Introduction

Workforce shortages drive nations and firms to vie for the same talent pool (Kerr, 2020). Skilled migrantsFootnote 1 comprise an essential part of the global labor market (Barnard et al., 2019) and optimizing their employment is important for multinational enterprises (MNEs) and international business (IB) (Fitzsimmons et al., 2021). Indeed, skilled worker migration has significant consequences for nations and firms. For example, in the US, a migrant arriving at age 25 with a college degree pays about half a million dollars more in taxes than they consume in government services over their life; the fiscal surplus nears $1 million for migrants with advanced degrees (National Academy of Sciences, Engineering and Medicine, 2016). In Europe, new immigrants represented almost 20% of entries into growing occupations, notably healthcare and STEM (Sohst et al., 2020), and talent flows from Asia are surging in the OECD, with 28% of immigrant inventors born in either India or China (Kerr, 2020). For firms, skilled migrants increase workforce diversity, enhance knowledge transfer and innovation, and stimulate FDI (Choudhury, 2022; Foley & Kerr, 2013).

Despite these benefits, highly skilled migrants commonly experience unsatisfactory outcomes in their destination countries (Guo et al., 2019; Harrison et al., 2019). For instance, migrants tend to be under-employed and under-paid relative to their native-born counterparts, are often first to lose their jobs in economic downturns, and face persistent insecurity about residence permits (Hajro et al., 2021). Rising immigration barriers, increasing anti-globalization, and xenophobia pose significant challenges to them even if their destination countries depend on foreign talent (Szkudlarek et al., 2021).

At the same time, competition for talent means that highly skilled workers have mobility options with emerging economies enticing returnee migration (Tung & Lazarova, 2006) and nations and cities actively attracting skilled migrants (Zikic & Voloshyna, 2023). Even in the face of mobility “frictions” (Choudhury, 2022), highly skilled workers that experience unsatisfactory outcomes in one location may leave for better options elsewhere. Attracting and retaining talent is thus a central issue for IB, and further scholarly inquiry is needed to identify factors that help highly skilled migrant employees settle successfully and remain in the destination country (Turkina & Van Assche, 2018).

Recognizing that macro societal factors likely play an important but understudied role in outcomes for highly skilled migrants (Hajro et al., 2019), we propose that destination-country institutional characteristics vis-à-vis those in the migrant’s origin country, along with migrants’ subjective experience of destination-country institutions, provide substantial and possibly contrasting influences on how successfully migrants settle and remain in the destination country. For example, a destination country may offer superior educational and economic systems but may also expose migrants to xenophobic harassment and discrimination. Thus, we propose that migrants’ experiences with the destination country’s formal and informal institutions as well as dissimilarity between the destination and origin-country institutional characteristics can provide opportunities and pose challenges that differentially affect migrants’ respective identification with their destination and origin countries. In turn, we detail how identification with these two target countries affects migrants’ destination-country embeddedness.

We focus on embeddedness in the destination country as the key outcome variable. Originally developed in organizational turnover research, the embeddedness construct reflects an array of forces that enmesh individuals within their organizations and/or communities (Mitchell et al., 2001). In the expatriation literature, destination-country embeddedness has been associated with lower employee turnover and intention to repatriate, and higher cross-cultural adjustment (e.g., Reiche et al., 2011; Tharenou & Caulfield, 2010). If migrants are more embedded in the destination country’s social and work contexts, they are more likely to stay and contribute to their organization and society.

Migration crosses geographic boundaries and levels of analysis (Buckley et al., 2017), but multilevel multi-target-country studies are uncommon. Specifically, there is limited work on how institutional characteristics shape individuals’ post-migration experience (Hajro et al., 2019). Addressing this concern and building much-needed macro-micro bridges, we demonstrate how institutional characteristics of migrants’ destination country and origin country, and their subjective experiences of the destination-country institutions, contribute to skilled migrants’ embeddedness.

Conceptual background and hypotheses

Given the importance of skilled migrant retention for destination countries and employing organizations, we identified embeddedness as our outcome of interest (Grzymala-Kazlowska, & Ryan, 2022). Embeddedness is a three-dimensional (fit, links, sacrifice) construct. Applied to migrants, the fit dimension refers to the match between the migrant’s values and goals with the opportunities provided in the destination country; the links dimension reflects informal and formal ties between the migrant and members of the destination country; and the sacrifice dimension captures a migrant’s perception of what they would lose if leaving the destination country.

In this paper, we focus on the sacrifice dimension of embeddedness for two reasons. First, compared to links and fit, sacrifice is better suited to understanding an important component of skilled migrants’ experiences as it captures migrants’ perceptions of what they can broadly lose by leaving their destination country (Cerdin et al., 2014). This should be particularly relevant for highly skilled migrants who, compared to low-skilled migrants, have more flexibility to decide where to move, therefore perceiving they would sacrifice less if they left a particular destination country (Ailon-Souday & Kunda, 2003; Lee et al., 2018). Second, destination-country fit and links may themselves be sacrificed upon leaving, making sacrifices critical for retention decisions (Reiche et al., 2011). Further, Halvorsen et al. (2015) assert that “sacrifice may be the factor that truly embeds [migrants] to an organization and the community” (p. 26). In contrast to expatriates, who are temporary destination-country residents and tend to maintain origin-country affiliations, migrants relocate for an indeterminate period and need to invest relatively more in their social relationships and professional career in the destination country, which means sacrificing them should be impactful. Such considerations make understanding what might raise or lower the sacrifice aspect of embeddedness especially important.

As individuals can be embedded in both their work context and non-work context (Mitchell et al., 2001; Tharenou & Caulfield, 2010), and given that embeddedness in both domains can shape decisions to stay in a job or a country (Singh et al., 2018), we differentiate between migrants’ community embeddedness, defined as the perceived value of community-related benefits that are forfeited by leaving the destination country, and their career embeddedness, capturing the perceived value of career-related benefits that are forfeited by leaving the destination country.

An inherent question, then, is what promotes or hinders skilled migrants’ community and career embeddedness? Acculturation theory provides a basis for understanding the mechanisms involved. Acculturation is a change in individuals’ cultural repertoire, value system, and self-meanings due to balancing two cultures while adapting to the destination country (Berry, 1997; Jacob, 2020). Although most acculturation studies focus on behavioral changes (e.g., Adler & Aycan, 2018), Schwartz et al. (2010) emphasize that acculturation is “a multidimensional process consisting of the confluence among heritage-cultural and receiving-cultural practices, values, and identifications” (p. 237), with changes in the person’s identity at its core. We draw three conclusions from Schwartz’s expanded view of acculturation in developing our theory about societal factors and skilled migrant destination-country embeddedness. First, identity-related processes are central to migrants’ embeddedness. Second, we need to consider identification in both destination country and origin country to understand migrants’ beliefs about the destination country. Third, like culture, societal institutions vary between the origin and destination countries in ways likely to highlight different identities. This led us to consider country identification as our key mechanism linking societal level factors with individual embeddedness.

Country identification and embeddedness

Country identification has its roots in Social Identity Theory (SIT) and its extension, Self-Categorization Theory (Tajfel, 1982; Vora & Kostova, 2007). SIT asserts that people perceive the social environment in terms of in-groups and out-groups based on their perceived membership in collectives. Individuals engage in “identity work” (Ibarra & Barbulescu, 2010) when in-group and out-group differences become salient, or their identity is challenged. In the migration context, this involves defending, re-evaluating, and revising one’s identity in light of the destination-country environment (Zikic & Richardson, 2016). Migrants balance identity preservation and identity dynamism (Petriglieri, 2011), but identifying with the destination country and origin country are largely independent processes (Berry, 1997). As such, identifying with the destination country does not mean decreasing identification with the origin country; identifying with both origin country and destination country is possible (Berry, 1997). Therefore, because destination and origin countries represent distinct social groups, we expect they have distinct effects on destination-country embeddedness.

Strong destination-country identification means that a migrant sees themself as destination-country in-group member, raising favoritism toward the destination country, and elevating its importance, meaning and significance for a migrant’s self-esteem and self-concept (Slotter et al., 2015; Stets & Burke, 2000; Tajfel, 1978). To maintain this positive distinctiveness, migrants often engage in identity-confirming cognitions (e.g., positive assessment) and socioemotional behaviors (e.g., investing in the group) (Burke & Reitzes, 1991). We expect that these identity-confirming cognitions and actions involve both career and community domains. Because skilled migrants typically move for career opportunities, strong destination-country identification includes alignment with its workplace norms and values, and a willingness to work hard to succeed (Cerdin et al., 2014; Gecas, 1982). Destination-country identification also involves perceiving locals as in-group members, increasing migrants’ willingness to take part in community and social activities (Cerdin et al., 2014). Conversely, people with weaker destination-country identification have lower cognitive alignment with and less drive to invest in the destination country (Schmitt et al., 2000).

As illustrated, destination-country identification likely encompasses elements of both career and community, so we expect that it relates to embeddedness in both domains. The connection between identification and embeddedness results from the loss frame migrants hold about leaving the destination country (Cerdin et al., 2014). Loss framing refers to the extent to which migrants perceive that the risks and costs of leaving are or are not worth taking. Identity research demonstrates that when facing the loss of a valued group membership (even if imagined), individuals anticipate self-esteem loss and self-concept uncertainty, and these increase as identification strengthens (Slotter et al., 2010). The identity-reinforcing cognitions and behaviors associated with strong identification elevate the losses anticipated from losing the destination-country identity. As a result, migrants with strong destination-country identification will perceive more sacrifices associated with leaving. Conversely, if destination-country identification is weak, a migrant’s self-concept is unfazed by the thought of leaving (Slotter et al., 2015), and their sense of sacrifice associated with departing will be lower.

Hypothesis 1a

Destination-country identification is positively associated with community embeddedness in the destination country.

Hypothesis 1b

Destination-country identification is positively associated with career embeddedness in the destination country.

Independent of destination-country identification, origin-country identification can play a role in the sense of loss expected from leaving the destination country. High origin-country identification means that a migrant retains a sense of attachment to the origin country in-group and derives self-esteem from their origin-country identity. High origin-country identification may prompt migrants to spend time and effort nurturing professional contacts with and maintaining community involvement in their origin country. Lending some support to this, expatriates experiencing a strong pull from one’s origin country due to high origin-country identification have higher intention to repatriate and increased job search in their origin country (Tharenou & Caulfield, 2010). Involvement in such origin-country activities across their personal and work domains likely leaves them with relatively less time and energy resources to invest in the destination country (Halbesleben et al., 2014), which should reduce their embeddedness in the destination country. Conversely, identity-confirming cognitions and socioemotional behaviors associated with the origin country elevate their sense that career- and community-related benefits exist outside the destination country. Together this should reduce migrants’ perceived sacrifices in leaving the destination country.

Hypothesis 2a

Origin-country identification is negatively associated with community embeddedness in the destination country.

Hypothesis 2b

Origin-country identification is negatively associated with career embeddedness in the destination country.

Institutional characteristics and embeddedness

Institutional factors have received little attention in research on highly skilled migrants (Hajro et al., 2019), and yet informal institutions such as socially shared norms and values, political ideology, and informal social systems, and formal institutions such as economic factors, human development opportunities, and labor markets (Dau et al., 2022) likely matter for migrants’ embeddedness. As illustrated in Fig. 1, we examine how two destination-country institutional characteristics affect destination-country and origin-country identification and, in turn, embeddedness: opportunities for human development and societal cultural practices. Migrants’ experiences with societal contexts can further impact their identification and embeddedness. For example, migrants might experience job insecurity and discrimination in the destination country, despite the greater economic stability on offer (Bretones, 2020; Zikic & Richardson, 2016). We therefore consider two common experiences linked to characteristics of the destination country: experienced ethnocentrism and experienced downgrading (Hajro et al., 2019).

Fig. 1
figure 1

Conceptual model and hypotheses. Note. The hypotheses in italics predict career embeddedness, and the hypotheses not in italics predict community embeddedness. The independent variables with single border lines are measured at the individual level, the independent variables with double borders are measured at the country-dyad level

Full size image

We propose that the formal and informal institutions in the destination country relative to the origin country, as well as migrants’ experiences with formal and informal destination-country institutions can serve as opportunities or constraints, promoting or hindering migrants’ embeddedness. Migrants moving from a less developed to a more developed country may enjoy greater economic stability and civil liberties, but they often also face a “liability of origin,” which may subject them to xenophobia and exclusion (Fang et al., 2013; Guo et al., 2019). Further, the extent to which an origin country’s poor economic development becomes a liability for a migrant depends on both the institutional distance between the origin and destination country and the direction of the move (van Hoorn & Maseland, 2016; Vora & Kostova, 2007), which is why we consider both effects.

Formal institutional characteristics and embeddedness

Human development opportunities in the destination country represent an important formal institutional characteristic for embeddedness because skilled migrants commonly relocate to find better jobs, improved standards of living, and better education (Harrison et al., 2019; Kerr, 2020). We define opportunities for human development as the extent to which a destination country provides greater opportunities for human development and standard of living relative to a migrant’s origin country. For example, Denmark scores nearly twice as high on the human development index (HDI) compared to Nigeria, so the experience of someone moving from Nigeria to Denmark would be different from that of a Dane moving to Nigeria. Greater human development opportunities act as a “pull motive” (Richmond, 1993) wherein migrants see value in the destination country (Cerdin et al., 2014); however, the direction of the difference matters – it is only when institutional differences directionally favor the destination country over the origin country that migrants perceive value associated with the destination country. Indeed, international migration typically flows from less developed towards more developed economies (United Nations, 2017).

When a destination country offers favorable human development opportunities compared with the origin country, migrants should perceive membership in the destination country as valuable. This should elevate the destination country’s significance for a migrant’s self-esteem and self-concept (Slotter et al., 2015; Stets & Burke, 2000; Tajfel, 1978). As a result, migrants will increase their destination-country identification. In turn, migrants will more likely display identity-confirming cognitions such as positively assessing the destination country and socioemotional behaviors like investing in the destination society, increasing their embeddedness. We expect that these effects are predominantly driven by migrants’ destination-country identification and less so by their origin-country identification. This is because without some level of destination-country identification, migrants are unlikely to perceive substantive sacrifices from leaving the destination country (Cerdin et al., 2014), no matter how much they identify with the origin country. We further expect that destination-country identification is associated with both career and community embeddedness, and this is particularly true for the connection between opportunities for human development and both domains because human development opportunities cover the economy and labor markets as well as community characteristics such as education opportunities.

Hypothesis 3a

Opportunities for human development is positively associated with destination-country community embeddedness, mediated through its positive effect on destination-country identification.

Hypothesis 3b

Opportunities for human development is positively associated with destination-country career embeddedness, mediated through its positive effect on destination-country identification.

Experienced downgrading. Even in countries with high human development opportunities, migrants commonly experience occupational downgrading (Hajro et al., 2019). Downgrading refers to a loss of occupational status between one’s job in the origin country and the job in the destination country (Crollard et al., 2012). Since this is specific to the work domain, we expect that the effects of downgrading will be concentrated on career embeddedness and that the negative effects will be mediated by both destination-country and origin-country identification.

When individuals feel their identity is threatened, they may engage in identity-protection responses such as derogation of the source of threat or creating a positive sense of self-worth by displaying pride in one’s origin-country identity (Petriglieri, 2011). In the case of migrants’ experience of downgrading, the destination country is the source of occupational identity threat, so migrants will be motivated to discredit the integrity or legitimacy of the source of threat, thus alleviating the potential harm. In their study of highly skilled migrants from Sri Lanka in the UK, Fernando and Patriotta (2020) found that sensemaking narratives often followed a similar pattern that represented the occupational world in terms of opposition between ‘us’ and ‘them’: migrants responded to occupational downgrading by adopting an antagonistic stance towards their British coworkers, positioning themselves in more positive terms as the employing organization was derogated. The negative stance toward the destination country as an out-group will lower identification with the destination country.

Professional or occupational identity is a source of achieved status – connected to a person’s abilities, merit, and accomplishments. Seeking positive distinctiveness – another identity-protection response (Petriglieri, 2011) – involves confirming the threatened identity, e.g., by expressing pride in one’s origin-country identity. Dissonance created by downgrading generates uncertainty and anxiety hindering a migrant’s ability to develop alternative identities (Mackenzie Davey & Jones, 2020). To enhance self-worth and status, migrants experiencing downgrading will attempt to reconnect cognitively with their past self. Desire for dissonance reduction motivates “identity work” (Zikic & Richardson, 2016), such as elevating their professional status and career achievements “back home” to gain cognitive continuity with their old sense of self. One of the migrants interviewed by Zikic and Richardson (2016:159) in their study of the professional identity dynamics experienced by migrants described herself as “a fish out of the water,” and highlighted her professional identity as a medical doctor in her origin country: “I was a physician back home. I can serve people and I can help people. Here I am nothing.” The positive association with the origin country reinforces it as the in-group, which, along with derogation of the destination country, promotes migrants’ origin-country identification.

Hypothesis 4a

Experienced downgrading is negatively associated with destination-country career embeddedness, mediated through its negative effect on destination-country identification.

Hypothesis 4b

Experienced downgrading is negatively associated with destination-country career embeddedness, mediated through its positive effect on origin-country identification.

Informal institutional characteristics and embeddedness

Value-congruent societal practices. Cultural differences between a migrant’s destination country and origin country are seen as a source of adjustment problems and acculturative stress (Berry, 1997; Schwartz et al., 2010), but the global mobility literature offers little support for the notion that cultural distance per se has an adverse effect on individual outcomes (e.g., Bhaskar-Shrinivas et al., 2005; Harrison et al., 2019). Shenkar (2001) has pointed to several conceptual flaws in existing cultural distance research, two of which are particularly relevant in the migration context. First, research on cultural distance suffers from the “illusion of symmetry,” i.e., the problematic assumption that the cultural distance from country A to country B is identical to the distance from country B to country A. Empirical research (e.g., Selmer et al., 2007; Yildiz & Fey, 2016) and common sense suggest that this assumption is erroneous: The cultural adjustment challenges facing a Nigerian migrant in Denmark are likely to be quite different to the ones facing a Danish migrant in Nigeria, although the cultural distance between the two countries (e.g., as conceptualized by Kogut & Singh’s [1988] index) is the same.

Second, cultural distance research suffers from the “illusion of discordance” (i.e., that cultural differences are inevitably a source of difficulties and conflicts). Although cultural distance can make cross-cultural interaction and adjustment more difficult, ample evidence suggests that cultural differences can also elicit positive outcomes (Stahl & Tung, 2015). In the foregoing example, a move from Nigeria – a culture characterized by high power distance practices (House et al., 2004) – to Denmark – a culture with low power distance practices – offers migrants the promise to become members of a society where hierarchy is de-emphasized, equality is promoted, and participation in decision processes is encouraged – practices that will be desirable by most migrants from Nigeria.

We propose that in the migration context, both societal practices and societal values need to be considered in predicting individual outcomes. The GLOBE project (House et al., 2004) makes a critical distinction between cultural practices (“culture as is”) and cultural values (“culture as should be”). Cultural practices tell us something about the way things are done in a society, whereas cultural values reflect societal members’ desires, preferences, and aspirations regarding the way things should be done. Importantly, cultural practices and values are not always closely aligned; in the GLOBE project seven of the nine national culture dimensions showed significant negative relationships between the values and practices scores (Maseland & Van Hoorn, 2010). This presents the possibility that an origin-country values–practices mismatch, from the point of view of a migrant, might be reduced (or heightened) by moving to another country. In the foregoing example, Nigeria has high scores for power distance practices but relatively low scores for power distance values, indicating that Nigerian society has a strong emphasis on hierarchy in terms of actual practices but a desire for less hierarchical practices. Hence, from the perspective of Nigerian nationals, migrating to Denmark would result in greater societal values–practices match, whereas in the less common case of a move from Denmark to Nigeria, migration would result in a societal values–practices mismatch on cultural dimensions such as power distance.Footnote 2 This underscores the asymmetrical nature of migratory moves between countries and the need to consider directionality in migrant research.

We expect that these dynamics have implications for migrants’ identification with the destination and origin countries and, ultimately, their embeddedness. A societal values–practices match should raise migrants’ perceived value of membership in the destination country because the destination culture is more aligned with what migrants’ origin culture aspires to be. Therefore, the migrants will view the destination country as more important for their self-esteem and self-concept, increasing their destination-country identification (Stets & Burke, 2000; Tajfel, 1978). At the same time, a societal values–practices match in the destination country will draw attention to misaligned values and practices in the origin culture, which should lower their origin-country identification.

We expect the mediating effects of destination-country and origin-country identification to be domain-specific, i.e., value-congruent societal practices will mainly affect community embeddedness, and less so career embeddedness. Migrants who feel that the move to the destination country resulted in greater societal values–practices match and who therefore feel attached to the destination country will more likely invest in the social fabric of their communities by building local friendships and contributing to social activities, increasing their community embeddedness. Conversely, migrants who experience a societal values–practices mismatch will perceive benefits from maintaining socioemotional attachment with the origin country. This may prompt them to focus their social efforts on members of their origin community in the destination country, remain more separate, and reduce their efforts to embed in the destination community (Berry, 1997).

Hypothesis 5a

Value-congruent societal practices is positively associated with destination-country community embeddedness, mediated through its positive effect on destination-country identification.

Hypothesis 5b

Value-congruent societal practices is positively associated with destination-country community embeddedness, mediated through its negative effect on origin-country identification.

Experienced ethnocentrism. Societal members’ attitudes toward migrants vary on a continuum from multiculturalism (i.e., accepting or even welcoming of cultural pluralism) to ethnocentrism (the belief in the inherent superiority of one’s own ethnic group or culture) (Berry & Kalin, 1995). Ethnocentrism is aimed at national, ethnic, and immigrant identities, which are not earned and form an essential part of an individual’s attachment to corresponding social groups. Experiencing ethnocentrism creates a persistent sense of being devalued for a quality that an individual cannot change and induces a need to hide or suppress their identity. Ethnocentrism also signals that the individual does not belong to the in-group. Because these threats arise in the destination country, they will likely be interpreted as beliefs and prejudices held by society as a whole, making the destination-country citizens a salient out-group. Migrants experiencing such ethnocentric attitudes often respond by psychologically distancing themselves from the natives (Wehrle et al., 2018), thereby inhibiting their destination-country identification. Furthermore, ethnocentrism makes natives less willing to interact with migrants, even if a migrant is motivated to engage with destination-country nationals. When this occurs, migrants may respond to the perceived identity threat with derogation (Petriglieri, 2011), which will lower their identification with the destination country and, ultimately, decrease their embeddedness in both the personal and the workplace domains.

Hypothesis 6a

Experienced ethnocentrism is negatively associated with destination-country community embeddedness, mediated through its negative effect on destination-country identification.

Hypothesis 6b

Experienced ethnocentrism is negatively associated with destination-country career embeddedness, mediated through its negative effect on destination-country identification.

Experienced ethnocentrism should also make it more likely that migrants strengthen their origin-country identification. Allport’s (1954/1979) seminal study on the nature of prejudice suggests that perceived discrimination against an ethnic group can increase group members’ identification with the in-group and reduce the willingness to identify with the out-group. A study by Spencer-Rodgers and Collins (2006) found that perceived group disadvantage – a concept closely related to perceived discrimination – among Latino-Americans was negatively associated with identification with the out-group, mediated by perceived public regard (the belief that destination-country nationals look down on Latinos). Perceived group disadvantage was also positively associated with personal self-esteem through the mediator of increased identification with one’s own ethnic group, as manifested in increased group centrality, attachment to the group, and liking for the group. Identity-protection responses such as creating a positive sense of self-worth by displaying pride in one’s national identity (Petriglieri, 2011; Zikic & Richardson, 2016) may further reinforce migrants’ origin-country identification and lead them to distance themselves from the destination country. Like for destination-country identification, we propose that the mediating effect of origin-country identification can be observed across domains, i.e., experienced ethnocentrism will affect both community embeddedness and career embeddedness.

Hypothesis 6c

Experienced ethnocentrism is negatively associated with destination-country community embeddedness, mediated through its positive effect on origin-country identification.

Hypothesis 6d

Experienced ethnocentrism is negatively associated with destination-country career embeddedness, mediated through its positive effect on origin-country identification.

Methods

Sample and data collection

We collected data in 2019 and the beginning of 2020 (pre-COVID) from 12 countries: Australia, Austria, Canada, Finland, France, Germany, Russia, Spain, Sweden, Switzerland, the UK, and the USA. In some countries, we used convenience and snowball sampling, approaching migrants directly through channels such as alumni organizations, migrant workers’ associations, researcher communities, nursing communities, and recruiting agencies. In other countries, we used services such as Amazon Mechanical Turk, Qualtrics, Bilendi and Purespectrum. Most data were collected via online surveys, but we also used other data collection approaches, such as distributing survey hard copies among members of the migrant community. Hence, the completion rates can be quantified only for some countries, namely 36% in Australia, 38% in Germany, 57% in Russia, 23% in Sweden, and 64% in Switzerland.

Our standardized questionnaire was developed after an extensive review of international migrant research and related fields (e.g., self-initiated expatriates). Where possible, we relied on existing, validated measures. All constructs were measured with multiple-item scales with established psychometric properties. The questionnaire was translated from English to local languages (French, Spanish, German, Russian, Finnish and Swedish) and then back-translated into English (Brislin, 1970).

Our second data source included the UNDP’s (2019) Human Development Index scores and the raw scores from the GLOBE database (House et al., 2004). Both primary and secondary data sources had some missing data. We listwise deleted participants for whom data from secondary data sources were missing (e.g., participants without country-dyad level information on Human Development). Also, we deleted cases where important demographic variables (e.g., age and gender) and all scale items of our substantive variables were missing. We imputed construct-level missing data for our control variables by means of the mice package in R.

We included a broad representation of origin countries, age groups, lengths of tenure in the destination country, and other demographic characteristics over a narrower range. Following Cerdin et al.’s (2014) definition of a skilled/qualified migrant, we only retained respondents who were employed and held a tertiary education or a similar degree, excluding those who were students or had retired. To ensure adequate exposure to the destination country, we included only migrants who spent at least one year therein. Our final sample comprises 1709 highly skilled migrants from 48 origin countries, living in 12 destination countries, and nested in 320 origin-country–destination-country dyads.

Measures

Please see the Online Appendix 1 for all variable definitions and a complete list of our measurement items. To establish the construct validity of our measures that are adaptations of existing scales, we collected additional data from migrants from the UK and the US, which we used solely for measurement validation purposes. Detailed results of our analyses are available upon request.

Community embeddedness and career embeddedness were measured with Tharenou and Caulfield’s (2010) three-item sacrifice scales. The Cronbach’s alpha for community embeddedness was .80, and for career embeddedness was .89.

Destination-country identification and origin-country identification were measured with a subset of three items each from Shokef and Erez’s (2006) identity scales to measure destination-country and origin-country identification. The Cronbach’s alphas were .86 for destination-country identity and .82 for origin-country identification. We tested the shortened identification scales in comparison to the full scales using data provided by Lee et al. (2018) and established concurrent and predictive validity of our measure using our construct validation dataset.

Opportunities for human development were measured with data from the UNDP’s Human Development Index (HDI). Using a geometric mean of life expectancy, expected years of schooling, mean years of schooling and gross national income per-capita values, HDI captures the extent to which a country caters for the needs of human development (UNDP, 2019). Higher levels of HDI indicate higher levels of development and standard of living. We subtracted the HDI value of the origin country from that of the destination country to measure opportunities for human development.

Experienced downgrading was measured with a three-item adaptation of the perceived overqualification scale by Maynard, Joseph and Maynard (2006). We tested our scale’s reliability, concurrent validity (against the full nine-item scale by Maynard et al., 2006) and predictive validity (using organizational commitment, based on Kehoe & Wright, 2013) with our construct validation dataset. Cronbach’s alpha was .71.

Value-congruent societal practices. We created a new index to capture the extent to which the societal practices in the destination country are more closely matched with the dominant values of the origin country than the societal practices in the origin country. We constructed the index using all nine cultural dimensions of the GLOBE study’s (House et al., 2004) cultural practices and cultural values scores. To compute the index, we aggregated the product of two components across each dimension. The first component is the difference in origin-country values and practices, which captures both the level and direction of mismatch along each GLOBE dimension based on the societal value system of the origin country. The second component is the difference between the societal practices scores of the destination country and the origin country, which captures whether an international move implies an upward or downward change along a dimension. To ensure that the differences in practices of all nine GLOBE dimensions have equal weights in our index, for each dimension we divided the difference in practices between two countries by the maximum difference range across all country cultural practices in the GLOBE dataset. This adjustment is denoted as “adj” in our formula below. We multiplied the two components to generate a dimension-specific indicator of societal value–practice congruence. The lower (higher) the value of the product term, the lower (higher) value–practice congruence along the specific GLOBE dimension. Finally, we aggregated these scores across all nine dimensions to generate our value-congruent societal practices score for each origin–destination country dyad. Our formula is as follows:

$$\mathop \sum \limits_{{{\text{i}} = 1}}^{9} ({\text{Origin Values}} - {\text{Origin Practices}}) \times {\text{adj}}\left( {{\text{Destination Practices}} - {\text{Origin Practices}}} \right)$$

where i is an element of the set of all nine dimensions of the GLOBE study. Online Appendix 2 explains the logic behind this construct and the mathematical foundations of this formative index.

Experienced ethnocentrism. We developed a new scale to measure experienced ethnocentrism based on scale development and validation prework on a sample of migrants in Switzerland (reported in Richardson, 2018). We further tested our scale’s reliability, concurrent validity (using a scale by Florkowski & Fogel, 1999) and predictive validity (using Tharenou & Caulfield’s, 2010 intention to return scale) with our construct validation set. Our results indicate a high concurrent validity and good predictive validity of the newly developed scale. Cronbach’s alpha was .84.

Control variables. We controlled for several factors that may be associated with our key theoretical variables. At the individual level, we controlled for: age of migration, since it is more likely for migrants to be entrenched in a destination country if they migrate at a younger age; gender, as gender is a critical determinant of personal identity; highest education level (as a categorical variable: other tertiary degree, bachelor’s degree, master’s degree, or PhD) given that level of education could impact career-related outcomes; tenure in the destination country, as those who have stayed in the country longer have more likely developed a destination-country identity and become embedded in the community and career; cultural intelligence (alpha = .88; Thomas, Liao, Aycan, Cerdin, Pekerti, & Ravlin etc., 2015), as it can influence one’s embeddedness in the destination country by, for instance, facilitating interaction with locals; and international experience (i.e., number of countries previously lived or worked in). Since organizational-level practices could influence migrants’ identification or embeddedness, we controlled for perceived diversity climate in organizations with a five-item measure developed for this study (alpha = .89). We also controlled for whether individuals had refugee status when they moved to the destination country. Finally, we controlled for participants’ survey language proficiency.

At the country-dyad level, we used two control variables. First, since language is an important source of identification (Reiche, Harzing, & Pudelko, 2015), we created a dummy variable named same language reflecting whether the official language of the origin country and the destination country are the same. Second, as colonial ties may influence how migrants perceive their destination country towards the post-colonial power, we created a dummy variable named colonial ties for those origin–destination-country dyads that had colonial links either for more than a hundred years or any time in the last century. We used two country-level controls: the country mean-score of experienced ethnocentrism (accounting for shared levels of ethnocentrism for destination countries), and destination-country HDI levels to distinguish profile effects from distance effects (Van Hoorn & Maseland, 2016). Finally, to account for the possibility that destination-country factors such as systematic discrimination impact migrants’ embeddedness, we included country-level data from the Migration Policy Index (Mipex, https://www.mipex.eu/). Table 1 shows the correlations among all variables (prior to data imputation).

Table 1 Individual level variables
Full size table

Preliminary analyses

We conducted CFA to assess discriminant validity for our latent variables using the full information maximum likelihood method to account for item-level missingness in the dataset. The eight-factor model provided the best fit values (χ2(532) = 2749.03, RMSEA = 0.049, CFI = 0.918, TLI = 0.908) compared with a seven-factor model in which experienced downgrading and experienced ethnocentrism loaded together (χ2(539) = 3847.98, RMSEA = 0.060, CFI = 0.878, TLI = 0.865), and other models with fewer factors. This provides evidence for the distinctiveness of our substantive variables. We then tested for the potential of common method bias by means of a common latent factor method (Podsakoff et al., 2003). Since the model was not identified when we tried to load all 30 items of our substantive variables on a common latent factor on top of items loading onto their corresponding constructs, we tested a model in which we loaded only one item per construct onto a common latent factor. The amount of variance corresponding to this factor was less than 1% of the whole variance, which is substantially lower than the 25% threshold suggested by Williams et al. (1989). Further, none of these items significantly loaded on the common latent factor. Thus, common method bias is unlikely to threaten the validity of our analyses. Finally, the highest generalized variance inflation factor among variables in predicting all four endogenous variables is 1.85, which is considerably lower than ten and indicates that multicollinearity is unlikely to be an issue (Table 2).

Table 2 Dyad variables
Full size table

Since individuals are nested within country-dyads, our hypotheses indicate cross-level direct effects. Two mediating variables, destination-country identification and origin-country identification, carry the effect of our two exogenous variables defined at the country-dyad level (i.e., opportunities for development and value-congruent societal practices) toward our two dependent variables, community and career embeddedness. We expect these four endogenous variables (destination-country identification, origin-country identification, community embeddedness, and career embeddedness) to have individual-level and higher-level components. Therefore, we followed the multilevel modeling approach of Aguinis et al. (2013) to check if the nested structure allows us to test our cross-level direct effect hypotheses.

We structured the dataset such that individuals are nested in origin-country–destination-country dyads in a unidirectional manner. That is, migrating from country A to country B is treated as a separate case from the reverse move. This structuring captures the asymmetrical nature of our opportunities for human development and value-congruent societal practices constructs. We created null models using maximum likelihood estimation in a cross-classified setting with three intercepts (country-dyad, origin country, and destination country), which varied freely with no predictors for each of the four endogenous variables. The total amount of variances associated with the three higher levels vis-à-vis the individual level are as follows: 10.8% for destination-country identification (40.3% of which is associated with the country dyad-level), 8.0% for origin-country identification (28.7% of which is associated with the country dyad-level), 12.0% for community embeddedness (6.7% of which is associated with the country dyad-level), and 22.4% for career embeddedness (14.3% of which is associated with the country dyad-level) (Table 3).

Table 3 Destination-country identification and origin-country identification
Full size table

We then tested whether a model in which intercepts did not vary was nested in a model in which country-dyad intercepts varied freely. ANOVA results support this nested structure for destination-country identification (χ2(1)  = 70.97, p < 0.001), origin-country identification (χ2(1)  = 26.60, p < 0.001), community embeddedness (χ2(1)  = 39.39, p < 0.001), and career embeddedness (χ2(1)  = 193.02, p < 0.001). Finally, we tested a model in which only the country-dyad intercept was nested in a null model in which all three intercepts freely varied. ANOVA results support this nested structure for destination-country identification (χ2(1)  = 31.71, p < 0.001), origin-country identification (χ2(1)  = 18.90, p < 0.001), community embeddedness (χ2(1)  = 54.83, p < 0.001), and career embeddedness (χ2(1)  = 95.31, p < 0.001). We have 51 destination-origin-country dyads, 36 origin countries, and 12 destination countries under which at least nine migrants are nested, which adheres to the requirements suggested by Maas and Hox (2005). In sum, our nested data structure is suitable for random coefficient modeling using random intercepts.

Analytical strategy

We used the lme4 package in R for our analyses (Bates et al., 2015) and tab_model function using Wald method to compute the degrees of freedom and to report the direct effects. While our theorization is at the individual and country-dyad levels and the data structure should take these two levels into account, all individuals in a certain destination country could be affected by their experience in a similar way (e.g., all migrants in Denmark). Similarly, individuals from a certain origin country (e.g., individuals who are emigrating from Nigeria) can have shared effects. To account for this complex nested data structure, we used a cross-classified model. Specifically, we introduced origin-country, destination-country, and origin-destination–country-dyad random intercepts into our analyses. To test our mediation hypotheses (H3a to H6d), we used the Monte Carlo method for assessing mediation (Selig & Preacher, 2008 – http://quantpsy.org/medmc/medmc.htm). This method uses 20,000 random draws from the joint distribution of the path coefficients of the two stages (a and b), and the asymptotic variances and covariances in a mediation model. The resulting distribution of a*b from these draws are used as an input to estimate confidence intervals (CI). For ease of interpretability, we standardized the variables with a high range (age of migration, tenure in destination country, opportunities for human development, origin-country HDI, destination HDI, and Mipex) prior to our analyses.

Results

Supporting Hypotheses 1a and 1b, our results indicate that destination-country identification is significantly positively associated with (H1a) community embeddedness (b = 0.51, p < 0.001, 95% CI 0.454, 0.556) and (H1b) career embeddedness (b = 0.26, p < 0.001, 95% CI 205, 0.321) (Table 4, Models 2c and 2d). Origin-country identification is significantly negatively related with (H2a) community embeddedness (b = − 0.08, p < 0.001, 95% CI − 0.129, − 0.036), but not significantly related with (H2b) career embeddedness (b = − 0.02, p = 0.496, 95% CI − 0.071, 0.034). Therefore, H2a is supported, but H2b is not.

Table 4 Career embeddedness and community embeddedness
Full size table

Providing support for both Hypotheses 3a and 3b, the mediating influence of opportunities for human development via destination-country identification on (H3a) community embeddedness is 0.041 (95% CI 0.016, 0.066), and on (H3b) career embeddedness is 0.021 (95% CI 0.008, 0.035). Our analyses show that the mediating effect of destination-country identification between experienced downgrading and career embeddedness is − 0.013 (95% CI − 0.025, − 0.002), which supports Hypothesis 4a. We did not find support for Hypothesis 4b that origin-country identification mediates between experienced downgrading and career embeddedness, as the mediating effect is − 0.006 (95% CI − 0.003, 0.001).

Hypotheses 5a and 5b state that value-congruent societal practices are positively related to community embeddedness via destination-country and origin-country identification, respectively. Our mediation analyses do not support H5a, for which the mediating effect is 0.020 (95% CI − 0.005, 0.047), whereas they support H5b, as the mediating effect is 0.006 (95% CI 0.001, 0.013). Supporting Hypotheses 6a and 6b, the mediating effect of destination-country identification between experienced ethnocentrism and community embeddedness is − 0.138 (95% CI − 0.165, − 0.112), and between experienced ethnocentrism and career embeddedness is − 0.070 (95% CI − 0.091, − 0.052). Finally, Hypotheses 6c and 6d suggest that origin-country identification negatively mediates between experienced ethnocentrism and community and career embeddedness, respectively. The findings from our mediation analyses support H6c, for which the mediating effect is − 0.006 (95% CI − 0.011, − 0.001) but do not support H6d, as the mediating effect is − 0.001 (95% CI − 0.006, 0.002).

We also ran post-hoc analyses to check if destination and origin-country identifications interact to predict embeddedness. Upon entering the interaction term into our regressions, the term is neither significantly related with community embeddedness (b = − 0.01, p = 0.509, 95% CI − 0.058, 0.029) nor with career embeddedness (b = − 0.00, p = 0.951, 95% CI − 0.051, 0.048). Interaction terms are also not significant when using polynomial regressions (b = − 0.01, p = 0.588, 95% CI − 0.055, 0.031; and b = − 0.00, p = 0.891, 95% CI − 0.053, 0.046), implying that there are no specific identity configurations at which our dependent variables are significantly higher.

Discussion

Because highly skilled migrants often have options for where they choose to live and work, understanding what embeds them in the destination country is vital to fully utilizing their talent and increasing their attachment and retention. Our principal message is that differences between the destination- and origin-country institutional characteristics as well as migrants’ experiences with formal and informal institutions in their destination country generate conflicting dynamics, with some characteristics (e.g., opportunities for human development) leading migrants to identify with their destination country and embed themselves in their communities and careers, and others (e.g., experienced ethnocentrism) compelling migrants to retain their origin-country identification and weakening their embeddedness in the destination country.

We theorized – and established empirically – that different institutional effects elicit different identification processes with significant implications for embeddedness. Destination-country identification is vital if highly skilled migrants are to become embedded in both the community and career domains. We found that pathways for this connection are affected by both macro institutional features and micro experiences of the destination country. Better opportunities for human development relative to the origin country is a key driver of both community and career embeddedness through destination-country identification, but if migrants experience occupational downgrading in the destination country, this lowers their destination-country identification and subsequently career embeddedness. This finding is consistent with research showing that the loss of status when their skills and credentials are devalued or unrecognized poses a major obstacle to highly skilled migrants’ career embeddedness (Fernando & Patriotta, 2020; Zikic & Richardson, 2016). It also suggests that even if a destination country has much to offer in terms of employment and career opportunities, if a skilled migrant does not experience access to those opportunities, their identification with the destination country may be truncated, hindering their career embeddedness.

Another important insight from our study is that if migrants experience ethnocentrism in the destination country, they have both lower community and career embeddedness, mediated through lower destination-country identification. Furthermore, experienced ethnocentrism highlights migrants’ origin-country identity (“I am different, and not one of ‘them’”), and we find that it negatively relates to community embeddedness in the destination country through elevated origin-country identification. Thus, it appears that ethnocentrism is particularly detrimental to migrants’ embeddedness in both domains.

We also find that value-congruent societal practices are important for community embeddedness in the destination country, an effect that was negatively mediated by origin-country identification. In other words, migrants who move to a destination country that is more aligned with their societal values (“culture as should be”) had lower origin-country identification and showed higher levels of community embeddedness in the destination country. A possible explanation is that immigrating to a country whose societal practices are more closely aligned with the values and preferences of the origin country may highlight the societal values–practices mismatch in the origin country.

Interestingly, we did not find support for the hypothesized positive effect of value-congruent societal practices on community embeddedness, mediated by destination-country identification. Together with other results this suggests that destination-country identification and origin-country identification appear to be largely separate processes, influenced by different antecedent factors.

These findings reveal important considerations that social identity processes offer for migration research. Formal institutional features related to human development opportunities may have their greatest influence on destination-country identification because motives to migrate involve improving achieved status and standard of living, raising destination-country identification, so losses may not trigger reevaluation of one’s origin-country identity (Bendersky & Pai, 2018). By contrast, one’s nationality or ethnicity are inherently connected to their origin country, therefore the associated institutional features – both formal and informal – may raise origin-country identification.

Another key insight is that origin-country identification does not seem to disturb career embeddedness in the destination country – even in the face of experienced ethnocentrism or occupational downgrading. We observed that origin-country identification mainly reduces migrants’ community embeddedness. This suggests that migrants who experience ethnocentrism and poor societal values–practices congruence in the destination country end up staying on the margins of society. Therefore, they may perceive sacrifices in leaving the destination country because of the career opportunities they could lose but not for what they have invested in the local community.

Overall, our findings suggest that whereas development of a destination-country identification is a sine qua non for achieving high levels of community and career embeddedness, origin-country identification also matters, but not in the way predicted by acculturation research. A key tenet of Berry’s (1997) model of migrant acculturation is that an integration mode (i.e., high destination-country identification combined with high origin-country identification) is associated with the most favorable outcomes. While this may be true for the integration outcomes commonly studied in acculturation research, such as acculturative stress and mental health (Schwartz et al., 2010; Yoon et al., 2013), “dual identification” (i.e., with both the origin and the destination country) may not be necessary for achieving high levels of embeddedness in the destination country. In fact, origin-country identification may even reduce embeddedness in cases where migrants form closed enclaves that demarcate exclusion.

Limitations and future directions

We worked to ensure a broad representation of origin countries, lengths of tenure in the destination country, and demographic characteristics to enhance the generalizability of findings. There was one characteristic, however, that most migrants in our sample had in common: Nearly 82% of the migrants had moved from an economically less developed to a more developed country. While this seems to characterize the experience of most migrants (United Nations, 2017), future studies should investigate the acculturation dynamics experienced by migrants moving from an economically developed to a less developed country and specifically test our proposed relationships between opportunities for development and country identifications. The acculturation challenges faced by this group might in some ways be the mirror opposite of those encountered by most migrants in our study, whose experience was characterized by the contrasting dynamics of being attracted to the new country (e.g., the better living conditions and job opportunities) and simultaneously rejected by it (e.g., being confronted with ethnocentrism and occupational downgrading). We know very little about this group of migrants, whose acculturation experience might resemble expatriates assigned to “hardship countries” – albeit without the support of a large corporation.

Our study adopted a cross-sectional design, which allowed us to identify associations between the key variables of interest but not to infer cause-and-effect relationships. To establish causality, future studies should conduct longitudinal research to examine how migrants’ country identifications and processes of (dis‐)embedding in destination and origin countries unfold over time. For example, longitudinal studies could examine whether migrants’ community embeddedness and career embeddedness may reinforce each other over time, and whether there is positive or negative spillover between the work and life domains for migrant retention in the destination country.

Longitudinal and qualitative research designs could also shed light on issues of migrant self-selection and destination selection, i.e., the assumption that certain push- and pull-factors may motivate specific types of people to migrate and/or move from certain origin countries to certain destination countries. In this study, we examined two “pull motives” related to the destination country – opportunities for human development and value-congruent societal practices. These factors do not only affect migrants’ destination country and origin-country identifications and, ultimately, their ability to achieve high levels of embeddedness in the destination country, but they may also influence the decision to migrate in the first place. Future studies should address issues related to self-selection – including what sets migrants apart from their non-migrating counterparts, what combination of ‘push’ and ‘pull’ motives influences the decision to migrate, and what factors determine migrants’ choice of destination country – to understand the effects of these pre-migration factors on acculturation dynamics and embeddedness in the destination country.

In response to recent calls for IB researchers to account for the asymmetrical nature of cultural distance (e.g., Maseland et al., 2018; Tung & Stahl, 2018), we developed a new measure to test whether a societal values–practices match (or mismatch) in the destination country affects migrants’ embeddedness. The application of this measure was a first – and, we believe, promising – attempt to address the “illusion of symmetry” (Shenkar, 2001) in the migration context. More research is needed to articulate how value-congruent societal practices may shape migrants’ country identifications, acculturation dynamics, and embeddedness in the destination country. To avoid ecological fallacy – a common problem with cultural distance measures – future research should also measure individual values to complement cultural distance or cultural values measured at the national level, on which we relied. This would also address another potential problem observed in values surveys, such as those employed by the GLOBE study, namely a possible conflation of values and marginal preferences. Maseland and Van Hoorn (2010) have argued that many of the observed differences in values survey scores may not be cultural in nature, but simply reflect differences in circumstances between groups of people. We encourage future research to utilize alternative approaches to measuring values (see Maseland & Van Hoorn, 2010). Further, given the known limitations of the GLOBE data (particularly, the small sample sizes for some countries), researchers may use alternative frameworks, such as the World Values Survey data, for determining cultural distance or differences in societal values (Beugelsdijk et al., 2018).

Implications for practice

Our study also offers important practical implications. Migration is “among the most emotive subjects in contemporary societies” (de Haas, Castles, & Miller, 2020 :1) and requires a diverse set of interventions to facilitate skilled migrants’ participation and retention in the labor market. At the macro level, governments may shape migration policies, such as skill-selective and skill-specific immigration. However, the efficacy of these policies remains highly disputed (Czaika & Parsons, 2017) and our results, too, suggest that policies on their own may not be sufficient to retain skilled migrants. Consideration must also be given to political, economic, and institutional arrangements in the destination countries to help migrants overcome their liabilities of foreignness and origin. Without national initiatives targeted at changing misapprehensions about and hostile attitudes toward migrants, this talent may not feel embedded and instead choose to move to countries where they feel more welcome (Turkina & Van Assche, 2018; Zikic & Voloshyna, 2023).

Employing organizations can facilitate embeddedness by nurturing a work environment, culture, as well as support and career development systems that buffer against the institutional sources of disadvantage identified in this study (experienced ethnocentrism and occupational downgrading) while at the same time helping skilled migrants to leverage the sources of opportunity available in the destination country (opportunities for career advancement, improved standard of living, well-being, etc.). Although the role of organizational support practices were outside the scope of this study, our findings suggest that equitable employment practices can play an important role in promoting skilled migrants’ career embeddedness and, possibly, also community embeddedness (Harrison, et al., 2019). For example, equality and diversity policies and a climate for inclusion signal the existence of a “level playing field” to migrants (Dietz et al., 2015) and can counteract experienced ethnocentrism and promote identification with the destination country, which is crucial for achieving high levels of community and career embeddedness.

Conclusion

This paper is intended to be a step towards a better understanding of how institutional characteristics of the destination country vis-à-vis those in the origin country as well as skilled migrants’ subjective experiences of the institutional characteristics in the destination country may contribute to their community and career embeddedness. Migration is a topic that is of great importance to society and organizations, and we know little about the factors and processes that govern migrant experiences in both the personal life/social domain and the workplace/career domain. In this article, we adopted an acculturation- and social identity-based approach to understanding the factors and mechanisms that influence migrants’ outcomes in the destination country, and the role that migrants’ country identifications play in this process. In addition, we have offered new constructs that help us investigate migrants’ embeddedness in the destination country, including a new, asymmetrical cultural distance measure that seems more appropriate in the context of migration than existing, symmetrical measures.