Remittances As Home Orientation Rooted in the Lifeworlds of Immigrants

Introduction

There are two dominant approaches to remittances in micro-level migration research. One considers remittances as an explanatory factor in the well-being of origin households (Özden and Schiff 2007: 4; Miller 2013; Ratha 2013); the other explains remittances in terms of altruism, pure self-interest, ‘tempered altruism’ (Stark and Lucas 1988) or other factors (Carling 2008a). Both of these approaches are relevant to the understanding of migration processes. However, focusing exclusively on these two approaches neglects the fact that remitting is part of interfamily processes and cannot simply be reduced to a ‘variable analysis’. Its full relevance for development cannot be derived from the pure ‘algebra of the variables’ (Blumer 1956). It has interpretation components that are strongly embedded in the ‘lifeworlds of migrants’ (Morawska 1984; Mau 2012) as given by their everyday lives and the way they perceive them. It is true that remittances are sometimes contextualised, but this is rarely the case, and is mainly limited to kinship relations (Vullnetari and King 2011). The research community, on the other hand, does feel a need to contextualise remittances (Page and Mercer 2012). The purpose of our study is to contribute to re-embedding remittances in the whole set of home-orientation behaviours of which they are a part.

Better contextualising of remittances is important for theoretical and policy reasons. Such a contextualisation is a key part of understanding: a) that for immigrants, remitting money is deeply embedded in their life strategies, and in their family and community life; and b) how the migration–development nexus is structured and, implicitly, how it could be influenced. Our study employs quantitative analysis including subjective variables, composite indices, typologies of transnational lifeworlds of immigrants, and multilevel models. Before describing the methodological details, we present the framework of the analysis, including first, a section on challenges in the analysis of remittances, and second, a section on the principles of re-embedding remittances in the lifeworlds of immigrants. The section on data and hypotheses will be followed by the results and conclusions.

The first contribution of the study is to propose and test both an index and a typology of the home orientation of immigrants, considering remittances in relation both to intentions to return home and to the intensity of communication with origin families. Second, the analysis tests a hypothesis on the role of different types of deprivation (at origin and destination, personal and collective) in shaping different types of home orientation. All these processes use comparative analysis of a large data set of immigrants in Spain (Reher and Requena 2009), with a small sample of Romanian immigrants in Madrid (Sandu 2009b) providing a more detailed set of measures for the variables of interest. Both data sets are unique in terms of their information content, allowing for the testing of the research hypotheses that are specific to this study. The data are appropriate for obtaining a better understanding of remitting behaviours of Eastern Europeans following the two most recent waves of enlargement of the European Union. Romania (with Spain as a preferred migration destination, after Italy) reached one of the highest rates of per-capita inflow of remittances in Eastern Europe in 2007; in the same year, Spain was sixth in the international ranking of remittance outflows (UNDP 2009).

In this study, monetary remittances are considered not in the standard opposition to social remittances (Levitt, Lamba-Nieves 2011), but as part of a home-orientation complex of economic, social and cultural components.

Current challenges in the analysis of remittances

The idea of understanding migration by contextualising is appropriate not only for its consequences (de Haas 2005), but also for explaining remitting behaviours (Page and Mercer 2012). The standard approach is to consider that unwritten contractual arrangements between migrants and their families involve intertemporal exchanges, of which remittances are a part, and reciprocal altruism creates an environment of low transaction costs as well as trust and loyalty (Stark and Lucas 1988). In this view, families invest in the education of children and, later, children as migrants send back money as compensation to their parents in the form of within-family social exchange (Gentry and Mittelstaedt 2009). The alternative to this solidarity with the left-behind family would be self-interest based on investment plans fostered by migrants remitting money to this family (Dustmann and Mestres 2010), waiting for property rights in the left-behind household, and consolidating their status and prestige in the home community (Stark and Lucas 1988: 470; Stark 1999). The planned behaviours of immigrants in relation to their home families and communities are part of this view, which regards remittances:

as a basket category that includes far more than just sending money from the place of domicile to a family. By remittances we also mean those contacts such as conversations on the phone that convey ideas, information and values, those journeys that move skills and knowledge around the world and the plethora of activities by which national and international connections are maintained (Page and Mercer 2012: 4).

A key challenge in the analysis of remittances, associated with the planned behaviours already mentioned, is how they relate to immigrants’ intentions to return home. The relationship is particularly difficult to analyse because remittances are measured as a period stock variable (how much money has been transferred home in a certain period of time) and return intentions are recorded as moment states (‘Do you intend to return to your home country?’). In most surveys that are not of the panel type, the researcher is put in the position of assessing the relationship between a past stock of remittances and a future-oriented behaviour as regards returning home. The usual hypothesis is that return intentions influence the probability and amount of remitting. Even if one distinguishes between different reasons for remitting (family support, saving for later, other reasons), the findings support the hypothesis. One of the methodological difficulties in testing the hypothesis is that the effect (stock of remittances) is measured for a time that occurs before recording the cause (intention to return). The challenge could be addressed in panel research by using lag correlations and imputing last-period remittances to return intentions at the beginning of the reference period. This is the procedure adopted for a large data set of immigrants living in Germany, using as dependent variables the probability of remitting and the amount of money sent home, and keeping under control relevant status predictors (Dustmann and Mestres 2010). This approach is not possible in non-panel surveys. In fact, several studies concur with the conclusion that ‘[w]hile the association between remittances and return or visits is clear, the causal mechanisms are complex’ (Carling 2008a: 590). Unfortunately, it is hard to find panel data that would enable the measurement of key variables for this study (remitting, deprivation at home and in the destination country, intentions to return, etc.) and allow for comparisons between Eastern European immigrants and those coming to Europe from other continents. Case studies and extended comparisons with cross-cultural data and multiple control variables could function as a substitute for panel data. This is why we worked with two complementary, cross-sectional data sets on immigration in Spain.

Even if the dominant determinant of remittances seems to be the intention to return, intention per se could be an effect of previous practices of sending money to those left behind, in a reverse causality pattern. What happens in real life is a continuous set of interacting processes of reciprocal adaptation of return intentions and remitting. Decisions in the interplay between remitting and return plans involve the continuous (re)interpretation by the migrant of the life space in terms of job, income, family, housing and social services in the origin and destination countries. The complexity of such a web of interactions could suggest that only qualitative research (of the type promoted by Marcus (1995)) has much to contribute to the illumination of the multiple interactions between remitting and returning plans. It is the view supported in this study that quantitative analysis could also elucidate such interactions by building composite indices and social typologies.

Another loop that complicates the analysis of this relationship is the possible influence of future estimated income and remittances as a conditioning factor for the current intention to return home. Its emergence is especially probable in times of crisis. Some immigrants come with rather precise plans regarding how much they will earn before returning home. Declining markets that bring fewer employment opportunities or lower incomes could impact expected incomes and, implicitly, return plans. It is difficult to say how frequently this occurs. However, the fact that the situation is plausible induces the probability of correlated errors between predicted remittances and return intention as a predictor.

Finally, it is relevant for the present discussion to note that remitting is more closely related to the probability of returning home than to the vaguely expressed intention to return home, or not (Sandu 2010a). This highlights the difficulties of measuring the relationship between remitting and planning to return home if the intention to return is measured in a very weak way, using only single questions such as ‘Do you intend to return home?’. Specifications indicating how structured the intention is in terms of probability of return and period foreseen could be very useful from this point of view.

Re-embedding remittances in lifeworlds

The complexity of measuring the relationship between remitting and intentions to return, together with the associated literature, indicates the need to expand the approach in order to integrate a new frame of reference. One such possible extension would be to adopt the lifeworld perspective (Schutz and Embree 2011). Lifeworld is ‘my world’, and ‘consists of my actual and previous experiences of known things and their interrelations (…) and certain more or less empty anticipations of things not experienced thus far, and therefore not known but nevertheless accessible to my possible experience’ (Schutz and Embree 2011: 170). It is formed by past and future experience, and by acts that are supported by ‘in-order-to’ and ‘because’ motives. Explicit or self-declared motives for the ongoing actions are of the in-order-to type. The because motives are inferred by the observer or the self after the action has been accomplished. The embeddedness view of remittances involves building scientifically on the two types of motive. Return plans are a proxy for in-order-to reasons for remitting. Past migration experiences or communication patterns with family left behind are a basis for inferring because motives.

The lifeworld perspective on immigration, with an explicit emphasis on in-order-to motives, was adopted many years ago in social history research in relation to the reasons that Eastern Central European peasants migrated to the United States more than a century ago (Morawska 1984). The lifeworld perspective in quantitative analysis, which is of primary interest in this study, can be located not so much in the area of in-order-to, but in the realm of because motivation. Research on the role of such factors as education, gender, ethnicity and duration of stay in the destination country (Carling 2008a, 2008b) frequently infers because motives from status predictors of remitting. State-of-mind variables at individual or at super-individual level – such as frustration or relative deprivation (Stark and Taylor 1991) – may be a good predictor of remittances. This would be in line with the requirements of the new economy of labour migration (Taylor 1999) that is the preferred framework for the theories used to explain remittances.

The key alternative approaches to exploring the lifeworlds of immigrants that are relevant for remittances involve the use of: in-order-to versus because motivations; correlates versus antecedent variables for remittances; typologies versus non-nominal variables; and one- versus two-level regression models. All these alternatives are used in the following sections of this paper.

More exactly, the lifeworld concept is operationalised in this article by: a) integrating economic remitting behaviour into a set of home-orientation variables together with return intentions and communication linkages; b) going beyond a linear measurement of home orientation and complementing it with a typology of orientations towards home (Tables 1 and 2); and c) using a large array of subjective variables (deprivation, identification, perceived effects of own migration on family, life satisfaction and life projects) as proxies for because motives relevant to home orientation.

Data and hypotheses

We used two complementary data sets to meet the objectives of this study. Encuesta National de Inmigrantes (ENI), a large survey of more than 15 000 immigrants in Spain, allowed comparison of the profiles of different types of structuring among remittances–return plans–communication patterns. Data were collected during the period November 2006 to February 2007 (Reher and Requena 2009) by a three-stage probability sampling with stratification in the first stage. The sampling frame for the survey was the population of non-natives over the age of 15 years. The sampling frame was the population register (Padrón Municipal). Post-data collection tests indicate this to be a representative sample of immigrants in Spain (Reher and Requena 2009). Unfortunately, the survey included only one question on return intentions (with or without plans to return to the country of birth during the next five years). For this reason, we also used a smaller but complementary data base of 832 Romanian immigrants in the Madrid area. This is known as the Romanian Communities in Spain (RCS) survey. It provides a more detailed measurement of return intentions, namely the probability and likely timescale of returning (for a description of this sample and its use see Sandu (2009b) and Şerban and Voicu (2010: 110)). The second advantage of this latter data set is that it allows for better proxies of lifeworlds to be considered in relation to home orientation behaviours. Many subjective variables that cover satisfaction – identification, perception of migration consequences, subjective probabilities to return home – are translated into survey questions only in the RCS and not in the ENI. A comparison between predictors in Table 5 (referring to ENI data) and Table 6 (based on RCS data) is relevant in this regard. The RCS survey collected data by respondent-driven sampling in September 2008 in the communities of Alcala de Henares, Arganda del Rey, Torrejon and Coslada. Comparisons between the two data sets are facilitated by the fact that the questionnaire for the RCS survey included questions adapted from or identical to the ENI survey.

The main dependent variable in the analysis is the home orientation of the immigrants, which is measured at a continuous level by an index of home orientation (IHORI) and in nominal terms by a typology of home orientation. The index is constructed from three indicators using the aggregation model proposed by Sandu (2010a, 2010b): a factor score of the logarithm of remittances sent home during the last year; the additive index¹ (with a range between 0 and 3) of intensity of communication with home by telephone, email and regular mail; and the intention to return home (3 – yes, 2 – undecided, 1 – no). Communication is considered to be of maximum intensity (3) if the immigrant declares that she/he uses all the three means of communication mentioned at least once a fortnight to contact people at home. The minimum would be the situation in which none of the communication means were used for at least a fortnight. The three means of communication are measuring the same latent dimension of communication intensity not only at the level of all the immigrants but also for each of the major groups of immigrants in Spain (Romanians, Moroccans, Colombians, Ecuadorians, Argentinians, Bulgarians, people from the EU-15 excluding the UK, and others) as recorded in the ENI. The three indicators that make up IHORI are not strongly related to each other in the case of the subsample of immigrants in Spain from the UK. Here, return intentions are independent of communication and remitting behaviours for British individuals, as they are mainly retired (25 per cent) or climate-attracted (54 per cent) immigrants.

The nominal variable for measuring home orientation was constructed by crossing the variables on communication, return intention and amount of remittances, after dichotomising them. This gives us eight types of home orientation in this property-space (Barton 1955), but if the least frequent cases are reduced to one category (labelled ‘other’), this produces six social types of home orientation (Table 1).

Table 1. Types of home orientation of immigrants in Spain, 2007

Source: ENI, 2007. N = 15 470. IHORI is the factor score rescaled to a range of 0–100 (as a Hull score).

The most frequent type of home orientation relates to immigrants who do not intend to return and do not send remittances at all (or only at a very low level), but communicate frequently with their home in the origin country. We called this ‘home-belonging orientation’, and assume that these individuals are keeping in touch with people at home not for pragmatic reasons (return plans or family arrangements summarised by remittances), but as a result of a well-structured feeling of belonging or similar symbolic reasons. The other two major types are ‘communication for remittances’ (with low values for intention to return associated with very high values for remittances and home communicating), and ‘generalised low home orientation’ in terms of remittances, intention to return and communication. The IHORI values are strongly differentiated between types, with a maximum value for ‘comprehensive orientation’ and a minimum, as expected, for ‘low home orientation’. The IHORI values are highest for the categories of ‘comprehensive home orientation’ and ‘communication for return home’. The social profile for each type will be reconstituted in the results section.

The first hypothesis (H1) relates to collective deprivation in remitting money. It states that immigrants from high collective-deprivation countries, irrespective of their personal deprivation at origin or in the destination country (Spain), will be more likely to have a high home orientation. The reverse should be the case for immigrants with a home orientation index that is lower than the sample average: they are more likely to come from low-deprivation countries. A collective deprivation index that measures the situation in the country of origin could be relevant from that point of view. The collective side of deprivation could favour the development of a culture of remitting in the diaspora communities of people coming from poorer countries. The expectation is in line with the approach contextualising remitting behaviours in terms of communities of practice and options setters (Page and Mercer 2012). It is likely that higher collective deprivation at origin contributes to a culture of remitting more money home and keeping more in touch with those left behind than immigrants from lower deprivation contexts. Hypothesis H1 is tested by ENI data.

The second hypothesis (H2) expresses the idea that home-orientation typologies are mainly differentiated in an agency space of life strategies and identities: immigrants who are focused on home return are more embedded in survival strategies; remittance-oriented immigrants act more in line with family development strategy; comprehensive home-oriented immigrants are those with high identification attitudes (Brubaker and Cooper 2000) and a greater perception of the negative consequences of their own migration on family members.

Push factors of dissatisfaction with life in the destination country are expected to be a reason for building a survival strategy rather than a development strategy and, implicitly, for returning rather than remitting. The identification typology of Romanian immigrants in the Madrid area differentiates between identification with Romania only, with Spain only, with both countries, and low country identification. This is consistent with the adaptation of the model of interethnic integration (Berry 1997) for describing the identities of immigrants by hybridisation, assimilation, segregation and marginalisation (Rother and Nebe 2009: 124). Hypothesis H2 is tested using the RCS data.

The third hypothesis (H3) assumes that home orientation as a quantitative variable is more relevant for the lifeworlds of immigrants than the remittances they are sending home. If this is correct, one would expect the same set of predictors to produce a higher percentage of explained variation for IHORI compared with remittances variables.

Deprivation as a key independent variable is measured in this study at a personal level at home – as material and cumulative deprivation – and in the host society with reference to investments and housing. A measure of collective deprivation in relation to the origin society is also devised. The package of deprivation variables are as follows:

• Aggregate data from the Survey on Income and Living Conditions in European Union, together with GDP per capita and life expectancy at birth, are used to assess the relationship between economic development and indices of material deprivation at society level. The resulting regression coefficients serve to estimate the material deprivation for all the countries that have immigrants in Spain.² As a result of the estimation method, collective deprivation for non-EU countries with immigrants in Spain is ‘collective deprivation in relation to GDP per capita and life expectancy at birth for the reference countries’.
• ‘Cumulative deprivation at origin’ is an additive index of the reasons for immigration (‘Why did you migrate to this country?’), in relation to job, education, quality of life, family, religion, politics, etc. It takes values between 0 and 9. The higher the value of the index, the higher the level of deprivation that motivated the person to migrate.
• ‘Material deprivation in the household at home’ is computed as an additive index of not owning a house, land, cattle, a business or a car (minimum deprivation coded by 0 and maximum coded by 5). The index is built by items that are relevant for relative deprivation (Stark and Taylor 1991) in the competition between non-migrants and returned migrants. The propensity of former migrants to invest in land, houses or businesses is systematically higher than for non-migrants (Sandu 2006: 158–159).
• ‘Housing environment deprivation’ in Spain is an additive index of the reasons for dissatisfaction with the environment of the individual’s house in Spain (‘Which of the following problems does your dwelling have?’): noise, bad smells, humidity, garbage on the street, vandalism in the area, poor communication opportunities, too small, building defects, etc. The index has a 10-point range.
• ‘Investment deprivation’ in Spain is indicated by assigning a number to reflect situations in which there has been no investment in dwellings, other durable goods, business, land, funds, etc. The index has a 7-point range.

Time predictors mark the distinction between duration of immigration in years and the period of arrival (up to 1989, 1990–1997, 1998–2001, 2002–2007). The distinction is adopted in line with attempts to investigate the specific effect of the period of arrival on remittances and transnationalism as distinct from that of the length of stay in the host country (Carling 2008a, 2008b; Sandu 2010b). The periods are delimited in order to consider important events or processes that could influence waves of emigration, such as the revolutions of 1989 in Europe, the opportunity for Romanians – one of the largest groups of immigrants in Spain – to circulate freely in the Schengen space after 2001, and the two most recent waves of accession to the EU in 2004 and 2007 (Carling 2008b). The annual rate of increase in immigration in Spain was highest, after 1998, in 2001 (42 per cent), 2002 (34 per cent) and 2003 (40 per cent), according to Eurostat figures. Immigration from Ecuador, Colombia and Romania had the highest rates of increase in this period, even if one disregards the effect of regularisation of illegal immigrants from 2001. Immigration from Morocco had a similar pattern (with a maximum increase in 2000). All these figures are indicative of the fact that 2001–2003 was a significant period for a new wave of immigration in Spain. The immigration wave of 2000/2003–2007 was in a ‘virtuous circle’ relationship with the economic growth in Spain during that period (Arango 2012).

Family lifeworlds and identification worlds are estimated by three typologies using data from the smaller sample of Romanian immigrants in the Madrid area. The first relates to the location of the majority of family members in the host or home country. About two-thirds of immigrants were in the host country with their families (Table A1 in the Annex 1). The second is the net perceived effects of emigration on the family members of immigrants. According to the data in Table A1, 47 per cent of Romanian immigrants to the Madrid area estimate that their emigration had predominantly positive effects on their family members. The proportion of those perceiving predominantly negative effects is 27 per cent, and the remaining 26 per cent perceive mixed effects. The proxy for the lifeworld at country level is a typology of dominant identification with the home, host or both countries. The largest group is that of immigrants who are mainly attached to their home country, Romania (40 per cent). The proportion of Romanian immigrants who are mainly attached to Spain is much lower, at 16 per cent. A significant proportion of immigrants (about one-third) have an ambivalent identification with Romania and Spain. The remaining proportion, about 10 per cent, is made up of people who have a low level of attachment to both Romania and Spain.

Mobility plans for Romanian immigrants in the Madrid area (Table A1), as a specific element of home orientation, are represented by five categories combining time horizon for return (soon or late) and the probability of return (high or low): no intention to return (29 per cent), late and unsure return (15 per cent), soon but unsure (14 per cent), late but sure (10 per cent), and soon and sure (32 per cent).

Data analysis and results

Understanding social types of home orientation

The proportions of each of the main groups of immigrants in Spain for each of the home orientation types are given in Table 2 (data from ENI). Moroccans were, at the time of the survey, the largest group of immigrants in Spain; the main home-orientation types for this group are communication for remittances (25 per cent) and low home orientation (24 per cent). Immigrants who send a large volume of remittances and have an intense communication with home come predominantly from five countries: Morocco, Ecuador, Colombia, Bulgaria and Romania. All these are societies with a high level of deprivation. Immigrants from societies with low levels of material deprivation account for a very small proportion of this category of remittance orientation; British immigrants, for example, who come from a society of low collective deprivation, have the highest and most specific concentration in the category of home-belonging orientation. Immigrants from other EU-15 countries with a low deprivation index are also significantly clustered in the same social type of home orientation. All these findings are clearly consistent with the expectations derived from the first hypothesis (H1): the ‘communication for remittances’ social type of home orientation is specific to immigrants coming from societies with high levels of material deprivation in Africa, Latin America and Europe. The focus on remittances in home orientation is not confined to immigrants from societies that have sent migrants to Spain recently. Moroccan immigration is a much older trend than those of Romania and Bulgaria. The average length of stay in Spain for Moroccans was about 14 years at the time of the survey, while it was only four years for Romanians. In spite of these dissimilarities in terms of time of arrival in the host country, the two groups of immigrants make up a large proportion of the communication for remittances social type. In contrast, the symbolic communication structured around home-society belonging is specific to developed, low-deprivation societies.

Table 2. Typology of home orientation by country/region of origin for immigrants in Spain, 2007

Source: ENI, 2007, weighted data (reduced by dividing weighting factor by its mean in order to keep the sample size constant). Shadowed cells indicate a significant association between column and row values – adjusted standardised residuals that are significant for p = 0.05. Own computations. * In a special cross-tab intersecting origin country/region and period of arrival, adjusted standardised residuals were computed. The periods specified in this column correspond to the cell of significant associations from the above-mentioned table.

Communication for returning home as a social type is specific to those immigrants from low-deprivation societies in Europe (EU-15 countries) or the Latin America macro-region (Argentina). Comprehensive home orientation is specific to some groups of immigrants from Ecuador, Romania and the new Member States that acceded to the EU in 2004.

It is clear that the probability of inclusion in different social types cannot be explained only with reference to collective deprivation or the time of arrival in the host society. Resources, personal deprivation and migration experience are also relevant factors. A multinomial regression model (not presented in the text)³ with types of home orientation as a dependent variable integrated predictors from all the above-mentioned areas (income, tertiary education, ability to speak Spanish very well, gender, young person, investment deprivation in Spain, cumulative deprivation before emigration, collective deprivation in the country of origin, immigration during the period 2002–2007).

According to the results of this multinomial regression on ENI data, collective deprivation continues to be a significant predictor for all five social types of home orientation even if all the other predictors are considered to be control variables.

Table 3. Role of different types of deprivation in explaining home orientation types

Source: ENI, 2007. Relations in multinomial regression between deprivation predictors and types of home orientation as dependent variables, controlling for income, education, age, gender and ability to speak Spanish: + significant, positive relation, - significant negative relation, 0 insignificant relation for p = 0.05. The reference category in the dependent variable is the residual one of ‘other categories’ of home orientation. Detailed data on the regression model are not included in the text.

High collective deprivation in the origin society increases the likelihood of inclusion in the comprehensive home orientation and remittance-structured communication categories. The likelihood of immigrants being included in all the other categories (return intention, home belonging and low home orientation) is increased by low values of collective deprivation.

Different types of personal deprivation affect home orientation types differently. A high level of dissatisfaction at the time of emigration (‘cumulative deprivation’) has a significant impact in terms of reducing the propensity for return intention and home-belonging orientation. A high degree of frustration in relation to opportunities to invest in Spain fosters comprehensive home orientation and return intentions.

Comprehensive home orientation is associated with high-income and materially successful immigrants coming from poor countries. The same analysis indicates that low home orientation is associated with low-income immigrants who have come from more developed societies or who reached the destination society earlier.

Immigrants who are home oriented in terms of remittances and communication are similar to those characterised by comprehensive home orientation (Table 4). Individuals in both categories come from high-deprivation countries and have higher incomes as immigrants. What is specific to remittance-oriented immigrants is their greater ability to speak Spanish. This is an easy-to-convert human capital resource that allows immigrants to earn more. The level of formal education per se is not relevant in terms of inclusion in the two categories. It is only for remittance-oriented immigrants that knowledge of the host country counts.

Table 4. Multinomial regression predicting types of home orientation for Romanian immigrants in the Madrid area, 2008

Source: RCS, 2008. * p ≤ 0.10, ** p ≤ 0.05, *** p ≤ 0.01 (two-tailed tests). Computing algorithm for positive or negative effects of own migration on family and also for the index of material goods are specified in Sandu (2009a).

The second hypothesis, on the role of survival–development–identification strategies, could be tested only for the sample of Romanian immigrants around the Madrid area (Table 4, Table A1, RCS data set).

Romanian immigrants who are return-home oriented are more influenced by the negative than by the positive consequences of their emigration on their family members. They are also significantly dissatisfied with their life in Spain. These findings suggest that they plan to return home as a kind of survival strategy, as a project to reduce the negative consequences of their emigration on the family, and to reduce their dissatisfaction with their life in Spain. The opposite is true for immigrants who are focused on remitting. They perceive their emigration as being more positive than negative because of its consequences for their own families, and they are also satisfied with the income they receive in Spain. Hence, their life strategies are more in line with the idea of family development than with survival. It is only for immigrants in the comprehensive home orientation category that country identification plays a significant role. There is a higher probability that Romanian immigrants in the Madrid area who identify with Romania or with Romania and Spain will be comprehensively oriented towards home.

Inclusion in the category of home-belonging orientation is the least understood phenomenon. It has the smallest number of recorded significant predictors in the multinomial regression model (Table 4). The immigrants in this category have in their specific profile the highest rate of internet use (79 per cent compared to the average of 52 per cent in the whole RCS sample) and a very low rate of identification with their area of residence in Romania (31 per cent compared to 46 per cent for the whole sample).

The five types of home orientation have family and national culture identification markers as predicted by hypothesis H2 (Table A1).

Understanding home orientation and remittances

The third hypothesis tests the idea that in reality, remittances function as part of a larger set of variables, not as a purely economic component. The comparison of the two regression models for two independent samples in Tables 5 and 6 supports the expectations derived from this hypothesis: the same set of predictors explains the variation of IHORI to a greater degree than the variation of remittances sent home. Table 5 presents the regressions for the large samples from different ethnic groups of immigrants in Spain (ENI). Table 6 uses data from the smaller sample of Romanian immigrants in the Madrid area (RCS). Although the predictors in the two tables are different, they refer to the same large categories of status variables (age, gender, education, income, ability to speak Spanish, type of family, etc.), frustration variables and arrival time in Spain. For the smaller sample of Romanian immigrants in the Madrid area, the set of available predictors is more extensive, and includes more variables relating to satisfaction, geographic identification and community location.

Multiple determination (R²) is 7 percentage points higher in terms of explaining IHORI than the variation of remittances for the large ENI sample of immigrants from different countries in Spain (Table 5). The difference is much higher between the explained variation for IHORI (R² = 0.41) and for remittances (R² = 0.17) as a dependent variable for the case of Romanian immigrants in the Madrid area (Table 6). This finding indicates that remittances are more meaningful in social life when they are considered together with behaviours of communication with home and intentions to return home. The regression models on the RCS data set are more clearly specified, as the survey in the Madrid area was explicitly focused on return migration projects and had a larger set of available predictors. The large data set of immigrants from all origin countries (ENI) was mainly descriptive by design and offered fewer opportunities to identify predictors that are relevant for IHORI or remittances.

Table 5. Predicting home orientation and remittances for immigrants of different ethnicity in Spain, 2007

Source: ENI, 2007. OLS regression in STATA using cluster option to correct for similarity profile of immigrants from the same province of Spain. 52 clusters. Suspicions of collinearity are dismissed by the very low values of VIF (mean value of 2.05, maximum value of 4.90).

Variables: * dummy variables; reference category for immigration period 1990–1997.

Time variables are particularly relevant for IHORI in the case of the analysis of all categories of immigrants (ENI): the elimination of three period-effect variables from the home orientation regression decreases its explanatory power by 10 percentage points (from 28 per cent to 18 per cent). The same type of elimination of the wave predictor in the regression of remittances produces a very small decrease in the explanatory power of the model (Table 5). This simple comparison is a sufficient indicator that the time of arrival of immigrants has a greater impact on the cluster of behaviours measured by IHORI than on the isolated component of it referring to remittances.

Time variables have no relevance in explaining variation for IHORI in the case of the Madrid area sample (Table 6). This could be because Romanians in the Madrid area are not as heterogeneous in terms of their arrival time as immigrants to Spain from the origin countries as a whole.

Table 6. Predicting home orientation and remittances for Romanian immigrants in the Madrid area, 2008

Source: RCS, 2008. OLS regression. Suspicions of collinearity are dismissed by the very low values of VIF (mean value of 1.38, maximum value of 0.97).* p ≤ 0.10, ** p≤ 0.05, *** p ≤ 0.01 (two-tailed tests).

Variables: * dummy variables. ** Coslada and Arganda del Rey are two of the four communities of Romanians around Madrid included in the survey. The Romanians in the local population are concentrated most heavily in these two communities (Şerban 2011: 149). Immigrants in Coslada come mainly from the historical region of Muntenia in Romania, and those in Arganda del Rey come mainly from Transylvania, another historical region of Romania (Sandu 2010b: 127).

The cluster of home-orientation behaviours are embedded in the contexts not only of time and level of deprivation but also of place of origin. IHORI tends to be significantly higher for immigrants from Latin America and significantly lower for those coming to Spain from Morocco. Places such as the old European Union (EU-15) or Romania do not condition per se, in a significant way, the values of the complex of home-orientation behaviours.

The ability of immigrants to speak Spanish has different impacts on home orientation. Its impact on the whole community of immigrants in Spain appears to be positive if one controls for ethnicity and other status predictors (Table 5). A more detailed analysis for each large group of immigrants produces a more nuanced picture:⁵ immigrants from Latin America or countries in the EU-15 are in the particular situation of being more home oriented if they speak Spanish better; the impact of Spanish-speaking abilities is insignificant for IHORI in the case of Moroccans. The more clearly specified regression model for the Romanian immigrants in the Madrid area indicates a higher home orientation for those with lower ability levels in Spanish (Table 6). The pattern could be specific to groups with less experience of migration.

Life dissatisfaction and individuals’ perception of the negative consequences for their family of their own migration tend to increase the home orientation of immigrants (Tables 5 and 6). Dissatisfaction with job and health in the host country in particular contribute to an increase in home orientation. The only type of dissatisfaction that seems to act in a different direction is that relating to income. Immigrants who have lower earnings and are dissatisfied with their income have lower home orientation as expressed by return intentions, communication frequency with home and remittances sent back.

The results of bivariate analysis (Table A1) on the role of cultural variables are also supported by regression analysis on the Madrid area data set. A higher level of identification with the origin country and a higher degree of ambivalent identification with both origin and host countries contribute to strong home orientation of immigrants. Return, communication and remitting behaviours have higher probabilities not only for those who are attached to their home country, but also for those who have an ambivalent cultural orientation towards the home and destination countries. The same types of behaviour are supported by different cultural attitudes in terms of national and transnational identifications.

Social ties at family and non-family level are significant predictors of home orientation and remittances: higher values for IHORI and sending remittances are associated with those immigrants who have a larger proportion of family members, a spouse and larger network capital in their home country (Table 6).

Conclusions

The index of home orientation of immigrants (IHORI) has a systematic variation under the influence of collective deprivation (as expected under H1), life strategies and identification (in accordance with H2). Irrespective of many other control variables, IHORI tends to be higher for immigrants in Spain coming from countries of higher collective deprivation. The third hypothesis (H3) is also supported by the data: home orientation behaviour, compared to its component of remitting money home, is much more deeply rooted in the social worlds of immigrants (as predicted by the H3 hypothesis). This is demonstrated by the much greater explained variation of home orientation compared with the explained variation of remitting behaviours when the same sets of predictors are used in multiple regression models. All these hypotheses are tested with positive results by running the same algorithms on two different data sets (ENI and RCS).

The testing results are important as they contribute to the expansion of the set of predictors of remittances and home orientation to areas that were previously neglected. Such neglected factors refer mainly to deprivation, perceived consequences of own emigration on family members, and time of arrival at destination. Some of these factors could be measured at different levels and proved to have effects function of measurement level. This is especially the case for deprivation when measured at national or collective level, in contrast to personal level. Migration and remittances are, in the light of the results of this analysis, multilevel phenomena, with specific variation function of individual, household, community and national scale.

The expanded explanation of home-orientation behaviours covers not only the variation of IHORI as a quantitative variable but also its nominal expression as given by the typology of home orientations. These types send to consistent clusters of behaviours.

Remitting behaviour is not only part of a home orientation set of behaviours. It is also indicative for social types of immigration practices. Immigrants who are focused on sending remittances home have a specific profile compared to other types of immigration practices (return-home orientation, home-belonging orientation, comprehensive orientation towards origin country, and low home oriented: Tables 1, 2 and A1). Remittance-focused immigrants are more inclined to be ambivalent in terms of their attachments to the origin and destination countries, have more of their family members in the origin country, and come from countries with higher levels of material deprivation (in accordance with hypothesis H1). Collective deprivation in the origin country is associated not only with the adoption of a remittance focus, but also with the comprehensive type of home orientation.

A social type that has high symbolic value is that relating to immigrants who are focused on home-belonging orientation. The immigrants in this category communicate intensively with home, though not for the purposes of returning or of sending remittances. They communicate for communication’s sake or, more exactly, for reasons not measured in the research, such as family solidarity or homesickness.

There is a high level of association between home orientation types and their geographic or national identification: return and comprehensive orientations are associated with immigrants who are attached to their origin country; remittance-focused immigrants are mainly ambivalently oriented towards their home and host societies; Romanian immigrants who are especially attached to Spain are characterised by home-belonging orientation or by practices of low home orientation.

Home orientation in behavioural and quantitative terms proved to be as consistent as its qualitative counterpart measure of home sense (Wiles 2008). These related concepts capture the symbolic universe that the migrants confer on their place of origin when in host countries.

The typological analysis of home orientations diverges from the standard approach that supports the view that ‘cross-border activities and exchanges do not cluster together’ (Waldinger 2008: 24). It argues for the fact that cross-border activities cluster together in different ways for specific social types. This clustering is frequently non-linear, by specific social types of home orientation.

Notes

¹ ‘Additive index’ in this article refers to a measure that sums several unweighted dummy variables (coded by 1 for the presence of the attribute and by 0 for its absence). The summing is equivalent to counting a set of pre-established values across several variables for the same unit of analysis.

² Material deprivation (MATDEPRIV) is estimated based on regression equation for 26 EU countries, EU-SILC data 2007: MATDEPRIV = 327.4 – GDPpc * 43.5 + LIFEexpectancy at birth for 2007 * 31.3. R2 = 0.81.The starting values of material deprivation (3 or more items) by country refer to 26 countries of EU (excluding Luxembourg with a very high value), for 2007 (source: Eurostat http://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=ilc_sip8&lang=en).

³ The multinomial regression model gave a pseudo R squared value of 0.12. It was run in STATA with cluster option to correct standard errors, function of province of residence of immigrant in Spain. The reference category for the dependent variable is the ‘other’ type.

4 The findings referring to the relationship between language abilities and IHORI derived from running the regression model from Table 3 by each specified group of immigrants with ENI data.

References

Arango J. (2012). Exceptional in Europe? Spain’s Experience with Immigration and Integration. Washington, DC: Migration Policy Institute.

Barton A. H. (1955). The Concept of Property-Space in Social Research, in: P. F. Lazarsfeld, M. Rosenberg (eds), The Language of Social Research, pp. 40–53. Glencoe, IL: Free Press.

Berry J. W. (1997). Immigration, Acculturation, and Adaptation. Applied Psychology 46(1): 5–34.

Blumer H. (1956). Sociological Analysis and the ‘Variable’. American Sociological Review 21(6): 683–690.

Brubaker R., Cooper F. (2000). Beyond ‘Identity’. Theory and Society 29(1): 1–47.

Carling J. (2008a). The Determinants of Migrant Remittances. Oxford Review of Economic Policy 24(3): 581–598.

Carling J. (2008b). Toward a Demography of Immigrant Communities and Their Transnational Potential. International Migration Review 42(2): 449–475.

de Haas H. (2005). International Migration, Remittances and Development: Myths and Facts. Third World Quarterly 26(8): 1269–1284.

Dustmann C., Mestres J. (2010). Remittances and Temporary Migration. Journal of Development Economics 92(1): 62–70.

Gentry J. W., Mittelstaedt R. A. (2009). Remittances As Social Exchange: The Critical, Changing Role of Family As the Social Network. Journal of Macromarketing 30(1): 23–32.

Levitt P., Lamba-Nieves D. (2011). Social Remittances Revisited. Journal of Ethnic and Migration Studies 37(1): 1–22.

Marcus G. E. (1995). Ethnography in/of the World System: The Emergence of Multi-Sited Ethnography. Annual Review of Anthropology 24: 95–117.

Mau S. (2012). Social Transnationalism: Lifeworlds Beyond the Nation-State. London: Routledge.

Miller A. (2013). Do Workers' Remittances Bring Economic Growth to Receiving Countries? A Macro Panel Analysis. Grin Verlag Ohg.

Morawska E. (1984). 'For Bread with Butter': Life-Worlds of Peasant-Immigrants from East Central Europe, 1880–1914. Journal of Social History 17(3): 387–404.

Özden Ç., Schiff M. W. (2007). International Migration, Economic Development & Policy. Washington, DC: World Bank Publications.

Page B., Mercer C. (2012). Why Do People Do Stuff? Reconceptualising Remittance Behaviour in Diaspora–Development Research and Policy. Progress in Development Studies 12(1): 1–18.

Ratha D. (2013). The Impact of Remittances on Economic Growth and Poverty Reduction. Policy Brief 8: 1–13.

Reher D., Requena M. (2009). The National Immigrant Survey of Spain. A New Data Source for Migration Studies in Europe. Demographic Research 20(12): 253–278.

Rother N., Nebe T. (2009). More Mobile, More European? Free Movement and EU Identity, in: E. Recchi, A. Favell (eds), Pioneers of European Integration. Citizenship and Mobility in the EU, pp. 120–155. Cheltenham, Northampton: Edward Elgar Publishers.

Sandu D. (ed.) (2006). Living Abroad on a Temporary Basis: The Economic Migration of Romanians: 1990–2006. Bucharest: Open Society Foundation.

Sandu D. (2009a). Return Migration, As a Project and As a Mentality, in: D. Sandu (ed.), Romanian Communities in Spain, pp. 43–69. Bucharest: Soros Foundation Romania.

Sandu D. (ed.) (2009b). Romanian Communities in Spain. Bucharest: Bucharest: Soros Foundation Romania.

Sandu D. (2010a). Home Orientation in Transnational Spaces of Romanian Migration. Studia Sociologia 2: 15–36.

Sandu D. (2010b). Lumile sociale ale migratiei românesti în strainatate. Iasi: Polirom.

Schutz A., Embree L. (2011). Collected Papers V. Phenomenology and the Social Sciences. Dordrecht, London: Springer.

Şerban M. (2011). Dinamica migratiei internationale: un exercitiu asupra migratiei romanesti in Spania. Bucharest: Lumen.

Şerban M., Voicu B. (2010). Romanian Migrants to Spain: In- or Outside the Migration Networks – A Matter of Time? Revue d’études comparatives Est-Ouest 41(4): 97–124.

Stark O. (1999). Altruism and Beyond: An Economic Analysis of Transfers and Exchanges Within Families and Groups. Cambridge: Cambridge University Press.

Stark O., Lucas, R.E.B. (1988). Migration, Remittances, and the Family. Economic Development and Cultural Change 36(3): 465–481.

Stark O., Taylor J. E. (1991). Migration Incentives, Migration Types: The Role of Relative Deprivation. The Economic Journal 101(408): 1163–1178.

Taylor J. E. (1999). The New Economics of Labour Migration and the Role of Remittances in the Migration Process. International Migration 37(1): 63–88.

UNDP (2009). Human Development Report 2009. Overcoming Barriers: Human Mobility and Development. New York: United Nations Development Programme.

Vullnetari J., King R. (2011). Gendering Remittances in Albania: A Human and Social Development Perspective. Gender & Development 19(1): 39–51.

Waldinger R. (2008). Between ‘Here’ and ‘There’: Immigrant Cross‐Border Activities and Loyalties. International Migration Review 42(1): 3–29.

Wiles J. (2008). Sense of Home in a Transnational Social Space: New Zealanders in London. Global Networks 8(1): 116–137.

Annex 1

Table A1. Home orientations by contextual factors related to family, migration effects and identification

Source: ENI, 2007, weighted data (reduced by dividing weighting factor by its mean in order to keep the sample size constant). Shadowed cells indicate a significant association between column and row values – adjusted standardised residuals that are significant for p = 0.05. Own computations. * In a special cross-tab intersecting origin country/region and period of arrival, adjusted standardised residuals were computed. The periods specified in this column correspond to the cell of significant associations from the above-mentioned table.