Keywords

The introduction stated that in the course of this work I would try to deliver answers to two main questions. First, to what extent do immigrants in Germany engage in transnational activities? Second, how do transnational involvement and immigrant integration relate to each other? The latter has been subdivided into a question on the determinants of transnational involvement and a question on the consequences of transnational involvement—both in relation to integration into the receiving society.

The answer to the first question is rather straightforward. Immigrants in Germany are transnationally active. But the degree of activity depends on the aspect we consider. Visits to the country of origin, a fundamental form of transnational involvement, are rather common. The majority of immigrants in the SOEP regularly pays visits to this country. In this regard, immigrants in Germany do not seem to be different from immigrants in other receiving countries, as, for instance, the US (Guarnizo et al. 2003; Itzigsohn and Giorguli-Saucedo 2002; Portes 2003; Waldinger 2008) or Australia (O’Flaherty et al. 2007). One of the most interesting aspects regarding these visits to the country of origin is that there is no decline across generational boundaries—at least from the first to the second generation. Second generation immigrants visit their (parents’) country of origin as frequently as first generation immigrants do, which links up with some evidence from the US. Haller and Landolt (2005), for instance, found that in some immigrant groups in the US, second generation immigrants visit their parents’ country of origin quite frequently. Remitting, i.e. monetary transfers to family and friends in the country of origin, on the other hand, is less common. The analyses show that only a small share of first generation immigrants remits and that this share is almost negligible among second generation immigrants. These results clearly distinguish between immigrants in Germany and immigrants in other receiving countries. Regardless of what previous study we look at (Portes 2003; Waldinger 2008), sending remittances appears less common in Germany. Since this work provides the first quantitative assessment of immigrants’ transnational involvement using representative data in Germany and Europe, it is unfortunately impossible to put the results into a European context.

Overall, it becomes apparent that border-crossing activities today are a normal part of the migration process, just as they have been in the past (e.g. Lucassen 2006; Wyman 1993). We can agree to Waldinger’s (2008, p. 24) conclusion that “[g]iven the centrality of migrant networks, the myriad of migration strategies, and the uncertain, transitional nature of the migration process, connections linking origin and destination places are ubiquitous.”

The picture is less clear when it comes to the second question, i.e. the relation between transnational involvement and immigrant integration. Let us recall what integration actually refers to. Individual social integration, as laid out in Chaps. 2 and 4, refers to the inclusion of a person into a group or a society. Within the setting of immigration, inclusion into the receiving society can take two forms: assimilation and multiple inclusion (Esser 2006, p. 25). I proposed to link the investigation of transnational involvement directly to the investigation of immigrant integration, arguing that while the latter focuses on the immigrant’s position in and interactions with the receiving society, the former emerges (partially) out of the opportunities and motives structured partly through the selfsame position and interactions. So, how are transnational involvement and immigrant integration linked? As argued earlier, one of the most pertinent questions in this context is whether transnational involvement and integration into the receiving society are concurrent or competitive processes. An assimilationist framework may conceive of transnational involvement as opposing integration if both processes compete for time and resources. To investigate this link, this work has split the overall question on the relation of transnational activities and immigrant integration into two sub-questions: What are the determinants of transnational activities? What are the consequences of transnational activities?

What are the main findings concerning the first of these two sub-questions? The answer is not straightforward. Regardless of what aspect of transnational involvement we look at, a simplistic conception that draws a dividing line between factors that promote or are already an aspect of integration into the receiving society, on the one hand, and immigrants’ border-crossing involvement, on the other hand, finds little support. In that sense, the results of this work mirror findings of earlier studies which conclude that predictions from a simple assimilationist perspective have to be rejected (Waldinger 2008, p. 25; Guarnizo et al. 2003, p. 1233, 1238; Itzigsohn and Giorguli-Saucedo 2005, p. 917; Portes et al. 2002, pp. 289–290). The main findings can be summarized in three points:

First, there are few uniform effects across the different aspects of transnational involvement. That is to say, a factor that promotes remitting (for instance, financial capital) does not necessarily have the same effect on visits to the country of origin. Thus, we have to conclude that just because these activities share one important trait—they link receiving and sending country—it does not necessarily imply that they are equal manifestations of an immigrant’s degree of transnational involvement. This conforms to previous research (Haller and Landolt 2005; Portes 2003; Waldinger 2008) and underscores the need to investigate in more detail how different types of transnational activities (e.g. Portes 2003) relate to immigrant integration.

There are, however, important conclusions to be drawn from the results. Remittances appear to be structured by family ties and obligations and are thus probably only partially voluntary (Taylor et al. 1996). This result is certainly all but surprising considering that the initial migration is often based on a household decision (Landolt 2001; Massey 1990; Stark 1991; Stark and Bloom 1985). It helps us understand, however, why we observe a stark inter-generational decline in remittances, but not in visits. Obligatory transnational involvement does not seem to extend across generations; more voluntary involvement, such as visiting the (parents’) country of origin, does. Yet, the results have also shown that conditions in the receiving country—foremost an immigrant’s labor market integration and the available financial capital—determine the material opportunities and restrictions of this type of border-crossing activity.

Second, however, this does not imply that there are no overall patterns. Quite the opposite is true. This applies foremost with regard to the temporal aspects of the integration process. Over time, i.e. with increasing years of residence, transnational involvement declines. This holds for both (the number of) visits to the country of origin and for sending remittances. Coming back to the earlier question of concurrent or competitive processes, this finding suggests that, over time, integration into the receiving society and transnational involvement do not go hand in hand. A uniform decline over time is not always found in other studies (Guarnizo et al. 2003; Itzigsohn and Giorguli-Saucedo 2002; Portes 2003; Snel et al. 2006; Waldinger 2008). It can, of course, be that this relates to differences in the sending and receiving contexts—a point I will pick up again later. But the longitudinal data of this study arguably allows for better inference regarding the temporal development of transnational activities than the cross-sectional data of aforementioned studies. What is more, there seem to be different trajectories of integration that are associated with different levels of transnational involvement (see also Haller and Landolt 2005, p. 1203).

This brings us to the third point. The analyses have shown that it is not sufficient to consider simple additive effects of factors that may determine transnational involvement. In particular, in the theoretical chapter, I argued that it is necessary to consider how configurations of different dimensions of integration (cultural, structural, social, and emotional) create particular opportunities and motives for transnational involvement. The main idea behind this reasoning is that aspects of structural integration (or assimilation), such as financial capital, will have differential effects, depending on the immigrant’s degree of integration on other dimensions. For instance, if integration on the structural dimension coincides with assimilation on another dimension, this will make transnational involvement less likely, whereas if structural integration coincides with segmentation on another dimension, this will make transnational involvement more likely. As the example of financial capital and network composition shows, this is indeed the case: Financial capital increases transnational involvement if immigrants have co-ethnically homogenous networks, but it decreases transnational involvement if their networks are heterogeneous (or comprise only Germans). Similar results are obtained on the interaction between citizenship acquisition and financial capital and it is important to note that this extends intergenerationally. As a consequence, although transnational involvement and integration into the receiving society may not be directly competitive processes, it looks as though there is a selective affinity between integration into the receiving society and lower levels of transnational involvement. Over time and with increasing integration into the receiving society, ties to the country of origin become weaker.

After having established how levels of integration can be determinants of transnational involvement, the next logical step is to investigate the second sub-question on the consequences of transnational involvement. This may be even more important, because it investigates if transnational involvement has consequences for the immigrants’ life chances in the receiving society. A general, and in a sense traditional, null-hypothesis is that transnational involvement itself will make integration into the receiving country (i.e. through multiple inclusion or assimilation) less likely. What are the results of this investigation?

Concerning first generation immigrants in Germany, there is little evidence that transnational involvement hinders integration. It seems as though immigrants’ integration is largely independent from their border-crossing activities. Other individual aspects, such as the stock of human and cultural capital, play a much more important role than transnational activities (Dustmann and van Soest 2001, 2002; Esser 2006; Kalter 2005, 2006; Tubergen et al. 2004). This is an important result. Not only does it show that loyalties to and engagements with the country of origin do not negatively impact integration, but also that integration is still heavily structured by the resources immigrants bring with them and the conditions they face in the receiving country. Thus, despite claims that focusing on the nation-state is inadequate when we attempt to understand and explain today’s migration and integration processes (Levitt and Jaworsky 2007; Pries 2005; Wimmer and Glick Schiller 2002a, b, 2003), selfsame nation-state is still of great importance. For second generation immigrants, the results are more ambiguous. They indicate that integration and transnational involvement may become competitive processes. There seems to be a negative impact of visits to the parents’ country of origin on the second generation immigrants’ structural integration.

Overall, this study’s main findings can thus be summarized as follows: (1) Most first and second generation immigrants in Germany engage in transnational activities. Regular and occasional visits to the (parents’) country of origin are rather common among both generations, sending remittances is less common among first generation immigrants and rare among second generation immigrants. (2) First generation immigrants’ transnational involvement declines over time with increasing integration into the receiving society. (3) There seems to be a selective affinity between segmentation and transnational involvement, although this is not necessarily a causal relationship. (4) The first generation’s transnational involvement appears additionally structured by family ties and obligations. (5) The first generation’s transnational involvement does not seem to lower its propensity of becoming integrated into the receiving society. (6) Regarding the second generation, there is some evidence that transnational involvement may impair integration into the German society.

However, these results should be treated as preliminary, since this study has a number of shortcomings. First, in light of conflicting, i.e. disconfirming and confirming, evidence, the theoretical model’s validity is, of course, called into question. A crude assertion that transnational involvement and immigrant integration simply oppose each other cannot be upheld, despite the fact that there seems to be a negative association between these two aspects of the migration process. Such an assessment appears too simplistic. Inclusion into the receiving society theoretically does not oppose simultaneous inclusion into the ethnic group (as argued in Chaps. 2 and 4). Theoretically, there are good reasons to argue that, just as the different dimensions of integration are positively linked, transnational involvement and integration are linked negatively. But this rests on some assumptions. We have to assume that transnational involvement actually hinders investments into receiving country capitals. This would be the case if transnational involvement and ethnic capital investments went hand in hand and if ethnic capital investments and receiving country capital investments were mutually exclusive. If the cultural distance between sending and receiving country is large—and thus the social production functions (Lindenberg 1996; Ormel et al. 1999) differ greatly—we have ample reason to assume that this is the case. Transnational involvement may then hinder integration into the receiving society, as it reinforces ethnic modes of production. However, if the cultural distance between sending and receiving country is small—and thus the social productions functions (Lindenberg 1996; Ormel et al. 1999) are rather similar—transnational involvement may not hinder integration into the receiving society. Vice versa, in a situation of lacking integration into the receiving society (i.e. segmentation), transnational involvement may be more likely. However, I would not say that this is the case. It rather seems as if ties to the country of origin are something very normal for first generation immigrants. They may wither with time and the intensity of these ties may systematically relate to how immigrants fare in the receiving country, but overall, these border-crossing ties and activities have a limited influence on immigrant integration.

Thus, the overall impression is that the theory performed moderately. Quite a number of hypotheses are not supported by the analyses. So what does that mean? Is the theory falsified? A model of intentional actions that tries to explain behavioral outcomes by focusing on how opportunities and motives structure trajectories of integration and transnational activities still seems very appropriate. However, in its present formulation it seems inadequate. If we are to draw one overall lesson from the analyses, it seems that the theory is too static. When it comes to the specification of the bridge hypotheses (Esser 1999), the model would profit from a stronger focus on the temporal aspects of immigrant integration and transnational involvement. The empirical analyses have clearly shown that both integration and transnational involvement depend on time. Years of residence or age at migration are but two prominent examples. In its current formulation, the theoretical model does not sufficiently account for the time-dependency of the processes investigated. But since time itself is unlikely to be a causal factor in the process of integration (Esser 1981), it lends itself to investigate how (individual decisions on) integration and transnational involvement are structured over the life course (Elder and Giele 2009; Diewald and Mayer 2009; Heinz et al. 2009; Huinink and Schröder 2008; Wingens et al. 2011; Schunck 2011). Up to now, most models of immigrant integration and transnational involvement—and this holds for this work’s model, too—have paid little attention to the fact that the life course structures the timing of events (exceptions are Levitt 2002; Smith 2002; Kobayashi and Preston 2007; Schunck 2011). From a theoretical point of view, it thus appears promising and necessary to bring together models of immigrant integration, transnational involvement, and life course research. Fortunately, we do not have to start from scratch. These bodies of research are well compatible (Heinz et al. 2009, p. 25; Kley 2010; Huinink and Schröder 2008) and using this compatibility to create a more dynamic theory on immigrant integration and transnational involvement could provide us with a better understanding on how both processes are linked. The link between these three bodies of research, moreover, has already existed from the beginning. In the introduction, I pointed to Thomas and Znaniecki’s The Polish Peasant (1919) arguing that one of the first sociological inquiries on immigrant integration already focused on border-crossing activities. But Thomas and Znaniecki’s work also inspired life course research (Elder 1985, p. 25).

Second, and this extends to theoretical as well as empirical issues, this work investigated only two aspects of transnational involvement, visiting the country of origin and sending remittances. Having only these two indicators of transnational involvement is surely a drawback. What is more, throughout the work, both examples of transnational activities have been treated as similar manifestations of immigrant transnational involvement. As the analyses have shown, this is not the case. Sending remittances, for example, seems less voluntary than visiting the country of origin. At this point, it should be clear that a stronger theoretical differentiation regarding the different transnational activities would benefit the investigation.

Third, regarding this work’s methodological approach, there are two main conclusions to be drawn. First, longitudinal data analysis with a focus on unobserved heterogeneity is a necessary endeavor. Second, longitudinal data analysis with a focus on unobserved heterogeneity is an insufficient endeavor. Let us consider the first point. After the lengthy discussion on (statistical) methodology, I hope that the reader is now as convinced as I am that Hausman and Wise (1981, p. 365) are correct when they state that the attempt to obtain unbiased parameter estimates is an illusionary endeavor in survey data analysis. Nevertheless, statistical models which control for time-constant unobserved heterogeneity are a great advantage—in particular, if we use hybrid or correlated random effect models (Allison 2009; Schunck 2013) that allow us to compare between- and within-person differences. As the analyses have definitely shown, a lot of what appears to be a causal effect is actually due to unobserved, time-constant factors. That is, quite a few associations we can observe cross-sectionally and even longitudinally, for instance between the intention to stay permanently and transnational involvement, are actually spurious.

At the same time, however, this works’ theoretical and empirical parts unmistakably presume a dynamic, bidirectional relation between immigrant integration and transnational involvement. An immigrant’s integration into the receiving society brings about specific opportunities and motivations for transnational involvement and, vice versa, transnational involvement is likely to influence individual decisions that shape integration outcomes. In other words, integration influences transnational involvement and transnational involvement influences integration. But this calls the statistical models into question. Fixed effects regression models (just as random effects models) assume that the independent variables are strictly exogenous. So what can be done about this? We could employ a related class of models that uses instruments (Arellano and Bond 1991; Wooldridge 2002, p. 299 ff.) to replace the respective predictors. This could be achieved through using first-difference models instead of fixed effects models. But is this really sufficient? If we in fact assume that immigrant integration and transnational involvement are dynamically linked, then the next step would require modeling these processes simultaneously. This would bring us to a very different class of models, i.e. structural equation models (Bollen 1989; Engel and Reinecke 1994), which at first may appear at odds with the models discussed and used in this work. Admittedly, structural equation models are motivated by very different problems of data analyses. But recent work (Allison 2009, p. 87 ff.; Giesselmann and Windzio 2012, p. 184 ff.; Bollen and Brand 2010; England et al. 2007) has demonstrated that both approaches are compatible. Combining these two approaches would bring about the great advantage of direct empirical assessment of the interdependencies between immigrant integration and transnational involvement. But, alas, this is beyond the scope of this work.

Fourth and lastly, despite its great advantages, the data used in this study has its particularities. The first generation immigrant population in the SOEP is rather mature. Most of these immigrants have lived in Germany for a considerable time before they became part of the SOEP. What is more, we are likely looking at a positive selection when it comes to integration outcomes, as argued in Chap. 5. Therefore, it might be that differences in transnational activities relate to the German context, but it might also be that immigrants in the German data are at a later stage in the life courses, which may impact activities linking receiving and sending country.

We have seen throughout the book that transnational involvement among immigrants in Germany in many instances resembles transnational involvement among immigrants in other receiving countries. But can we compare the results in such a simple manner? There are many differences between the studies and the data used, on the one hand, and between the receiving and sending countries and their migration regimeson the other. Without comparable data, as I have argued before, it is hard to compare the results. But even if we had comparable data, the question on how different contexts of reception (Portes and Rumbaut 2001, p. 46 f.) mutually shape immigrant integration and transnational activities would be a topic by itself. If we briefly consider a few contextual aspects in Germany, this becomes evident. Germany is a conservative welfare state (Esping-Andersen 1990) and this has been shown to shape integration outcomes (see for instance Kogan 2006). Moreover, German immigration policies, such as restrictive citizenship laws (Kivisto and Faist 2010, p. 68; Brubaker 1992), may increase the hurdles for immigrant integration (Kogan 2007) and create additional motives for transnational involvement. Germany is located within the European Union, which grants its members relatively large freedoms regarding movement and settlement within its boundaries while, at the same time, trying to seal its outer borders (Kivisto and Faist 2010, p. 67 ff., 253 ff.). This may result in reduced costs for intra-European transnational involvement, but may increase costs for border-crossing activities that transcend the EU’s borders. These are but a few points. A coherent investigation into the influence national and supra-national contexts exert on immigrants’ transnational involvement begs for a stringent, comparative research design, which, obviously, is a promising avenue for future research.

Future research could thus: (1) Pay more heed to the dynamic relation between immigrant integration and transnational involvement; (2) Investigate more closely how different forms of transnational involvement relate to immigrant integration; (3) Investigate from a comparative perspective if and how transnational involvement, in its scope, determinants, and consequences varies with different contexts of reception.

I hope I have convinced the reader that it is worthwhile to study immigrant integration and immigrant transnational involvement jointly. Both are different sides of the same coin: migration. A division thus seems defective. It is counterproductive to investigate immigrants’ border-crossing activities without relating them to the nation-state(s) and the communities they are embedded in. But it is equally counterproductive to investigate immigrant integration without paying attention to the fact that immigrants’ embeddedness may span across borders.