Return Migration as Failure or Success?: The Determinants of Return Migration Intentions Among Moroccan Migrants in Europe.

Different migration theories generate competing hypotheses with regard to determinants of return migration. While neoclassical migration theory associates migration to the failure to integrate at the destination, the new economics of labour migration sees return migration as the logical stage after migrants have earned sufficient assets and knowledge and to invest in their origin countries. The projected return is then likely to be postponed for sustained or indefinite periods if integration is unsuccessful. So, from an indication or result of integration failure return is rather seen as a measure of success. Drawing on recent survey data (N = 2,832), this article tests these hypotheses by examining the main determinants of return intention among Moroccan migrants across Europe. The results indicate that structural integration through labour market participation, education and the maintenance of economic and social ties with receiving countries do not significantly affect return intentions. At the same time, investments and social ties to Morocco are positively related, and socio-cultural integration in receiving countries is negatively related to return migration intentions. The mixed results corroborate the idea that there is no uniform process of (return) migration and that competing theories might therefore be partly complementary.

Introduction

Although the focus of migration studies has traditionally been on immigration towards Western countries, the issue of return migration has recently been receiving increasing attention in the migration literature (cf. Asiedu 2005; Cassarino 2004; Rodriguez and Egea 2006). Much academic writing on return migration, however, has focused on the contribution which return migrants can make to economic development in countries of origin (McCormick and Wahba 2001; Thomas-Hope 1999; cf. Diatta and Mbow 1999). There has been comparatively little research on the individual and contextual factors which determine return migration. It is therefore not very surprising that most policy measures taken to encourage return migration have failed (Dustmann et al. 1996).

The underlying behavioural mechanisms of return migration is an important topic to explore further empirically, since different migration theories offer radically opposed interpretations of return migration (Constant and Massey 2002; Fokkema and de Haas 2011). Neoclassical migration theory (NE) models migration as individuals’ behaviour to maximize their utility by moving to places where they can be more productive (Harris and Todaro 1970; Massey et al. 1998; Todaro and Maruszko 1987) or where they can expect the highest returns on their human capital investments (Bauer and Zimmermann 1998; Sjaastad 1962). Because of the implicit assumption that migrants have access to full information on opportunities abroad prior to migration, NE and, in general, conventional push-pull models have difficulty in explaining return migration: after all, migrants whose expectations are met are expected to integrate successfully and be more productive than in their origin countries, so that there will be no rationale for returning. However, if we reject the latter assumption, it becomes possible to imagine situations in which expectations do not come out. For instance, migrants might fail to find a job and to improve their lives through migrating, in which case they are more likely to return. So, within a NE perspective, return migration is mainly interpreted as a result of failure to find a place in receiving societies. Put differently, while “winners” settle, “losers” return.

If migration is perceived as an individual cost-benefit analysis as in NE, it makes little sense for successful migrants to maintain economic and social ties with people living in origin societies, because these ties would only raise the financial (and psychological) costs of staying abroad and lower the costs of return migration. The other way around, economic and social ties at the destination will decrease the costs of staying and increase the costs of returning. This interpretation also fits within a classical immigrant assimilation theory (Castles and Miller 2003; Portes et al. 1980) which predicts that immigrants will gradually assimilate into receiving societies, going along with a concomitant decline of transnational ties and a declining inclination to return.

However, we can cast doubt on these interpretations for two fundamental reasons. First, there is reason to question the assumption that orientations on origin and destination societies are necessarily substitutes. This has been challenged by recent studies on migrant networks and transnationalism (Glick Schiller et al. 1992; Portes 2003), which have questioned the idea that the maintenance of economic and social ties with origin societies and return migration is necessarily a manifestation of their inability to integrate. Some empirical evidence supports the idea that integration and transnational ties are not necessarily substitutes, but can be complements (Portes et al. 1999; Snel et al. 2006). Second, the new economics of labour migration as pioneered by Oded Stark assumes a radically different rationale behind migration and, hence, an opposing effect of immigrant integration on return migration.

The new economics of labour migration (NELM) interprets migration as a livelihood strategy employed by households and families instead of individuals to spread income risks and to overcome sending country (credit, insurance) market constraints. This co-insurance model is a radical departure from neoclassical models which conceptualize migration as income or utility-maximizing behaviour by individuals. The central NELM idea is that households send out the best-suited individuals to gain an income elsewhere. The money they remit serves to spread income risks, to increase income, to improve living conditions and to enable them to invest (see several key papers compiled in Stark 1991). If the prime motive for migrating is to improve the situation at the origin, migrants will only return once they have succeeded to amass, save and remit enough financial and human capital in order to realize their investment plans. Importantly, this turns NE interpretations of return migration upside down, that is, from an indication or result of integration failure (according to NE) to a measure of success (according to NELM). The projected return is likely to be postponed for sustained or indefinite periods if integration is unsuccessful. Permanent settlement then becomes the end result of repeated postponement of return because of integration failure. Such prolonged stays can go along with the maintenance of strong transnational ties with origin societies and can become inter-generational, which seems to contradict conventional assimilation theory.

There is still a lack of pertinent studies which simultaneously test these conflicting hypotheses on return migration. There is some empirical evidence suggesting that there is a positive correlation between integration and return migration intentions (Waldorf 1995). However, analyzing survey data from Denmark, Jensen and Pedersen (2007) observed that various variables measuring labour market involvement had a negative effect on return migration. In addition, Dustmann (2008) found evidence that educational investments in children as well as their permanent wages are associated with a higher probability of permanent migration of the father. A study of European experiences with return migration by Dustmann et al. (1996) showed that, in general, return propensities increase with the age at entry, but decrease with the number of years of residence. They also found that, among those migrants who decided to return, the remaining years in the destination country decrease with years of residence. However, such results do not provide a direct test for NE- and NELM-derived hypotheses on the effect of integration on actual or intentional return migration.

Two of the rare simultaneous tests of NE and NELM theories are the studies by Constant and Massey (2002) and Fokkema and de Haas (2011) on the probability of return among migrants in Germany and return migration intentions among African migrants in Spain and Italy, respectively. Their studies found mixed support for both hypotheses, suggesting that there is no unitary process of return migration because of heterogeneity in the background and motivations of migrants. The latter interpretation does seem to make a lot of sense: the relation between integration processes and return migration is likely to depend on initial motivations to migrate, livelihood opportunities in origin and destination societies as well as educational, cultural and other specific features of immigrant groups. So, it is likely that there is no one-size-fits-all theory, and depending on the specific context, both theories might provide powerful explanations of observed patterns.

However, more analyses of micro-level data are desirable to test the further applicability of these apparently competing theories for different migration contexts. This study aims at contributing to filling this empirical gap by analyzing return intentions among Moroccan migrants living in Europe. As they form the second largest non-EU immigrant group living in the Europe, the case of Moroccan migration is particularly relevant to test competing theories on return migration in a European migration context. This study draws on data collected by the 2005 survey “Les Marocains Résidant à l’Étranger” (“Moroccan residents abroad”). More than 2,800 Moroccans living in a range of different European countries were interviewed about their current living conditions, their link to their origin country, and their intention to return. Hence, this survey provides a unique opportunity to study a variety of factors in sending and receiving countries impacting on return migration intentions among Moroccan migrants across Europe.

This study focuses on return migration intentions. Although the study of actual return migration behaviour has obvious advantages, there is an added value to studying return migration intentions. As Waldorf (1995) argued in her study on return migration, the conventional focus on actual return is based on the implicit assumption that observed behaviour is preceded by a desire to migrate and that the factors influencing actual behaviour affect migration intentions in a similar fashion. We know that this is not the case and that also among Moroccan migrants, the return is often not driven by an actual desire to return (de Bree et al. 2010). So, there might be discrepancies between intentions and actual migration behaviour, which means that it is also relevant to study migration motivations that may or may not precede actual migration behaviour.

We will first give background information on the migration history of Moroccans to the EU and the role of return migration. The next section will address the data and methods of this empirical study. The fourth section will present the results of the analysis, and we will conclude by summarizing the theoretical inferences of this study and by suggesting future avenues for research on this topic.

Moroccan Migration to and from Europe

Since the 1960s, Morocco has evolved into one of prime source countries of migrants to Europe. In the late 1960s and early 1970s, the combination of high economic growth in West Europe, the lack of immigration restrictions and active recruitment led to a boom of labour migration to France, and, to a lesser extent, the Netherlands, Belgium and Germany. The economic recession following the 1973 oil crisis and rapidly increasing unemployment led to a recruitment freeze.

Initially, it was widely expected by sending and receiving countries as well as the “guestworkers” themselves that their migration was temporary. However, the economic recession following the 1973 oil crisis and rapidly increasing unemployment did not lead to the widely anticipated return of Moroccan migrants. Contrary to expectations, and despite the recruitment stop and the gradual tightening of immigration policies after 1973, relatively few Moroccan migrants returned, and many ended up staying permanently, a process which was accompanied by large-scale family reunification and family formation through new transnational marriages between migrants’ children and Moroccan residents (Hooghiemstra 2001; Lievens 1999; Reniers 2001).

This family migration explains much of the continuation of Moroccan emigration over the late 1970s and 1980s. Increasing immigration restrictions did little to stop migration, but rather reinforced the reliance on family migration through networks. It also led to an increasingly irregular character of migration and the exploration of new migration destinations in the booming economies of southern Europe and, to a lesser extent, North America. While since 1990, Moroccan low-skilled emigration has increasingly focused on Italy and Spain, the higher skilled have increasingly migrated to the USA and Canada (de Haas 2007).

Moroccans now form not only one of largest but also one of the most dispersed migrant communities in Western Europe. Over 3 million people of Moroccan descent (out of a total population of 30 million) are currently believed to live abroad. According to figures from Moroccan consulates, in 2004, France was home to the largest legally residing population of Moroccan descent (more than 1,100,000), followed by Spain (424,000), the Netherlands (300,000), Italy (299,000), Belgium (293,000) and Germany (102,000). Smaller but rapidly growing communities of high-skilled migrants live in the USA (100,000) and Canada (78,000). Actual numbers may be rather higher, due to substantial undocumented migration (de Haas 2007).

While most labour migrants who reunified their families ended up staying permanently, in the late 1980s and early 1990s also a return movement occurred mainly consisting of relatively elderly, retired or jobless Moroccans migrants (de Haas and Fokkema 2010). Between 1985 and 1995, some 314,000 migrants returned to Morocco from France, the Netherlands, Belgium, Germany, the UK and Denmark. Since 1994, return migration has fallen to less than 20,000 per year. Return migration among first-generation Moroccans has been low compared to other immigrant groups in Europe (Fadloullah et al. 2000).

The predominantly permanent character of Moroccan migration is also testified by a high tendency towards naturalization (Berrada 1990; Fadloullah et al. 2000). However, this trend towards dual citizenship does not mean that Moroccan immigrants and their descendants have entire blended into “mainstream” culture. On the contrary, the Moroccan heritage population in EU countries seems to increasingly develop a new, distinct diasporic (cf. Cohen 2008) group identity in which identification with origin and settlement societies coexist and transnational ties with family in Morocco are maintained. At the same time, the political climate towards Moroccan (and other predominantly low-skilled immigrants) has become harsher in several European countries, particularly in the Netherlands. This has been part of a general backlash against multiculturalism and a political shift towards assimilationism (Vasta 2007).

Within this climate, politicians and the media have increasingly portrayed the maintenance of socioeconomic ties and identification with origin societies as a manifestation of the inability or even unwillingness to “integrate”. In the same vein, dual citizenship, which used to be seen as facilitating integration, is increasingly negatively interpreted as “double loyalty” which would actually block integration. However, prior empirical evidence has cast serious doubt on the assumption that “integration” and transnational ties are necessarily substitutes and has shown that they can be complementary (see above). Moreover, it is still unclear whether and to what extent the decreasing tolerance towards diversity and transnationalism has affected return migration intentions. Although there is a lack of pertinent data to assess this, macro-trends do not suggest a significant increase of return migration to Morocco. There has apparently been an increase in return migration to other countries such as Turkey, a trend which seems primarily related to the radically improved economic circumstances in Turkey (de Haas 2009).

While Moroccan return migration is comparatively low, the Moroccan heritage population in the EU faces relatively high rates of socioeconomic marginalization compared to several other immigrant and minority groups, as is testified by high unemployment, relatively low incomes, high school drop-out rates and residential segregation (Aparicio 2007; Ireland 2008; Snel et al. 2006). At face value, these macro-level observations seem to confirm the idea that the lack of return migration is a corollary of integration problems and, hence, apparently confirming the NELM hypothesis on return migration (see above). However, we would commit a classical ecological inference fallacy to derive inferences on individual migrants’ behaviour from aggregate, macro-level statistics. Such reasoning only holds if immigrant groups are homogeneous, which seems an unrealistic assumption.

After all, return motives can be highly diverse, ranging from socioeconomic exclusion in the country of destination to successful business investments in Morocco. Hence, appropriate micro-data is needed in order to adequately assess the determinants of return migration, thereby providing a test for the competing theories and hypotheses outlined above. Particular attention will be paid to the effects of integration, transnational ties, experiences and perceptions of discrimination and overall climate vis-à-vis immigrants and differences between destination country contexts.

Methods

Data

Data stem from the survey “Les Marocains Résidant à l’Étranger” (“Moroccan residents abroad”), conducted by the Haut-Commissariat au Plan–Centre d’Études et de Recherches Démographiques (HCP 2007). In August-September 2005, face-to-face interviews were carried out among 2,832 Moroccans, who were head of the household, aged 15 and over and living in Europe. Due to financial and logistic constraints—lack of money to interview Moroccans in the destination country and high mobility of Moroccans when they visit their origin country—the interviews were held in the four Moroccan port cities of Tanger, Sebta, Nador and Al Hoceima. In the absence of a reliable sampling framework of Moroccans living in Europe, quota was applied according to migrants’ destination country and region of origin.1 Of those approached, 94.4 % successfully completed the interview. We excluded 120 persons who were born outside Morocco. The sample is further reduced to 2,633 cases due to missing values on one or more of the relevant variables.

The survey collected extensive information on respondents’ past and current socio-demographic and socioeconomic situation in Morocco and the country of destination. This provides the unique opportunity to study a variety of origin and destination country factors that potentially affect return migration intentions. The survey also covers Moroccans living in various European countries, including those countries with a long-standing “guestworker”-migration history (e.g. France and the Netherlands) and more recent destination countries in southern Europe (mainly Italy and Spain), allowing us to examine between-country differences and similarities. The main drawback of the survey is its inherent male bias, because only the (predominantly male) household heads were interviewed. Besides this gender bias, as the interviews were conducted among migrants visiting Morocco, this study is inevitably biased towards migrants with a higher than average attachment to Morocco.

Measures

Apart from return migration intention, integration, ties with receiving and destination country and residential quality of life (i.e. experiences and perceptions of discrimination and overall climate vis-à-vis immigrants), several background characteristics were included in the analyses. These latter were gender (0 = male, 1 = female), age (in years), length of stay (in years), prior migration experience (0 = no, 1 = yes), educational level (from no or incomplete education (reference group) to above secondary) and religiosity (0 = not (using facilities for) worshipping, 1 = using facilities to worship). Furthermore, dummy variables for the main countries of destination were constructed, with France as reference group. For descriptive information on the variables, see Table 1.

Table 1

Descriptive statistics of the variables (N = 2,633)

	Percent	M
Return migration intention	57.9
Background characteristics
Female	5.4
Age (19–84)		42.8
Length of stay (in years 0–56)		18.8
Prior migration experience	11.0
Level of education:
No education	18.5
Preschool/Primary	25.2
Secondary	38.6
Above secondary	17.6
Religiosity (% using facilities for worshipping)	75.2
Socio-cultural integration
Objective (0–7)		2.3
Subjective:
Integrated	62.0
In-between	29.6
Excluded	8.4
Structural integration
Paid job	84.0
Occupational status current job:
Without qualification	30.8
Low-skilled worker	32.6
High-skilled worker	24.7
Technician	5.2
Management	6.7
Occupational status current job (0–4)		1.2
Economic ties receiving country
Owner of a house	30.2
Investment in receiving country	27.2
Project in receiving country	21.0
Economic ties Morocco
Investment in Morocco	45.5
Project in Morocco	60.7
Social ties receiving country
Partner living in receiving country	77.9
Children in same household	73.2
Children in receiving country	14.0
Social ties Morocco
Partner living in Morocco	9.8
Children in Morocco	6.5
Frequency of visiting Morocco during the past 3 years:
0–1	10.4
2	17.5
3	59.9
4+	12.2
Residential quality of life
Discrimination of Moroccans in public services	21.3
Discrimination of Moroccans: number of public services (0–9)		0.7
Degree of satisfaction with facilities to worship:
Satisfied	80.7
Neither satisfied nor dissatisfied	17.4
Dissatisfied	1.9
Feelings of racism	30.2
Feelings of racism: number of settings (0–4)		0.6
Host country
France	42.3
Belgium	7.4
the Netherlands	7.4
Italy	15.5
Spain	23.9
Other	3.6

Return Migration Intention

To assess the intention of returning home, respondents were asked “Do you consider to return to Morocco?” The answer categories were 0 “no” and 1 “yes”. 2

Integration

Both structural and socio-cultural integrations were examined. Structural integration refers to the acquisition of rights and status within the core institutions of the receiving society (Heckmann 2005). In this study, it was measured by labour force participation (0 = no, 1 = yes) and occupational status of the current job (range 0 “without qualification” to 4 “professional/management” with the mean score to those unemployed). Socio-cultural integration pertains to the cognitive, behavioural and attitudinal changes in conformity to the dominant norms of receiving societies (cultural integration or acculturation); social intercourse, friendship, marriage and membership of various organizations (interactive integration); and feelings of belonging, expressed in terms of allegiance to ethnic, regional, local and national identity (identificational integration) (King and Christou 2007; King and Skeldon 2010).3

In the questionnaire, these different dimensions of socio-cultural integration were represented by the following variables: watching Moroccan television channels (0 = frequently; 3 = never), attitude towards mixed marriage (0 = disagree, 1 = agree), membership of organizations (0 = no; 5 = membership of five different organizations), participating in past election (0 = no, 1 = yes), frequency of contact with non-migrants (0 = never; 3 = often), having a non-migrant as best friend (0 = no, 1 = yes) and feelings of belonging (0 = Morocco, 1 = Morocco and receiving country, 2 = receiving country). Instead of studying these indicators separately, we constructed one index of socio-cultural integration (α = 0.63): the seven indicators were summed after the range of each indicator was revalued with a minimum of 0 and a maximum of 1. As a result, the score on the index variable is at least 0 and at most 7: the higher the score, the more the respondent is integrated socio-culturally. In addition to this index, respondents were asked straightforward about their degree of integration in receiving country. The answer categories were “integrated”, “in-between” and “excluded” (reference category).

Ties with Receiving and Origin Countries

A distinction is made between economic and social ties. With regard to economic ties, respondents were asked whether or not they have either an “investment” or a “project” in the country of destination and Morocco, respectively, and being “owner of a house” in the country of destination. Social ties with receiving country is represented by the dummy variables (0 = no, 1 = yes) having “a partner”, “children in the same household” and “children outside the household” living in the country of destination. For social ties with Morocco, we included the dummy variables (0 = no, 1 = yes) having “a partner” and “children” living in Morocco and the frequency of visiting Morocco during the past 3 years (range 0–4+).

Residential Quality of Life

The survey data comprises three different indicators of respondents’ residential quality of life. First, showing a list of nine public services respondents were asked how Moroccans are treated compared to non-migrants. Response options ranged from “better” to “worse”. Based on this information, the dummy variable (0 = no, 1 = yes) “Discrimination of Moroccans in public services” was created. The second indicator concerns the degree of dissatisfaction with facilities to worship, running from 0 “yes” (including those not (using facilities for) worshipping) to 2 “not at all”. The third indicator, feelings of racism, is a dummy variable (0 = no, 1 = yes), indicating whether or not the respondent ever experienced racism in (1) working place, (2) residential neighbourhood, (3) administration and/or (4) other places.

Results

Table 2 displays the results of the logistic regression analysis in the form of odds ratios. Return intentions do not seem to be significantly affected by respondents’ age and gender. Contrary to expectations, length of stay has a positive impact on the likelihood of intending to return. This might indicate that many Moroccan migrants intend to return after retirement.

Table 2

Logistic regression of return migration intention (N = 2,633)

Background characteristics
Female	0.82
Age	1.01
Length of stay	1.01*
Prior migration experience	1.42*
Level of education:
Preschool/Primary	1.45**
Secondary	1.26∼
Above secondary	1.41∼
(ref. no education)
Religiosity	1.56***
Socio-cultural integration
Objective	0.84***
Subjective:
Integrated	0.65*
In-between	0.86
(ref. excluded)
Structural integration
Paid job	0.94
Occupational status	0.98
Economic ties receiving country
Owner of a house	1.10
Investment in receiving country	1.00
Project in receiving country	0.93
Economic ties Morocco
Investment in Morocco	1.32**
Project in Morocco	3.58***
Social ties receiving country
Partner living in receiving country	1.31
Children in same household	0.88
Children in receiving country	1.03
Social ties Morocco
Partner living in Morocco	1.79**
Children in Morocco	1.40
Frequency of visiting Morocco during the past 3 years	1.10∼
Residential quality of life
Discrimination of Moroccans in public services	1.16
Degree of dissatisfaction with facilities to worship	1.23∼
Feelings of racism	1.37**
Host country
Belgium	1.06
the Netherlands	1.47*
Italy	1.35*
Spain	1.31*
Other	1.49∼
(ref. France)

***p < .001; **p < .01; *p < .05; ∼p < .10

Experienced migrants, who have moved several times in their life, are more likely to intend to return, possibly reflecting the less settled nature of their life histories. The effect of educational attainment on return intentions is positive, albeit non-linear. The highest likelihood to intend to return can be found among Moroccan migrants who have completed preschool or primary education, followed by the highest educated migrants, while unqualified migrants are the least likely to express a return intention. Religiosity significantly influences intentions to return. Moroccan migrants who use facilities to worship are more likely to intend to return in the future.

In line with neoclassical migration theory, there is a clear negative association between socio-cultural integration and return intentions. This holds for both the subjective and more objective measure of socio-cultural integration. With regard to the objective measure, further analyses showed that having a non-migrant as best friend and feelings of belonging to host country have a negative effect on return intentions.

Structural integration, however, does not seem to have a significant effect on return intentions. Neither having a paid job nor occupational status has a significant effect. Economic ties to the destination country (owning a house, having an investment or a project) do not decrease return intentions. The maintenance of economic links with Morocco, on the other hand, shows an unequivocally positive correlation with return intentions. Those who invest in Morocco are significantly more inclined to return. It goes without saying that these investment variables are likely to be partly endogenous. For instance, investments are likely to mirror return intentions to a certain extent.

Attachments to people in the destination country (presence of partner, children in and outside household) do not seem to play an important role in the likelihood of intending to return either. Social ties to Morocco do have the expected positive effects on return intentions—an increased likelihood to intend to return when having a partner or children in Morocco and visiting country of birth frequently—but only the partner-effect attains statistical significance at the 0.05 level.

Conditions that negatively influence “residential quality of life” pull Moroccan migrants back to their country of birth: feelings that Moroccans are less well-treated than non-migrants in public services, being dissatisfied with facilities to worship, and experiences of racism have a positive effect on return intentions, although only the latter attains statistical significance at the 0.05 level.

Finally, the country dummies indicate that Moroccans living in the Netherlands, Italy and Spain are significantly more likely to intend to return than their peers living in France. As the analysis controls for duration of residence, the higher likelihood to consider returning among Moroccans in the Netherlands, Italy and Spain cannot be attributed to their possible shorter stay in destination countries. Neither country differences in migrants’ composition are likely to be the explanation because of the inclusion of a large number of socio-demographic characteristics of migrants in the analysis. The relatively high inclination of Moroccan citizens in Italy and Spain to return possibly reflects the more recent, less established migration history of these two countries. The increasingly negative social and political climate towards migrants in the Netherlands might have contributed to the relatively high likelihood to intend to return among Moroccans in the Netherlands. The fact that France and Belgium (Wallonia) share linguistic links with Morocco might also explain the relatively low return intentions among Moroccan migrants in those countries.

As noted above, especially the relationship between the two investment variables and the intention to return is subject to reverse causality as those with a return intention are more likely to invest in their home country or they only intend to return for investment purposes. To examine whether inclusion of the two investment variables has led to serious biases in the odds ratios of the other explanatory variables in our model, we re-ran the analyses with exclusion of these investment variables. The results clearly show hardly any changes in the other variables’ effects (results available upon request). Two issues are worthy to note: in the model without the investment variables, Moroccan migrants are significantly (at the .05 level) more likely to intend to return when having children in Morocco and the higher likelihood to consider returning among Moroccans in the Netherlands, compared to their peers in France, is not significant anymore.

Conclusion

This study examined the determinants of return intention among Moroccan migrants across Europe. Special focus was on the effect of migrant’s integration in the country of destination as different motivations for international migration and hence, competing integration effects on return migration are claimed in the migration literature. Within the neoclassical migration theory (NE), according to which migration is an income or utility-maximizing behaviour by individuals, a negative effect of integration can be expected. According to the new economics of labour migration (NELM), migration from developing countries should be understood as a household livelihood strategy and migrants will return once they have been successful in earning sufficient income to accumulate assets and to invest in the origin country.

Our analyses were based on data from the survey “Les Marocains Résidant à l’Étranger” (“The Moroccan residents abroad”), including more than 2,800 Moroccans living in a wide range of European countries. Although to our knowledge this sample is the largest one of its kind to date, the sample is likely to be biased towards men as well as to those migrants who have a relatively high attachment to Morocco. Hence, the respondents may not be representative of Moroccan migrants living in Europe.

The findings support neither NE nor NELM completely. In line with NE, socio-cultural integration in destination countries has a negative effect on return migration intentions. However, structural integration through labour market participation and the fostering of economic and social ties in destination societies do not significantly affect return intentions. Investments and, to a lesser extent, attachments to people in Morocco, on the other hand, are positively related to return migration intention. This suggests that structural integration in receiving countries on the one hand and the maintenance of ties with origin countries on the other hand are not necessarily substitutes. This might also explain the absence of a relation between age and the positive relation between length of stay and return migration intentions. This seems to provide support for the NELM hypothesis that return can be the result of success rather than failure to integrate economically and that such success might even be a condition for returning.

The mixed support for the NE and NELM-derived hypotheses suggests that there is no such thing as “the” (Moroccan) migrants and that initial motivations to migrate (and return) differ strongly among migrants and that they might change over time due to personal experiences and contextual factors such as discrimination, social exclusion and access to labour markets. This confirms, in line with the conclusions of Constant and Massey (2002) and Fokkema and de Haas (2011), that there is no uniform process of (return) migration and that the competing theories might be complementary in explaining return migration intentions and behaviours occurring between and within specific migrant groups and within specific origin and destination contexts. Hence, future research should pay more attention to the heterogeneity of migrants. In addition, more research would be warranted, preferably using fully representative data, to increase insights in the influence of contextual factors in order to explain the observed differences in return intentions within the several destination countries.