of defining research questions related to political issues. The first step is to analyze the discourses and political systems in order to identify the underlying assumptions (or ideas) under the decisions. Second, demographic and statistical analysis are mobilised to test how these preconceptions, which justify the policies implemented, are checked when studying the behavior of people, especially those migrants. Finally, third, scientific results are put into perspective by 'practitioners', such as migrant associations, non-governmental organisations (NGOs), development assistance and support to migrants, technocrats). The conclusion offers a summary record of the debates raised by these results during a roundtable of the project bringing together researchers and African and European actors of international migration.
2. 'Migration between Africa and Europe'. Ed. Cris Beauchemin. The book is using the MAFE working papers that are submitted to a peer review process. It should be published by end 2013 at Springer'. It is the first book to assemble systematically quantitative evidence on Sub-Saharan migration.

Wider societal impact

The WP 9 of the project dealt with the policy dialogue activities of the project, with the intention to both:

1. consult with the policy-making communities to elaborate on the MAFE analysis and identify key questions that should be tackled by the research teams;
2. account of project results and advancements to a non scientific audience.

The project relied on previous linked with a Senegalese based NGO (Enda-Europe) and developed ties with several stakeholders gathered in an advisory board. Throughout the project duration, the project partners updated a 'boundary partners' list that was used to establish the project mailing lists and contact participants to policy dialogue events.

Non-academic activities

The MAFE project developed four main tools for non-academic dissemination:

1. From June 2009, the MAFE website was made available in two languages as much as possible.
2. A registration tool allowed the team to collect mailing contacts of interested stakeholders in the project.
3. Five newsletters were distributed to a mailing list of more than 500 contacts from non-academic arenas, such as journalists, members of the European Parliament (MEPs), administration and ministry officers, NGO representatives etc. in the nine MAFE countries.
4. Five policy briefs were distributed to the mailing list and through the partners communication offices.

List of Websites:

Until 31 December 2013, the project website is accessible at the following address: http://www.mafeproject.eu. From 1 January 2014, the project website will be accessible at http://www.mafeproject.ined.fr.

For more information on the MAFE project contact the coordinator of the project Cris Beauchemin.

Cris Beauchemin

Institut National d'Etudes Démographiques (INED)

133 boulevard Davout

75980 Paris cedex 20

France

Tel : +33-156-062128

cris.beauchemin@ined.fr