Transnational ties: Resource or stressor on Peruvian migrants' well-being?

This paper explores the role of types and the intensity of transnational ties for migrants' well-being from a global perspective. Based on a literature review, two competing hypotheses are formulated — transnational resources versus transnational stress — according to which transnational ties have either a positive or a negative effect on migrants' well-being. Drawing on data from a large-scale survey of Peruvian migrants worldwide, this paper examines the strength and direction of the relationship between Peruvian migrants' transnational ties and poor well-being, the latter measured as depression/loneliness as a principal concern. While the multivariate regression results do not support the transnational resources hypothesis, partial support is found for the transnational stress hypothesis: more intense transnational ties are positively associated with poor well-being. Our study points the importance of considering transnational ties in research on migrants' well-being and indicates the relevance of developing adequate measurements and longitudinal research designs to explore the causal relationships between migrants' well-being and transnational ties.

migrants encounter within but not beyond the borders of their country of residence. Yet, as findings from still-scarce research suggest, using single or multiple indicators of physical health (including self-rated health and chronic condition) and/or mental health (including depressive symptoms and loneliness), transnational ties might also be of relevance for migrants' well-being (Alcántara, Chen, & Alegría, 2015;Klok, van Tilburg, Suanet, Fokkema, & Huisman, 2017;Torres, 2013;Torres, Alcántara, Rudolph, & Viruell-Fuentes, 2016).
Conflicting theoretical arguments have been developed on the implication of transnational ties as either protective or risk factors for migrants' well-being (Torres, 2013;Torres, Lee, Gonzalez, Garcia, & Haan, 2016). According to Torres, Lee, et al. (2016), transnational ties may serve as a source of either resilience or risk for migrants' well-being as they may mitigate as well as exacerbate the negative well-being outcomes of migration-related stress. The present study builds on this theoretical foundation to investigate the role of transnational ties on Peruvian migrants' well-being, the latter measured as depression/loneliness as a principal concern.
By exploring the role of transnational ties on Peruvian migrants' well-being worldwide, this study provides the first global approach in the field. Hitherto empirical evidence stems primarily from research on migrant populations residing in the same nation-state and is often limited to rather specific groups such as migrant mothers, younger adults or older adults (Klok et al., 2017;Miranda, Siddique, Der-Martirosian, & Belin, 2005;Murphy & Mahalingam, 2004;Ward & Styles, 2012). Hence, the findings cannot be extrapolated to the wider migrant community or to migrant populations residing in different host societies. Moreover, while not focusing on a specific age group or type of migrant (e.g., mothers), the study offers a more inclusive picture of the role of transnational ties for migrants' well-being. Unlike most previous studies, it further takes systematically into account the intensity of migrants' transnational ties. We define intense as frequent engagement in transnational practices (family visits and remittances) as well as a strong sense of belonging to the country of origin (permanent return intention and purchase of various Peruvian articles). Finally, the study contributes to theoretical advancement by operationalising transnational ties not only as a structural condition and/or transnational behaviour but also as belonging.

| THE GLOBALISATION OF PERUVIAN MIGRATION
International Peruvian migration represents a rather recent development. Although the first migration flows of Peruvians can be traced back to the 1920s (Altamirano, 1999), it was not until the late 1990s that they turned into a mass phenomenon. The outflow peaked between 2006 and 2010, with more than 200,000 Peruvians leaving their country per year on average (INEI, 2015). A particular characteristic of international Peruvian migration is the diversity of destinations (Paerregaard, 2010): unlike migrants from other Latin American or Caribbean countries, Peruvians are dispersed across many countries.
The United States was the preferred destination of Peruvian migrants, counting at present for one third of the approximately 2.6 million Peruvian migrants worldwide (INEI, 2015), but has become less accessible since the mid-1980s due to tightened migration policies and stricter border controls (Takenaka, Paerregaard, & Berg, 2010). At the same time, countries like Japan, Italy and Spain began to facilitate the entry of Peruvian migrants in order to satisfy those countries' growing need for labour in the manufacturing industry and the care and domestic work sector (Escrivá, 2003;Takenaka, 2005;Tamagno, 2003). More recently, Argentina and Chile have also become popular destinations, especially for working-class Peruvians.
So far, the transnational ties of Peruvian migrants in different host countries have been explored; this includes the United States (Berg, 2015), Spain (Díaz Gorfinkiel & Escrivá, 2012), Chile (Nuñez Carrasco, 2010) and Italy (Skornia, 2014). These qualitative observations point to a variety of modes through which Peruvian migrants maintain contact with their significant others in Peru, including regular information and communications technology (ICT)-based communication, the sending of gifts and remittances and frequent visits to the country of origin (Horn, 2019). Peruvian migrant women have been shown to feel particularly responsible for the lives and fates of their family members 'left behind' (Leinaweaver, 2010;Skornia, 2014).
These findings align with literature on transnational motherhood (Ehrenreich & Hochschild, 2003;Erel, 2002;Hondagneu-Sotelo & Avila, 1997), which emphasises the gendered nature of caregiving within transnational family contexts. Quantitative studies on Peruvian migrants' cross-border activities support these findings, revealing a positive association between being female and frequent visits to the country of origin (Horn, 2017). Similarly, Peruvian migrant women have been shown to remit a larger share of their monthly income than their male counterparts do (Horn, 2016).

| TRANSNATIONAL TIES AND MIGRANTS' WELL-BEING: EMPIRICAL AND THEORETICAL CONSIDERATIONS
This study draws on research at the intersection of transnational family and well-being literature, including both studies on mental and physical health. As argued by Mazzucato (2015), combining these two bodies of literature allows for a holistic analysis of migrants' well-being in transnational (family) contexts. The value of this combination derives from the very different theoretical and methodological foundations underpinning research in the two fields.
Studies on migrants' well-being are largely coloured by methodological nationalism according to which nation-states are the quasinatural container or fundamental frame for social practices (Wimmer & Glick Schiller, 2002). As a consequence, theoretical premises and methodological approaches in the field tend to overlook dimensions in people's lives that go beyond the borders and boundaries of a defined national territory. In other words, the focus of well-being literature lies primarily on factors related to the migrants' country of residence, whereas transnational ties and family formations often go unnoticed. Nevertheless, a small but increasing body of literature exists, which examines and theorises about the implications of transnational ties for migrants' well-being. This set of studies comprises mainly quantitative (Afulani, Torres, Sudhinaraset, & Asunka, 2016;Alcántara et al., 2015;Murphy & Mahalingam, 2004;Torres, Alcántara, et al., 2016) but also a few qualitative (Viruell-Fuentes, 2006;Viruell-Fuentes & Schulz, 2009) contributions.
Transnational family studies, by contrast, are inspired by the transnational paradigm in migration research, according to which migrants, 'through their daily activities, forge and sustain multi-stranded social, economic and political relations that link together their societies of origin and settlement' (Basch, Glick Schiller, & Szanton-Blanc, 1994, p. 6). Methodologically, transnational family scholars predominantly employ qualitative research approaches to examine the well-being of migrants and their significant others (Boccagni, 2015;Dreby, 2010;Menjívar, 2012). In doing so, they tend to scrutinise the complexities of transnational family dynamics and practices and how they transform in the context of spatial separation (Nyberg Sørensen & Vammen, 2014). Accordingly, large-n studies on migrants' well-being are still rare. A notable exception is a recent study of Mazzucato et al. (2017) that compares transnational Angolan parents in Portugal and the Netherlands with their nontransnational counterparts.

| Transnational ties as a protective factor
Several scholars found a positive relationship between transnational ties and migrants' well-being, suggesting a protective function of personal cross-border relationships. Studying younger Latin adults in Los Angeles, Torres (2013) found that having a relative in the country of origin was positively associated with self-rated health among foreign-born but not US-born respondents. The study further showed that parents' past engagement in sending remittances was associated with an increased probability of good overall health of US-born Latino adults compared with their American-born counterparts without a parental remittance history. Alcántara et al. (2015) also found a positive association between the sending of remittances and migrants' well-being. Analysing Latin American migrants in the United States, these scholars showed that an increase in the burden of remittances to relatives is associated with a decrease in the odds of past-year depressive disorder.
A possible explanation for the positive relationship between transnational ties and migrants' well-being is that maintaining links with the country of origin contributes to a stronger sense of ethnic identity (Torres, Lee, et al., 2016). Besides a potentially direct, protective effect, a strengthened ethnic identity may also serve as a buffer against perceived discrimination and other psychosocial stressors by improving self-esteem (Mossakowski, 2003;Torres & Ong, 2010).
Theoretically, this mechanism can be explained by the rejectionidentification model (Branscombe, Schmitt, & Harvey, 1999), according to which the negative effect of perceived discrimination on self-esteem can be buffered or compensated by strong identification with the in-group. In transnational migration studies, this mechanism has been conceptualised as 'reactive transnationalism' (Itzigsohn & Giorguli-Saucedo, 2002). There is empirical evidence that migrants experiencing discrimination, particularly men, show a higher propensity to engage in ties with their country of origin (Itzigsohn & Giorguli-Saucedo, 2005). This suggests that migrants facing rejection by the majority society tap social resources not only within the local ethnic community but also beyond national borders.
In a similar vein, Viruell-Fuentes and Schulz (2009, p. 2171) argue that keeping in touch with those living in the country of origin may offer migrants 'an alternative space of belonging' . This sense of belonging is nurtured and fostered primarily through support exchanges between migrants and those relatives left behind (Baldassar, 2008;Wilding, 2006). It leaves migrants feeling understood, valued and cared for by others, and empowered to cope with inequity, feelings of social isolation, discrimination and other challenges they might face in the host society-all of which lead to improved well-being (Torres, 2013). Remittances, in turn, are one of migrants' most common and important types of support (Carling, 2014). Sending money and goods is a way of showing the attachment migrants have to their left-behinds and may serve as a marker of having fulfilled one's obligation of being the family's financial provider (Torres, Lee, et al., 2016), with an improvement of wellbeing in the bargain. Finally, those migrants who maintain close links with their country of origin develop a keen awareness of their privileged socio-economic situation and perspectives compared with their left-behinds (Nieswand, 2011). This may lead to improved self-esteem, a heightened sense of control and a greater 'sense of purpose' (Torres, Alcántara, et al., 2016, p. 6) and meaning in their lives, which in turn have positive implications for well-being outcomes (Alcántara et al., 2015;Jin, Wen, Fan, & Wang, 2012).

| Transnational ties as a risk factor
The finding of a positive impact of transnational ties on migrants' well-being is not consistent across the studies in this field. Investigating Sub-Saharan migrants in metropolitan France, Afulani et al. (2016) found that women who sent remittances had 50% higher odds of reporting a chronic condition than nonremitting women. As argued by different scholars, the persistent connection to the country of origin may be indicative of long-term cross-border separation from significant others (Afulani et al., 2016;Menjívar, 2012;Miranda et al., 2005), such as parents, children, a favourite brother or sister or close friends. Similarly, Mazzucato et al. (2017) found that geographical separation from children is negatively associated with migrants' well-being. This study did not control for the relationship between well-being and other forms of transnational belonging or the engagement in particular transnational practices though. The negative relationship with spatial separation is not surprising, as it is well-known from loneliness literature that different types of relationships are not or only barely interchangeable, as they serve different, more or less unique functions (Weiss, 1974). Hence, social contacts in the host society do not fully remedy the emotional hardship of the absence of beloved persons who live in the country of origin and the negative effects on well-being, especially for women (Alcántara et al., 2015). It could also be argued that intense transnational ties (e.g., frequent communication or sending remittances on a regular basis) may hamper migrants from building up significant social relationships in the host society (Baumeister & Leary, 1995). In this scenario, migrants' well-being would be negatively affected by a lack of reliable social support structures in their immediate environment.
Another possible explanation for a negative effect of transnational ties on migrants' well-being is that being engaged in transnational practices is not always entirely voluntary. Left-behind family members more or less impose a moral demand on their migrant relatives to make frequent calls, send substantial amounts of money on a regular basis and visit their homeland regularly, bearing gifts and attending social and familial activities during their stay (Krzyzowski & Mucha, 2014;Mazzucato, 2008). This is confirmed by the positive association of frequent visits to the country of origin with the risk of depressive disorders among migrant women (Alcántara et al., 2015).
As argued by Baldassar (2015), migrants' transnational practices such as family visits are intrinsically linked to the guilt caused by the act of leaving family behind. This feeling of guilt is likely to impact the migrants' well-being and their families by association (Ward & Styles, 2012). In this context, transnational practices often serve as an assessment of the quality of relationships: when migrants fail to meet left-behinds' expectations, they are likely to be labelled as unreliable or irresponsible and the left-behinds are likely to feel ignored and abandoned (Licoppe, 2004). High expectations can also lead to a withdrawal from transnational ties when migrants feel ashamed or overburdened by not being able to fulfil expectations (Mazzucato, 2008).
In addition, engaging in transnational practices bears an inherent risk of ambivalence and emotional distance, as shown in the literature on ICT-based communication (Marchetti-Mercer, 2017;Nedelcu & Wyss, 2016;Tarrant, 2015). Information asymmetries, coupled with resource inequality, between the migrant and those left behind can lead to misunderstandings, disagreements and conflicts (Carling, 2008).
Moreover, to protect significant others, there is a reluctance to share personal problems or bad news (Zentgraf & Stoltz Chincilla, 2012). As a result, remote interactions are likely to become gradually more superficial, limited to a handful of routine exchanges (greetings, pleasantries, good wishes and passing orders). Finally, besides positive impacts on well-being, maintaining a strong ethnic identity might also cause feelings of uprootedness or in-betweenness, and therefore a decreased sense of belonging either 'here' or 'there' . Moreover, if ethnicity is the core part of one's overall identity, discrimination that particularly targets the specific characteristics of this ethnic identity might pose a threat to one's well-being (Yip, Gee, & Takeuchi, 2008).

| Hypotheses
Previous findings and theoretical approaches give rise to two hypotheses with opposite predictions regarding the impact of transnational ties on the well-being of Peruvian migrants. On the one hand, transnational ties improve migrants' self-esteem and contribute to retaining their ethno-identity. Transnational ties further serve as reference points that enable migrants to gain a favourable status through comparisons with their left-behinds. Finally, transnational ties provide migrants with a source of identity and social support alternative (or additional) to those provided (or not) by their immediate environment. When these effects dominate, then transnational ties lead to good well-being. We call this the transnational resource hypothesis. On the other hand, transnational ties stir the emotions of long-term separation from relatives and friends, nurturing feelings of loss, longing and nostalgia through the recurring awareness of one's absence. At the same time, they amplify feelings of financial and social obligations, putting pressure on migrants to act according to their transnational families' expectations. Moreover, keeping transnational ties causes or strengthens feelings of 'uprootedness' and 'identity crisis' ('betwixt and between' identities and 'double absence'), and therefore a decreased sense of belonging to host society.
When these effects dominate, then intense transnational ties lead to poor well-being. We call this the transnational stress hypothesis. We expect the transnational stress hypothesis to apply to more intense transnational ties, such as frequent family visits or permanent return intentions, and the transnational resource hypothesis to apply to less intense transnational ties.

| Data source
This study draws on data taken from the first worldwide survey on The principal selection criterion for our sample was whether respondents were first-or second-generation migrants. Although only a small proportion (0.2%) was born abroad, these respondents were filtered out to increase the homogeneity of the sample. Missing values on the independent variables (which ranged between 0.3% and 3.3% for single variables but would have reduced the overall data with listwise deletion by 13.2%) were imputed with multiple imputation in Stata, by chained equations (m = 20) and including all variables (White, Royston, & Wood, 2011). The results obtained via listwise deletion were largely similar to the results with imputed data (results available upon request). Excluding those cases that had missing values on the dependent variable (7.9%), we reached a sample of 11,305 first-generation Peruvian migrants.

| Well-being
The dependent variable is operationalised by the question 'In the country where you are living, what are your principal concerns?' Respondents could choose multiple answers, including 'employment' , 'migration status' , 'language' , 'discrimination' and 'depression/loneliness' . Around half of the respondents who filled in 'depression/loneliness' did not choose any other answer option, one third filled in one and one fifth two or more additional principal concerns. As many respondents did choose multiple answers, it can be assumed that depression/loneliness has not been obscured by other possible principal concerns.
Loneliness refers to the unpleasant experience of sensing a discrepancy between either the number or quality of the social relationships a person has and desires (Perlman & Peplau, 1981). Depression has been defined not just as an unpleasant state but also as a psychological disorder with clinical implications (Davey, 2014). The two concepts are usually treated as separate theoretical constructs and used as independent measures for well-being (Koropeckyj-Cox, 1998). That both formed a single answer category in the WSPCA hearkens to the mutual relationship between the two constructs. It is widely acknowledged that depression and loneliness often go hand-in-hand (Rokach & Philibert-Lignières, 2015), with correlations ranging from 0.4 to 0.6 (Russell, Peplau, & Cutrona, 1980;Weeks, Michela, Peplau, & Bragg, 1981). Moreover, the link between the two concepts has been shown to be present at all ages and to be relatively stable over time (Victor & Yang, 2012). In terms of their causal relationship, research considers loneliness to be a precursor as well as a consequence of depression (Barg et al., 2006;Cacioppo & Cacioppo, 2014;Green et al., 1992). As the analysis will look at the risk and protective factors for poor wellbeing, the dependent variable was coded 1 if the respondent reported 'depression/loneliness' as a principal concern, and 0 otherwise.

| Transnational ties
To measure transnational ties, a distinction is made between variables referring to transnational behaviour and to belonging (Klok et al., 2017).
Transnational behaviour is operationalised by two variables: by the frequency of family visits with the categories 'regularly' (at least every 2 years), 'occasionally' and 'never' (=reference category) and by the frequency of financial remittances sent, assessed through the question 'How often did you send money to Peru in the last twelve months?' with six response options. We distinguished three categories: 'regularly' (at least every month), 'occasionally' (at least once a year) and 'never' (=reference category). Transnational belonging is operationalised by four variables: first is by the absence of relatives in the host country (no = 0, yes = 1) as a proxy for the presence of close relatives in Peru.
Second is by the respondents' return intention with the categories 'permanent return intention' , 'for periods of months' and 'no return intention' (=reference category). Third is by the variable 'information about Peru on the internet' , indicating whether respondents visited four specific websites (Peruvian Consulate, Ministry of External Affairs, El Comercio newspaper and RPP news) in the last 3 months (=1) or not (=0). Fourth is by the variable 'purchase of Peruvian articles' , assessed by the question 'In the country where you live, did you buy any of these Peruvian products?' with five multiple answer categories (e.g., clothes, food and beverages). Respondents who marked 'yes' in a category were coded with 1, resulting in a score of 0-5.

| Control variables
A first set of control variables includes the demographic characteristics of age (continuous), gender (male = 0, female = 1) and marital status, where married/cohabiting respondents (reference category) were distinguished from those single, divorced and widowed. A second set includes the migration-related characteristics of duration of residence (in years) and migration status (irregular = 0, regular = 1). The category 'regular migration' was constructed clustering different types of legal residence such as citizenship, permanent residence or working visa.
Respondents reporting that their visa has expired fell under the category 'irregular migration status'. A third set comprises four socio-economic variables: education ranging from no schooling (=1) to postgraduate studies (=6), having a paid job (no = 0, yes = 1), homeownership in the host country (no = 0, yes = 1) and proficiency speaking the host country's language with the categories 'well/very well', 'moderate' and 'none' (reference category). Third, 'region of destination' was controlled for, distinguishing between Latin America (reference category), North America, Europe and elsewhere. Finally, we included whether the data were collected in a face-to-face interview (=1) or via a self-completion questionnaire (=0), as the latter method of data collection generally yields higher levels of loneliness (De Jong Gierveld & Van Tilburg, 1999). Multicollinearity was tested, with results showing that none of the independent variables was highly correlated (correlation matrix available upon request).

| Descriptive statistics
Descriptive statistics are displayed in Table 1, which lists the means and ranges of both categorical and continuous variables. Overall, 8% of respondents reported poor well-being, with feelings of depression and/or loneliness as their principal or one of their principal concerns.
The relatively low percentage indicates that the dependent variable rather measures severe than mild or moderate depression/loneliness. On average, respondents were 41 years old, and women were slightly more represented than men were. Most respondents were married/cohabiting (52%) or single (40%), whereas comparatively few were divorced (6%) or widowed (2%). Duration of residence ranged from 1 to 60 years with a mean of 10 years. A total of 88% had a regular migration status. The average educational level was around 4, which corresponds to nonuniversity higher education. One fifth (21%) of the respondents owned property in the host country, and three quarters (73%) had a paying job. Most described their host country language proficiency as well/very well (72%) or moderate (25%), compared with 3% who reported that they did not speak the new language. Respondents were almost equally spread over the three continental regions of Latin America, North America and Europe, and 70% of them were interviewed face-to-face.
According to the indicators for Peruvian migrants' transnational behaviour, 40% of the respondents visited family in Peru at least every 2 years, 27% occasionally and 32% never. Regular financial remittances were sent by 36% of respondents, compared with 24% who remitted occasionally and 40% who had not transferred money to Peru in the year prior to the survey. With respect to respondents' transnational belonging, as a proxy for the presence of close family in Peru, 23% had no relatives living in the host country. Around 40% preferred to either return permanently to Peru or to alternate periods of months in the two countries, whereas 18% preferred to stay in host country. Most had visited a Peruvian website during the last 3 months (72%), and on average, respondents had purchased two Peruvian articles since their arrival.

| Regression analysis
Do transnational ties protect or threaten the well-being of Peruvian migrants? To answer this question and test the corresponding transnational resource and transnational stress hypotheses, we used logistic regression with poor well-being (depression/loneliness as principal concern) as outcome variable. Table 2   country, in case of a permanent return intention, and the more vPeruvian articles were bought, respectively. Similar results were found when looking at the transnational ties variables separately, with one exception: occasionally sending remittances is then significant at a 95% confidence level (results not presented, available upon request).
Additional analyses further show that the associations of transnational ties variables with poor well-being were very similar for both men and women (results not presented, available upon request).
As for the control variables, the odds of poor well-being increased linearly with age, so that older respondents were more likely to report depression/loneliness as principal concern. Women and widowed respondents were more likely to have poor well-being than men and married/cohabiting respondents, respectively. With respect to migration-related characteristics, the longer respondents lived in the destination country, the less likely they were to experience poor wellbeing. Migration status also played a role, but not in the direction one would expect: lacking a regular migration status was associated with a decreased likelihood of poor well-being. As additional analysis shows, this particularly holds for women. Of the socio-economic characteristics, only being employed is associated with lower odds of poor wellbeing. Compared with respondents who lived in Latin America, those in continents other than North America and Europe were more likely to report depression/loneliness as principal concern. Finally, the odds of poor well-being were lower when being interviewed face-to-face. This is not surprising, as respondents are less likely to report their problems, worries and so forth when asked directly by the interviewer.

| CONCLUSION AND DISCUSSION
This study aimed to explore the effects of transnational ties on the well-being of Peruvian migrants residing around the globe. Two conflicting hypotheses were formulated-transnational resources versus transnational stress-according to which transnational ties either reduce or increase well-being among migrant populations. Wellbeing was measured by a binary indicator for whether or not depression/loneliness is a principal concern. A wide array of indicators of transnational ties was included, capturing both transnational practices (family visits to Peru and sending remittances) and transnational belonging (absence of relatives in host country as a proxy for the pres-   (Lulle, 2014). However, the different explanations could not be tested due to lack of available data about the reasons behind intense transnational engagement. The mechanisms underlying the positive association between frequent transnational practices and poor well-being of migrants therefore remain an open question.
The observed association between migrants' poor well-being on the one hand and permanent return intention and purchase of various Peruvian articles on the other provides more evidence to support the argument that maintaining a strong ethnic identity threatens rather than fosters migrants' well-being. Studying Turkish return migrants, Kunuroglu, Yagmur, Van de Vijver, and Kroon (2017) revealed that a strong sense of belonging to their earlier communities was important to their decision making. Negative experiences, especially perceived discrimination, were a major issue why these return migrants did not develop a feeling of belonging to the host country or the host society.
In this sense, permanent return intentions can be seen as part of 'reactive transnationalism' (Itzigsohn & Giorguli-Saucedo, 2002) that explains more intense transnational practices and belonging as a migrant's reaction to feeling rejected by the majority society. For the same reason, migrants may adapt or maintain their consumption behaviour in order to express their exclusive belonging to a wider ethnic group as well as to the country of origin. Food consumption, for example, has been shown to play an important role in the maintenance and formation of migrants' local and transnational identities (Chapman & Beagen, 2013;Weller & Turkon, 2015). Still, while the consumption of ethnic products may help migrants to develop a positive sense of identity, they may also be evoking memories of past people and places (Abarca & Colby, 2016). The latter assumption is supported by our study, although more in-depth qualitative research is needed for a better understanding of why migrants purchase a variety of ethnic products. Another finding that calls for future research is that particularly male migrants and those without a regular migration status show a lower likelihood of poor well-being. One might have expected that lacking residence rights would cause higher levels of distress due to fear of deportation and limited access to formal employment, housing and social services, and so forth (Aroian, 1993;Cavazos-Rehg, Zayas, & Spitznagel, 2007). In addition, irregular migrants do not dispose of mobility rights and cannot-at least not successfully-apply for family reunification. This severely diminishes their ability to visit family and postpones a reunion with close family into an uncertain future. It might however be precisely the prospect of a longer period of spatial separation from close family members, which has a bearing on migrants' coping strategies. Hence, in order to cope with absence and uncertainty, they might seek to reduce the emotional consequences of a situation that is largely beyond their control. Another explanation could be that migrants without a regular migration status possess or develop resilience differently than migrants in a regular situation.
Being 'used' to multiple types of stress related to their migration status, they are more able to face the challenges and personal implica- To develop adequate measures, more knowledge is needed about the causal relationship between migrants' transnational ties and wellbeing. Indeed, one of the limitations of this study is that we cannot draw conclusions about causality because of the type of data analysed (cross-sectional). For example, it cannot be said whether it is through a strong sense of belonging that permanent return plans are positively associated with poor well-being or that poor well-being is the reason why migrants plan to return permanently and therefore invest more in transnational social ties. In other words, the question remains open as to whether more intense transnational ties cause poor migrant well-being or whether poor migrant well-being leads to more intense transnational ties.
In addition, the dependent variable may not be seen as the ideal method of measurement of depression and loneliness. However, the aim of this study was not to provide reliable and valid estimates of the prevalence of either construct but to examine the association between transnational ties and migrants' well-being approximated by a binary mental health indicator. Still, the indicator undeniably comes short in grasping the multifaceted nature of well-being. Future research is needed to explore the relationship between transnational ties and the various dimensions of well-being (psychological, social and physical).
Concerns about the representativeness of the sample have to be considered as well. For example, data collection in Peruvian consulates has not been random, as only those individuals were interviewed who, personally or as a proxy, had to deal with consular issues. Finally, important information about migrants' type of transnational family relations are missing. It may be assumed that specific dyadic relationships (e.g., mother-child) have a particular effect on the association between the intensity of migrants' transnational ties and well-being, in conjunction with the perceived quality of these relationships.