The Selective and Diffusion Functions of Social Capital: Why So Many Migrants Out of So Few Places?
Abstract and Keywords
Despite a high migration potential in most developing countries which send many migrants abroad, there is a high place selectivity. Typically, in the few places out of which many emigrate abroad, migration as a behavioural pattern rapidly diffuses. Understandably, propitious conditions such as labour recruitment will raise desires among potential migrants. Relative frustration ensues. Yet changing preferences concerning exit, voice, and in situ adaptation do not automatically translate into growing levels of migration. In contrast to changing macrostructural opportunities and increasing desires, access to information, control, and other people's resources do not immediately improve to the same extent. Thus, there then becomes a lag between rising opportunities and expectations, on the one hand, and actual migration, on the other hand.
Despite a high migration potential in most developing countries which send many migrants abroad, there is a high place selectivity. Typically, in the few places out of which many emigrate abroad, migration as a behavioural pattern rapidly diffuses. Understandably, propitious conditions such as labour recruitment will raise desires among potential migrants. Relative frustration ensues. Yet changing preferences concerning exit, voice, and in situ adaptation do not automatically translate into growing levels of migration. In contrast to changing macro-structural opportunities and increasing desires, access to information, control, and other people's resources do not immediately improve to the same extent. There is a lag between rising opportunities and expectations, on the one hand, and actual migration, on the other hand. The ties that bind people involved and the resources inherent in these ties play a crucial role in overcoming this lag and, consequently, lead to the self-feeding migration dynamics of chain migration. The mechanisms and benefits of social capital are central in processes of cumulative mobility.
We know from research on social movements that recruitment of participants in collective action usually proceeds by means of pre-existing social and symbolic ties. Thus, mobilization is more likely when the members of the beneficiary population are linked by such ties. Personal and organizational networks are important for recruiting participants, and a fair amount of empirical evidence supports this claim (McAdams, McCarthy, and Zald 1996). Although international migration rarely occurs as organized social movement, the mechanisms used to study mass collective action shed light on geographical mobility, in particular the actualization of migration potentials. We need to identify the features which account for endogenously induced chain migration out of very few places.
Four propositions guide the analysis in this chapter:
(p. 144 ) 1. Migration processes only go beyond pioneer migration if brokers such as pioneer migrants succeed in mobilizing specific kinds of resources via social and symbolic ties, such as economic and human capital, reciprocity, and solidarity. In particular, specific reciprocity and focused solidarity do not suffice to stimulate chain migration. They merely foster migration in rather closed kinship circles. Instead, generalized reciprocity and diffuse solidarity beyond kinship groups is necessary. Only then do the various mechanisms of social capital reduce transferral costs, stimulate a critical mass of migrants and reach an ever-growing supply of potential migrants. Otherwise, migration quickly stagnates.
2. If migrant networks are available, migration tends to become self-perpetuating because each act of migration strengthens the necessary ties and creates additional resources that promote and sustain more migration in processes of cumulative mobility. The steady expansion of networks yields feedback loops that are particularly vigorous. Importantly, factors external to this dynamic become less relevant the more an international migration system matures. The endogenous dynamics of migration flows then become predominant.
3. Despite these cumulative effects, migration flows tend to reach turning points at which migration does not increase further. This is somewhat independent of factors outside these processes themselves, such as wage equalization between the country of emigration and immigration in the case of economically motivated migrants. In the long run, exogenous factors intertwine with endogenous ones: The immigration states are likely to implement restrictions that do effectively curb older patterns of flows but do not prohibit new flows along established migrant networks.
4. The first three propositions usually do not apply to the same extent to asylum seekers due to the overwhelming importance of external force in producing and halting flows of refugees.
The first part of this analysis explains the function of social capital in selecting potential migrants and the diffusion effects. Second, the analysis presents a stage model of international migration processes as diffusion, ranging from the ties mustered by pioneer migrants to the formation of migrant and migration networks, the diffusion of chain migration and cumulative mobility, and finally, to the decline of specific international migration flows. Third, the Turkish–German case study exemplifies the sequence of endogenous processes in international migration, constantly halted and modified by immigration state policies.
(p. 145 ) The Selective and Diffusion Functions of Social Capital in Migrant Networks
Selection of migrants comes in three forms: organizational recruitment, personal contact, and brute force or indirect violence. First, organizations recruit workers abroad. For example, this happened in the early stages of European contract worker recruitment in the 1960s when immigration country governments set up agencies abroad to select labour migrants in cooperation with emigration country institutions. Other cases are travel bureaus or smuggling organizations who arrange trips for those who cross borders illicitly. Second, personal contact of potential migrants or refugees perform decisive selections. For example, one member of the family works abroad and others join them as tourists and later legalize their status, if possible. Third, brute force or indirect violence act as a cause or a catalyst of refugee movements in cases of unrest, persecution of minorities, civil war, and ecological disaster. Then force majeure primarily propels migrants.
In all three situations, potential migrants and refugees do not choose in isolation from one another. They are usually not aware of what everyone else in the broader groups or in networks surrounding them is doing. Rather, they tend to take their cues from those in social proximity—friends, family members, neighbours, fellow-villagers, or colleagues. Each member of the group or network takes into account what others do before deciding to join in. Ties operate in all three situations. In the first one, formal ties matter: organizational recruitment offers pioneer migrants a chance venture abroad. In turn, many pioneer migrants later provide the ties for emergent migrant networks. Moreover, personal contacts are sometimes decisive to get on recruitment lists of organizations pursuing selection. The second type of selection through personal ties characterizes the overwhelming majority of migration after the initial start-up phase. Once pioneer migrants have established a migration road, more and more potential migrants are drawn into it by information from former migrants. Veritable migration avenues then sprout. The third type is a limiting one because it makes us aware of severe structural constraints in making choices whether to stay or to go, exercise exit, voice, or in situ adaptation. In all three situations ties in networks carry obligations, reciprocal patterns, and solidarity. Ties provided by networks reduce the risks associated with international migration because individuals can expect help from earlier migrants to find a job and housing.
Thus, resources may congeal in networks of cooperation. These mechanisms do help migrants to muster the benefits of social capital— (p. 146 ) information, control, and scarce resources of others, such as money. Because of lack of information about the labour and housing market in the country of destination, pioneer migrants and early refugees experience even higher risks and costs than those who follow later. For refugees, information about reception centres in potential destination countries are valuable resources. Often, only informal agreements and not legal contracts undergird these kinds of transactions between movers and intermediaries. Migrants then tap the resources of pioneer migrants and brokers. Very often, movers know who awaits them and many probably already know their prospective employer (Tilly 1990). When migrants arrive in the country of destination on prepaid tickets, they are expected to pay back the expenses defrayed beforehand. In addition, migrant networks help those who stay behind. Eased access to work and housing abroad helps migrants to send back remittances. After the migration of the first individual, the monetary and psychological costs of migration are substantially lower for friends, relatives, and selected members of communities in the country of origin. Information is most valuable when there is little of it; and it is always worth more, the more heterogeneous the network is.
Only when brokers initiate migration can international South–North migration evolve on a larger scale. Brokers work very selectively and do not cover all places in any given nation-state in the South. Therefore, migration does not start from all places, and not at the same time. This is a preliminary answer to the second part of our question: Why so many international migrants out of so few places?
When the selection of migrants reaches a critical mass, the diffusion function of social capital comes to the fore. In general, as the stock of social capital resources increases in networks of cooperation, the self-feeding processes of migration strengthen. The network expansion affects both the magnitude and the composition of additional flows, and it supports the direction of migration already given by the structural linkages between emigration and immigration countries.
We can discern the following stages in migration processes in a schematic way:
Phase I: Start and acceleration
Phase II: Climax
Phase III: Deceleration
• thresholds: the effects of pioneer migration
•diffusion and chain migration
culture of migration: migration as a way of life
•formation of migrant and migration networks
•self-feeding effects: cumulative mobility
•the exhaustion of mass migration
(p. 147 ) Phase I: Start and Acceleration
Threshold Effects: The Importance of Ties in a Critical Mass
If we compare places of origin which are very similar regarding both people's desires to move or stay and the opportunity structures they face, we often find widely varying migration rates. It has been repeatedly observed that the number of people moving abroad from—in this regard—two most similar villages can be very different (Struck 1984). We have mentioned part of the answer: different stocks and forms of social capital. To start with, a threshold model of collective behaviour (see Granovetter 1978) gives situation-specific explanations of moving and staying that do not explain outcomes solely in terms of structures, desires, and expectancies of agents before the movements begin. Instead, applied to international migration, this model can be used to look at how migratory processes unfold. This means that migration is dependent upon the number or proportion of other potential movers who must make the decision before another agent does likewise.
Threshold models are a special instance of critical mass models. One of the fundamental assumptions is that as the rate of knowledge in a system increases up to a certain level, there is very little adoption, but once this threshold is passed, further increases in awareness lead to increases. As to migration, it will not start unless there is an initial number of aware potential migrants. In this way, the threshold model specifies the stress-awareness model discussed earlier. The threshold model is superior to game-theoretic models in which the players must choose in parallel, without knowledge of others' strategies; with the partial exception of iterative games. The threshold model imagines serial rather than parallel choices: serial choices mean, in essence, that each potential migrant looks around to see how many others are participating before deciding to join in. By contrast, parallel choices imply that all agents know what the respective others around them are doing (Macy 1991). In the case of parallel choice, actions are solely a by-product of an underlying interest in marginal utility. Following the more realistic assumption of serial choice, we need to focus on the distribution of thresholds and social and symbolic ties through which members learn about the actions of others. Hence threshold models are of particular importance in understanding situations where opportunities and desires are favourable to international migration but no movement occurs. Immobility in the face of favourable macro-structural opportunities are such cases.
There is a serious limitation to the threshold model that can be (p. 148 ) overcome with a relational analysis. It shares with game-theoretic models the assumption of rational agents with complete information. This means that all participants know what all those involved do. In this view, the analysis then proceeds in a classical rational choice fashion—the overall risk level determines the expectations of returns. In other words, the central question is: How much does a migrant or refugee risk when migrating, being the first person to do so? According to this model, responses to the actions of a small group in social proximity would not appear to make much sense by any reasonably demanding criterion for rationality. Given this generally unrealistic assumption of complete information, rational actors are concerned primarily with the overall risk level, and not with the actions of their immediate circle. However, cooperation depends not only on the level of risk and the strength of collective interests—for instance, the interest of a kinship group—but also on the network of social and symbolic ties that channel the necessary chain reactions.
Hence, threshold models need to be extended in order to understand how participation in migrant networks follows the contours of the social structure and the resources inherent in the ties: potential migrants usually do not choose in total isolation from each other, nor are they completely aware of what everyone else in the group is doing. If they are not first-time pioneer migrants, they rather take their cues from their ties with significant others, those in social and/or symbolic proximity—kin, friends, neighbours, or colleagues. Each member of the group or network takes into account the ties with others and what others are doing before deciding to join in serial transactions. Whether and how migration then takes place crucially depends on how the mechanisms of social capital operate and what benefits they generate. It follows that the spread of migration crucially depends on the properties of the network of the respective social and symbolic ties, linking the members of the relevant groups. Only when taking into account the web of ties, can we hope to trace how the responses of potential migrants might lead to a critical mass of migrants. However, the lower the degree of freedom of the people involved, the less applicable the threshold model becomes. At the extreme, very brute forces leave practically no choice to observe what other people do.
According to a critical mass model, the central issue in migrant network formation is the problem of overcoming the start-up costs. The ties mustered by brokers are important for overcoming these costs, viz. thresholds. Brokers are the third persons who benefit, the tertius gaudens. The broker position denotes the freedom a person derives from conflicting group affiliations on the basis of multiple ties reaching into (p. 149 ) various networks (Simmel 1995: 297). Brokers derive benefits from negotiating and facilitating ties between other players (Burt 1992: 30–4). Here, the term refers to being the third between two or more parties trying to use the same set of ties. Because of their positions in networks, brokers have a high degree of information and exert a great amount of control. Brokers can be recruiters who work for companies abroad or employment agencies in the emigration country. They can also be pioneer migrants who capitalize on their experience, and thus respected individuals in the emigration or immigration communities or professionals in organizations concerned with labour recruitment. Alternatively, they can be commercial migration merchants who transport undocumented or falsely documented migrants across nation-state borders. Another type of tertius gaudens are gatekeepers. These are persons—migrant or native—who provide new migrants with access to employers, landlords, or public authorities. They are intermediaries who enable contacts of potential and actual migrants to employers, legal authorities, and other institutions.
Many brokers act as transnational entrepreneurs supporting mechanisms which include exchange-based obligations and reciprocity. They benefit from money or social debts incurred to them during the process of migration—instances of exchange-based obligations. For example, villagers in developing countries are likely to reap the best results from international migration if they all agree to sponsor selected individuals for graduate studies at a university in the North. The individualized strategy would be illegal entry in the country of destination. Sometimes brokers are themselves constrained by social norms when responding to legitimate claims for assistance in reciprocal transactions. Of course, reciprocal ties do not matter for those migration merchants and their helpers who simply abandon their human cargo whenever feasible.
The Formation of Migrant Networks: The Ties of Brokers
Once the number of movers reaches a critical threshold, capable of sustaining an extensive network of social and symbolic ties, migrant networks can emerge. Migrant networks are important overall mechanisms for enabling exits because they are characterized by low jointness of supply. Jointness of supply or non-rivalry means that the benefit of a collective good such as a network of cooperation does not diminish by other people also sharing it. Low jointness of supply is a fundamental characteristic of reciprocity and solidarity. With reciprocity and solidarity, embedded in migrant networks, the transaction costs of movement become lower. It decreases maintenance and adaptation costs. The (p. 150 ) question then is how migrant and migration networks come into existence. Networks that extend beyond the nuclear family and tightly knit kinship groups are particularly interesting because they are needed for chain migration to unfold.
Two assumptions are necessary: first, a higher overall prevalence of social and symbolic ties and a higher density of such ties promotes collective action in migrant networks. Note that a crucial proviso has to be added: the content of these ties needs to extend beyond specific reciprocity and focused solidarity to include exchange-based obligations, generalized reciprocity, and diffuse solidarity. Otherwise, the benefits of social capital remain limited to the parochial few. Only then does migration reach people outside kin groups who have pioneer migrants. Second, network centralization through organizers such as brokers tends to have a positive effect: the concentration of ties around, say, pioneer migrants, has an accelerating effect on migration. Additionally, the value of centralization increases as the heterogeneity of benefits derived from social capital grows (for a micro-level formulation of the same thought, see Marwell and Oliver 1993: 122–3). This is important because resources necessary to migrate such as money, travel information, accommodation, work, and other necessities are unevenly distributed among members of kin groups, friendship circles, and communities. When resources are distributed heterogeneously among the group, limitations on the key broker or organizer—the one with the highest amount of social or symbolic ties, viz. the largest ego-centred network—set by the amount of resources she might devote to mobilizing become much less important. Instead of ties that eventually induce the whole large network into migrating, brokers in a group with heterogeneous resources usually serve to influence others in their decision to move or to stay by transacting with a selected subgroup of the larger network or collective. Such brokers with large networks are very likely to be in contact with the few key people in the group, such as a village community or a city neighbourhood; for example, teachers in villages usually are in contact with both pioneer migrants abroad and local persons who are multipliers, such as the village notables and others, often via indirect contact with their students. What mechanism of social capital is operative—obligations, reciprocity, or solidarity—depends on the structure and content of ties within groups. Note that the people reached by the brokers need not be tied to each other directly and thus form circles or cliques. All that is necessary for them to be reached by ties is their common position in a network—a structural equivalence to the positions of brokers.
Often, migrants consider personally-known brokers more trust (p. 151 ) worthy than anonymous agencies (for many, see Spaan 1994: 109). From the potential migrants' point of view, these intermediaries range in decreasing benevolence from relatives and friends in their own, ego-centred networks, through casual smugglers and border guides, to sophisticated organizational recruitment which organizes the whole process, purveys documents, and provides guidance abroad.
Broker-induced territorial exits are the rule rather than the exception (see the examples, albeit using a different terminology, in Castles and Miller 1993: chs. 4 and 5). In most instances, a relatively small cadre of highly interested and resourcefully tied people produces migrant networks. It is not the effort of the average member of the community. Central is the importance of a critical mass of large contributors— ‘large actors’ (Olson 1965)—to the formation of emergent migrant networks. This small group of necessary inducers consists of pioneer migrants, recruiters, migration merchants, and other facilitators. The evolving networks of cooperation then give benefits such as information, they increase the awareness space of potential migrants and widen access to resources of pioneer migrants and brokers in the country of destination.
The question now is what kind of ties are necessary and what role they play in network formation. And it is important to point out that it is not only the strength of ties that matters but also their content. An influential argument is that strong ties tend to form cliques, while weak ties tend to bridge cliques and bring everyone into the same network (see Chapter 4). Therefore, weak ties can often be a better basis for collective action (Granovetter 1973). In our case, this concerns the formation of migrant networks. Here, the argument is that decentralized nets and centralized wheels of weak ties are equally effective network structures for migration.
The weak ties argument has to be qualified in two respects to be useful for our purposes. First, not weak ties per se are useful, but their tendency to be centralized. This is a plausible assumption because without a centralization of ties by pioneer migrants and brokers, many potential migrants could not leave their country, and new migrants abroad could not adjust because they lack referrals. Second, the exclusive focus on social ties, whether strong or weak, misses important connections between potential migrants, pioneer migrants, and brokers. We have repeatedly come across case studies showing that ties extend beyond kinship groups and friendship circles and involve the whole community (e.g., Klaver 1997). Symbolically tied collectives that matter range from small village communities to ethnic or even transnational religious communities. For example, symbolic ties such as a (p. 152 ) common understanding of a long history can form the basis of solidarity in village communities and, in strong cases, give access to migration networks to virtually all willing members of the village; while the absence of a shared history, ancestry, or otherwise shared understandings can indicate a low degree of symbolic cohesion and is indicative of the fact that only members of smaller groups have access to their own particular migration networks. Thus, symbolic ties can significantly enhance the effect of social ties. In addition, symbolic ties carry a role of their own. This becomes most obvious in the virtual absence of strong social ties. For example, on the basis of a shared understanding concerning regional origin, persons in the immigration community may act as substitutes for pioneer migrants or early refugees and contribute to both the selective and adaptive functions of social capital in networks.
Phase II: Climax
Chain Migration as a Diffusion Process
It is likely that migrant networks of cooperation, a regular circuit in which migrants retain claims and contacts and routinely return to the country of origin, brings in its wake the migration of related individuals or households. This is a typical ‘friends and relatives effect’. When looking at the dynamics, this process is akin to snowballing: the more immigrants of a given place stay in the destination region, the more want to come. It takes time to develop the chain and this is the reason why we see it full-fledged only in later phases of international migration. When brokers and gatekeepers find worthwhile benefits in advising and channelling movers, when norms of reciprocity such as remitting money home can be enforced and when forms of mutual aid among migrants create broad commitments to other migrants, then networks of cooperation that encompass movers and stayers begin to diffuse rapidly.
In order to understand the importance of networks of cooperation for the diffusion process, let us first consider its opposite, namely migration in the context of atomistic structures. Banfield (1958), for example, reports on a town in Italy's Basilicata region where prospective emigrants could not leave because their numerous fellow-townspeople abroad had severed all ties with the country of origin. We can infer that social ties in this town must have been extremely fractured—even more than is the rule in Southern Italy. This fact made chain migration (p. 153 ) impossible and stopped the diffusion process after it had taken off haltingly a few decades before. The reasons can be found in the absence of networks of cooperation that could continue the migration process. Associations, community organizations, clans, or other forms of generalized solidarity and diffuse solidarity were conspicuous by their absence. Moreover, the nuclear family household, the multilateral kinship system, and dyadic patronage—the basic forms of social organization in this part of Italy in the first half of this century—were precarious. Therefore, chain migration to the United States in the late nineteenth and early twentieth century, based on a ‘hometown’ society of the Southern Italian type necessarily ran the risk of leaving some prospective emigrants out on a limb (Macdonald and Macdonald 1964: 90–1). In such a situation, a potential sponsor abroad may desert her family, friends, and clients, when she adapts to the immigration country without keeping up ties with home, or when the frequent conflicts in this kind of social organization rupture bonds. Mechanisms of control reaching beyond the territorially absent nuclear family are missing. Also, a brokering sponsor in the immigration country may fulfil her duties of focused solidarity to those few fellow-townsmen who are close relatives, friends, or clients, and not give any thought to the majority to whom she has no customary obligations. This behaviour prevents migrant networks from expanding and exit options only extend to family members. Consequently, no chain migration evolves.
The transmission or diffusion of migration can be conceptualized as a two-stage process. The first stage is awareness—as suggested by the threshold model—in which brokers send signals by means of social and symbolic ties; sometimes reinforced through mass media portrayals of immigration country conditions. The second stage is the readiness to migrate. The diffusion effect is the cumulatively increasing degree of availability of obligatory, reciprocal, and solidary ties to migrate, resulting from the activation of networks. Then the reduction of transaction costs is positively related to the rate of adoption of migration if brokers exist, and negatively related if social capital can only be transferred under conditions of great risk or uncertainty.
Typically, diffusion processes are characterized by a change from an accelerative to a decelerative pattern. In an accelerative pattern the rate of adoption is proportional both to those who have adopted an item and to those who have not. Decelerative means that the rate of adoption is proportional only to those who have not yet adopted an item (see also Coleman, Katz, and Menzel 1966).
It is most likely that accelerative processes operate in the middle stages of migration courses, including all those persons connected by (p. 154 ) existing networks of cooperation. Only when international migration starts to reach beyond the well-entrenched circle of relatives of pioneer migrants, do we expect accelerative diffusion to blossom into chain migration. Once the accelerative processes have reached the majority of the potential migrants, diffusion is likely to turn into a decelerative pattern. Migration keeps going on but slows down over time—until the supply of potential migrants is exhausted. Accelerative diffusion implies exponentially growing chain migration, decelerative diffusion steadily declining flows through migrant networks. Thus, the nature of the process is bound to change as the diffusion proceeds, accelerative diffusion being most operative in the middle and decelerative diffusion when migration is dropping off.
The facility with which chain reactions spread through networks of cooperation depends on the kind of social and symbolic ties that link people, groups, and organizations. While strong ties embody greater potential for influencing behaviour, weak ties may be more effective as diffusion channels (Rogers and Kincaid 1981). Chain migration then evolves from one reference group in a community, extends across communities, and even across the boundaries of isolated clusters. As mentioned above, diffusion processes require types or mechanisms of social capital that extend beyond narrow kinship clusters.
As the stock of social capital inherent in social and symbolic ties grows over time—and as international migrant experience increases—migration becomes progressively less selective and thus spreads, for instance, from the middle to the lower segments of the socio-economic hierarchy (Massey et al. 1987). The socio-economic distribution of migration is tightly interlinked with life-course patterns. Once networks have developed to the point where a foreign job is within easy reach, international migration becomes a preferred strategy among poor families seeking to alleviate pressing economic needs caused by many dependents and few workers (Balán 1988: 56).
Gender-specific networks add to the dynamic of international migration. Here, we can see how migration networks overlap with social support networks in the area of origin. It is not true for many cases of today's international migration that women are the ones ‘left behind’ while their husbands go abroad (Sassen 1988 and Fernández-Kelly 1994). In some cases, network diffusion gives relatively immobile women a chance for in situ adaptation in a gender-specific way. For instance, single mothers from Mexico leave their children with their mothers, sisters, or other relatives for varying lengths of time while they work in the USA (Chavez 1992: 122). Child-fostering provides a stable source of income for non-migrant women, and a woman who fosters the (p. 155 ) children of relatives and friends abroad may be able to increase her stock of ties. At some point in time she may draw upon them when sending her children, other relatives, or friends abroad (Simon and Bretell 1986: 45–6). This strategy is gender-specific because in many southern and northern countries women are known to be more active in support networks related to kin and friends (see also Bott 1957: 138).
Migration as a Self Feeding Process: The Mechanism of Cumulative Mobility
The diffusion of migration is a typical instance of cumulative mobility as positive feedback causation. Once the number of network ties in migration flows reaches a certain level, migration becomes self-perpetuating, because it reinforces and augments the ties and social capital necessary to sustain it. Once migration processes are underway, they fuel themselves. At some point in international migratory processes, networks of cooperation sustain population flows quite independently of objective economic conditions in the areas of origin and destination (Hugo 1981). Existing research has described processes of cumulative mobility from a micro-level and a macro-level vantage point. Despite contrary claims, a meso link is still missing and is provided here.
(1) The Macro-Level Vantage Point
On the emigration side, migration originates in profound transformations of developing societies, involving processes of mechanization, capitalization, and commercialization in agriculture and mechanization and profound rationalization in industry. Over time this produces socio-economic changes that encourage these trends and make subsequent migration more likely. For example, the mechanization of agriculture that began in many developing countries during the 1950s displaced peasants and agricultural workers. Similar processes apply to industrialization. Given access to work in the former colonial centres, in former white settler colonies and foreign labour recruitment programmes during the 1960s, industrial workers who searched for economic betterment and, later on, displaced peasants from the countryside who sought informal insurance and economic survival took up international migration. When networks evolved during and after the open door or recruitment phase, they enabled migrants, their kin and significant others to use the gates of family reunification, marriage migration, and illicit entry to immigrate in increasing numbers.
Effects of these chain migrations in the areas of origin have been manifold. Particularly in small villages, some migrants invested in (p. 156 ) mechanization and used more intensive production methods so that other villagers were finally displaced. More households were able to invest in capital-intensive production methods, and more families withdrew from cultivation. In this way, migration exacerbated the falling demand for agricultural labour and accelerated the shift to commercial agriculture. Temporary migration has been a strategy of risk diversification in rural households. Foreign wages have sometimes led farmers and peasants to cultivate their land less intensively than before or even let it lie fallow. If these migrants buy land, the outcome might be that there is less land under intensive cultivation in the community, that local food production is reduced, the price of staple crops raised, and the demand for labour decreased. This, too, gives incentives to the remaining members of the community to move (see e.g. Cornelius 1991).
If, on the other hand, land is more intensively cultivated, as farmer migrants can now afford more capital, more emigration evolves since less manual labour is needed (Massey 1990). However, remittances spent on agriculture could actually increase agricultural profits. They can help to develop productivity and output, and migrant farmers will keep marginal land under production (Cornelius 1991: 48). In this latter case we could not expect economic feedback effects to encourage further migration. Instead, forms of familial obligations and specific reciprocity are kept alive and undergo in situ adaptation more slowly than under circumstances of large-scale emigration.
(2) The Micro-Level Vantage Point
There is much empirical evidence that a history of prior migration increases the likelihood of future migration. This is true in two ways: for the repeated migration of individuals and for new migrants who choose the exit option. First, in the case of ready transferability of social and other forms of capital, we would expect that, ceteris paribus, each trip to the immigration country increases the likelihood of subsequent trips (Massey et al. 1993). An economistic focus upon the sunk costs—costs that have been expended in any long-standing line of action—suggests that the more investments go into certain lines, the higher is the stake (Becker 1975). Migrants can be thought to invest time and energy in social and symbolic ties, as well as more tangible resources in pursuing migratory careers. The costs of exit from such a line of behaviour are thus substantial—such as costs of readapting completely to life in the country of origin. Second, for potential migrants to exit, their desires need to change. For example, role theory predicts that the process by which people learn any new role may (p. 157 ) evolve into a stable pattern (Dahrendorf 1960). If we apply this to the problem at hand, new migrants are gradually socialized into their role as international movers, emulating the behaviour of earlier migrants. Ultimately, the ‘Tocqueville effect’—desires rising more rapidly than opportunities (Chapter 2)—makes more and more potential migrants choose the exit option, even if they have to resort to unauthorized migration.
(3) The Meso Link
The mechanisms of positive feedback cumulative mobility can be summarized as shown in Figure 6.1.
Recruitment policies of immigration countries do indeed induce pioneer migrants to exit and move from South to North. Many pioneer migrants who establish themselves abroad but return continuously or intermittently, maintain ties with people in their kin groups and (former) communities. Thus, potential migrants see the advantages of also moving abroad. The diffusion of migration beyond narrow and small groups occurs through transaction cost reducing migrant networks. The rate of migration keeps going up. After a while, the admission policies of the countries of destination do not allow the reception of as many immigrants as are willing to come, partly because the desires
This state of affairs gives rise to family reunification, marriage migration but also illicit entry. Family reunification is an avenue for those who enjoy intra-family reciprocity. Marriage migration is a strategy used by those, for example, who try to recur to obligations of friends and relatives in sending their son or daughter abroad. Migration occurring under weak ties in exchange-based obligations carries the highest risks for migrants. Especially so if migration merchants are involved. The migrants then have practically no chance to enforce social or legal norms should the migration merchant renege and break the informal contract to channel the migrant into the desired country of destination. Given this high level of risk, transaction costs can be exorbitant in the absence of trustworthy ties.
Once set in motion, migration flows develop under their own steam. In particular, generalized reciprocity and solidarity help to broaden the scope for actual migrants, driving up both emigration, immigration and the stock of immigrants in the places of destination. These endogenous processes are, at times, brought into being and altered by exogenous factors which contribute to initiate, to accelerate, or to decelerate flows. Typical examples of exogenous factors accelerating the endogenous dynamics are family reunification and non-refoulement policies. In general, they help to bring close kin into the country of destination, or entice refugees to select certain abodes rather than others as places of refuge. Stricter immigration controls have a decelerating effect. Because of generally efficient border controls and the high transaction costs involved in illicit migration regulated by migration merchants, organized and large-scale circumvention of official policies is unlikely to mushroom. Much more likely are new illicit channels via direct ties within webs of kin, friends, and fellow-members of collectives.
The various mechanisms of cumulative mobility work in a different manner for those whose personal freedom is significantly restricted. For example, persecuted activists, members of target groups, or simply victims of civil wars and environmental disasters are most often involuntary movers. Their exit occurs under the threat of sometimes massive physical and psychic violence and danger for their personal integrity and life itself. Hence, we then expect the process of cumulative mobility to be much less governed by endogenous but more by exogenous factors, such as a single event or a row of events that provoke or force an exodus. Therefore, the internal dynamics of flows of involuntary (p. 159 ) migrants tend to be less self-feeding than in the case of rather voluntary migration. Nevertheless, the distinction is often one of degree. The more migrants have active disposal over their fate, for example, in situations when they do not face acute persecution, the higher the chance that self-feeding processes operate, albeit within much more restricted bounds than in many cases of migration characterized by higher degrees of freedom for the potential migrants. The two limiting cases are countries with civil strife over decades that have either produced a trickle of refugees or forced expulsions of whole religious or ethnic groups at one point in time. Potential migrants in the former situation often tend to have a higher degree of freedom than those in the latter.
Phase III: Deceleration
Towards a Culture of Migration
In some instances of continued chain migration, a culture of migration develops which encompasses migration as an accepted and desirable way towards achieving social and economic mobility, a higher income, a lifestyle which could not be sustained by dependence on local resources only, or a way to challenge the political regime in the country of origin from exile abroad. For such cultures, the experience of geographical movement across nation-state borders and resettlement is formative. A culture of migration often spans several generations, as in the case of the Caribbean islands, where migration to North America and the British Isles has constituted an accepted lifestyle. The social and symbolic ties to the country of origin run on (for many, see Goulbourne 1991). This finding seriously challenges cherished conceptual distinctions, such as the one that migrants break off ties they entertain with these countries and regions after a certain period of time. They either remain sojourners and return to the emigration country—or another destination—or they become settlers and stay in the immigration country. Migration experiences such as the Caribbean islands-USA rarely conform to this dictum: there is, for example, abundant evidence on long-term circulators who work and invest abroad while their kinship base remains on the Caribbean islands (Pessar 1997: 2). This applies to groups among Puerto Ricans since the 1940s and Dominicans since the 1960s. In the West Indies, where the international migration experience dates back much further than in the rest of the Caribbean, people are calling such movers ‘travellers’. In all these cases (p. 160 ) migration has not been decelerating but has proceeded on a stable and high level.
Decelerating International Migration Flows
So far, we have stressed that cumulative mobility—migration begetting more migration—works in a snowball fashion. The model of cumulative causation does not allow for the possibility of turning points, for reaching a maximum amount of migration, when migration stabilizes or falls, well before wages or job availability between the emigration and immigration countries equalize or political conditions in the countries of origin improve. The considerations pertaining to the decelerating stage of international migration processes are particularly difficult to ascertain because the empirical referents are tenuous. Very few of the migration flows induced after World War Two from South to North have reached their zenith. None the less, migration equilibria have evolved in some cases: emigration and (return) immigration have reached a rough balance. Examples include migration between Puerto Rico and the mainland USA—a sort of quasi-international migration, and between Portugal and Germany in the late 1960s, well before the termination of labour recruitment in 1973. Average wage differentials of a magnitude of 10 and 12 to 1 continued to exist between the mainland USA and Puerto Rico, and Germany and Portugal throughout the 1970s, 1980s, and 1990s. Puerto Ricans enjoyed freedom of movement, as did the Portuguese since 1986 in the EU. In addition, migration to alternative destinations did not grow. For example, the main destination of Portuguese workers, France, did not experience an increase in the immigration of Portuguese workers in the past thirty years. On the contrary, a decline occurred as well.
Economic arguments say that it is not the absolute differential between emigration and immigration country that matters. There are turning points that depend, among other things, on wage differentials. One of the key elements has been the so-called inverted U-curve thesis: economic development often first enhances and thereafter reduces the scope and incentives for migration (Faini and Venturini 1994). The thinking behind this hypothesis is as follows: there is usually very little emigration from countries with very low per capita incomes, where people work mostly in agriculture. If at all, there is internal migration. As soon as the per capita income reaches a higher level, we tend to find a higher percentage of persons who can afford the financial means to migrate abroad. The migration potential is highest at this stage. When a still higher income level is attained, the means to migrate increase but the (p. 161 ) actual emigration rates abroad are lower because wage differentials are not as stark any more (Figure 6.2). In this latter echelon, always to be thought along a continuum, human capital migrants are the most likely group to migrate abroad, since they could reap the highest relative profits (see Fischer et al. 1997b: 100).
However, such a view neglects a significant macro-level and a meso-level consideration. First, the inverted U-curve thesis is unsatisfactory because it assumes that a change in macro-structural conditions directly impinges on and translates into changing preferences of potential migrants. As the preceding analysis has shown, we need to take into account the intermediate structures and processes to complement the one-sided analysis of exogenous factors with decisive endogenous variables. Once a recognizable percentage of potential migrants starts moving, others will follow. And once the supply of potential migrants is exhausted, migration rates fall drastically. As usual, when the degree of freedom of choice declines, the applicability of the turning point model diminishes.
Second, the inverted U-curve argument does not include other determinants such as the endogenous dynamics of political regulation of exit and entry. For instance, more restrictive admission policies on the part of the immigration countries have tried to curb flows. Therefore, active labour recruitment has become rare. The more restrictive policies are usually part of post-migration developments. Large numbers of immigrants often raise the fear among the native and settled populations that the newcomers do not learn the mainstream language and practice religious beliefs that are alien to believers of the respective mainstream religions. Often, political parties evoke these dangers in election campaigns, followed—if successful among the populace—by
Deceleration and Restrictionist Policies: The Politicization of Welfare Provisions and Security
Restrictive immigration policies decisively shape decelerating migrations flows. Immigration states have employed both extensionist and restrictionist policies to regulate immigration and adaptation of labour migrants and asylum seekers. In turn, these policies have over time shaped political conflicts over welfare state provisions, external and internal security, and the fear of being culturally or politically over-whelmed by intruding newcomers. These conflicts have not endangered the consensus upon which nation-states are built but have increased the level of politicization of social rights and perceived security, and contributed to restrictionist immigration policies.
When dealing with perceived costs and benefits of immigration, it is of particular interest to look at the interplay of governments, political parties, interest groups, and voters. Politicians relate the shortages in the housing and labour markets, and problems of border control, among other things, to immigration. Although these problems may have other causes, they have become one of the prime targets for voicing dissent. Indeed, immigration has turned into a ‘meta-issue’ (Lasswell 1948): it can be referred to by intermediate organizations such as political parties in explaining many social and economic problems—such as unemployment, and housing shortages—without having to give concrete evidence because the effects of immigration are empirically hard to establish. In referring to these fears and in being responsive to the expectations of their constituency, especially politicians from populist parties have in fact introduced and reinforced xenophobic tendencies. This is not to say that threats to welfare and security in the immigration country are without any real-world foundation. However, through meta-politics low-level threats usually gain out-of-porportion significance. We cannot understand the saliency immigration has gained for public debates if we limit our discussion to the structural deficiencies of border control, internal policing, certain labour market (p. 163 ) segments characterized by a high representation of immigrant workers, and to certain social services in which we find a high proportion of immigrants. Instead, the broader question is how issues of material interest—border control, internal security, wage depression, job substitution, industrial relations, but also dependency of immigrants on social services—have fuelled debates over social justice, ethnic and national identity, and multiculturalism. Debates over identities and norms are generally much harder to resolve than conflicts over interests.
Sovereign states have crucial functions such as welfare, security, and the maintenance of national identity. Regarding welfare provisions, the standard argument for restriction is that continually high levels of immigration could lead to competition between immigrants and immigration country natives for jobs, housing, and social services and eventually even increase xenophobia (Ahfeldt 1993). More dispassionate observers have noted that high levels of immigration can be a factor leading to conflicts over scarce goods not simply because of direct competition between newcomers and the settled population. This version holds that high levels of immigration allow for the politicization of the immigration issue. Immigration can create an overabundance of cheap labour. This could threaten the balance of regulation of labour and housing markets that rests upon the cooperation of labour unions, employer associations, and state institutions. Given an abundant supply of labour willing to work under substandard conditions, individual employers, for example, would have more incentives to opt out of the regulation of labour markets. Unions, in turn, would lose economic and, finally, political power. Ultimately, this could lead not only to specific labour market segments for immigrants but to a levelling of social rights for all, immigrants and natives alike. And continuing high levels of immigration would not allow for the integration of immigrants who arrived earlier.
We could think of another version of this thesis. While the argument just made takes account of the importance of political conflicts as a driving force in the further development of immigration and integration policies, it does not consider immigration or immigrant policies themselves as a main factor contributing to the problems migration and the integration of immigrants pose for solidarity and social rights in receiving welfare states. It could be, for example, that restrictive immigration and integration policies entail a growing politicization of welfare state politics along immigrant-native lines. When potential immigrants from certain groups are rejected more often than others and migrants already present in the territory are barred from social rights, this could be interpreted to mean that even those who have gained entry have no (p. 164 ) legitimate claims to admission and membership. We thus have two versions of the argument, one concerning the effects of settlement, the other relating to very restrictive policies. In both cases we could find increasing levels of politicization over immigrant social rights. Both arguments are plausible, the first one regarding the adaptation of labour immigrants, and the second one with respect to restrictive policies towards labour migrants and asylum seekers.
The first version of the argument has been used to diagnose the ‘Americanization of welfare state politics’ in Europe. Mass migration of foreign labour ‘reduced the political clout of those social strata that have traditionally been the chief source of support for welfare state development’. Specifically, the post-war mass migration ‘diminished the power of organized labour by dividing the working class into national and immigrant camps, by easing the tight labour market conditions that would have enhanced labour's strategic resources and provoking a resurgence of right-wing and nativist political movements’ (Freeman 1986: 83).
This thesis has the merit of linking immigration policies to the broader issue of welfare state integration. Yet there are two reasons to be sceptical about the claim that mass migration has undermined the consensus in European welfare states. First, although nationalist-populist parties in various European countries have run on anti-immigrant platforms, they usually do not attack the welfare state as such—in contrast to the neo-conservative rhetoric and onslaught during the 1980s. Rather, they demand strict restrictions on immigration and the repatriation of immigrants. Furthermore, there seems to be no conclusive empirical evidence that the political support for the welfare state as such has waned, despite various forms of ‘welfare backlash’ that have occurred since the 1980s. Contention has been limited to public assistance programmes while contributory insurance programmes still enjoy wide support among voters (Clayton and Pontusson 1998). Conflicts over immigrant social rights have been limited to tax-financed assistance programmes and benefits. Second, if the mass migrations since the 1960s had pitted the native working class against the immigrant working class we should see more signs of union opposition to continuing immigration. However, the latter has also not been the case (Faist 1995).
In order to explain how immigration and adaptation policies become contentious and lead to restrictionist measures, we have to analyse welfare state integration in three dimensions, functional, moral, and expressive (see also Figure 4.2). Temporary labour migrants are aliens hired to fill niches in the labour market for which employers cannot find (p. 165 ) domestic workers at prevailing wage rates and working conditions. Employers usually prefer to hire temporary labour, often on a rotational basis, i.e. migrants are supposed to return to the sending country after a fixed time period and an exchange of workers is supposed to take place. There are various reasons why temporary labour is hired, for example because it is assumed to be cheaper, more docile and because it may be used by employers in splitting the domestic workforce. Since temporary workers are mostly young, single, and healthy, costs to the welfare state are relatively low and benefits prevail in the eyes of the actors who initiated their recruitment, such as government bureaucracies, employers, and unions. In the course of the 1960s and 1970s migrant workers with secure residence and work permits have gradually lost their function as a ‘conjuncture buffer’ in Europe. Empirical studies show that in Western Europe they cannot simply be classified as part of secondary labour markets (Böhning 1984). And the native population started to perceive the costs of immigrant adaptation in areas such as housing, education, and health.
This is even more visible in the case of asylum seekers. The costs for asylum seekers and refugees to the coffers of receiving welfare states are visible in the short term. Their admission is justified on humanitarian and/or political grounds. Generally, asylum seekers are not allowed to work in the beginning stages and exclusively depend upon social provisions. Therefore, if the number of asylum seekers increases suddenly, we would expect political conflicts to focus on welfare provisions for this group. Indeed, while the numbers of asylum seekers sharply increased since the mid-1980s, recognition rates of refugees declined all over Europe (Joly, Kelly, and Nettleton 1997). The trend has been to grant proportionately more inferior categories of recognition to asylum seekers—for example, humanitarian status. These statuses confer considerably fewer social rights than full refugee status alienship and denizenship.
The important point is that once immigrants are in the territory they have access to a modicum of social rights and, unless a strict rotational principle is enforced, there is a high chance that they improve their legal status. In the post-1960 period there seems to have been a movement from alienship to denizenship among immigrants (Figure 6.3).
As to the security functions of immigration states, the control over external and internal borders can be interpreted as a sign of external and internal sovereignty of nation-states (Weiner 1995: ch. 6). Serious threats to the stability of states through migration are, however, exclusively limited to South—South migration. By contrast, the immigration countries in the North usually effectively control external borders and (p. 166 )
Issues of external and internal control connect, among other things, through the welfare function of immigration states in the North. External control capacities of immigration states and the degree of internal regulation closely interlink. The lower the capacities and willingness for internal control, the higher the call for restrictive border control policies. By contrast, the higher internal control capacities of immigration states, the lower efforts to engage in very rigorous border control. This means that highly regulated welfare states tend to focus their activities on internal control, while less highly regulated welfare states are characterized by a partial absence of rigorous internal controls. For (p. 167 ) example, the increased efforts to control the Mexican–US border since the early 1990s is not simply a function of a long, contiguous, and therefore potentially porous border line. Anyway, such a border could only be sealed completely if the country of immigration was willing to use authoritarian police state methods. Institutions such as the Immigration and Naturalization Service (INS) and the Social Security Administration (SSA) do not exchange information on a regular and sustained basis. In Germany, by contrast, institutions such as health insurances, the German Employment Service (BA), and various institutions active in social security do indeed exchange information. It is therefore potentially much harder for undocumented immigrants to work in Germany with fraudulent papers than in the USA (for more examples, see, Vogel 1996).
The linkage of external and internal control is played out in the activities of interest groups. In less highly regulated welfare states, employer associations have a higher ability to consistently advocate special regulations for categories of workers. This leads to a constant balancing of restrictive and expansionist immigration policies. For example, part of the 1986 USA legalization programme—US Immigration Reform and Control Act (IRCA)—foresaw sanctioning employers for employing undocumented immigrants, and, at the same time, gave room for a Special Agricultural Workers (SAW) programme which brought in more workers from Mexico and the Caribbean. It attracted many first-time migrants from Mexico into the US labour force, and thus furthered more undocumented migration on the long run. The policies did not deter additional migration from within Mexico (Cornelius, Mertin, and Hollifield 1994: 34–5), simply because new migrant networks materialized and ran their course. Moreover, employer sanctions and the SAW programme furthered the fraudulent documents industry. This means that restrictive policies can be counter-balanced by measures that further more international migration in the long run. In more highly regulated welfare states consensus over immigration restrictions is usually easier to achieve because of more consensual policy-making (see Chapter 3).
It is no coincidence that—once migration flows have taken full speed—certain categories of international migrants become the object of intense debate over moral issues of acceptance into immigration states. Over the past decades two categories of international migrants have occupied the centre stage of restriction: asylum seekers and undocumented immigrants. Public debates about asylum seekers showed a clear-cut distinction between allegedly legitimate political refugees, on the one hand, and illegitimate economic refugees, on the other hand. In (p. 168 ) short, the claim of bogus asylum seekers has entered the debate over restrictions in all countries of immigration. Needless to say that such distinctions can rarely be established because the documentation of persecution is often hard to come by, given the irregular nature of refugee flights. Again, the connection of welfare and security figures prominently. This applies to demands to curtail the access of unauthorized immigrants to tax-supported public services, including education and health care. California's disputed Proposition 187 in the mid-1990s is just one prominent example.
In the expressive realm, migration can be seen as a threat to collective identity. This is not necessarily synonymous with threat to the immigration state (Huysmans 1995: 54–6). Rather, it concerns threats to religious and ethnic collectives. The nation involves the most complicated phenomena in this respect: while many countries of immigration base their collective self-conception on the basis of nationhood, the concept and understanding of nation in various states varies widely in its cultural-collective inclusiveness towards newcomers—ranging from states who regard themselves ‘nations of immigrants’, such as the former white settler colonies, to those who see themselves as ‘not a country of immigration’, such as some Central European states.
The expressive integration of ethnic, religious, and national collectives in ‘we’-groups is certainly not simply constituted by identifying the ‘other’. This would be a purely negative form of self-identification which could not integrate ‘we’-groups in the long run. Nevertheless, responses to immigration tend to touch upon unfulfiled and problematic characteristics of self-identification. The paradigmatic case is the Chinese exclusion movement in California in the late nineteenth century (Saxton 1971). Considering the relatively minor role of actual economic competition between Chinese and European gold-diggers, and later, the Chinese enclave economy consisting of grocery stores, textile and cigar-making factories, and laundry businesses, the rabid and violent exclusionism directed at the Chinese, ranging from zero immigration (1880s until the end of World War Two) to segregation in all realms of life, calls for explanations going beyond economic competition. It is remarkable that the ideological justification of Chinese exclusion, over and over again, insisted upon their alleged status as contract or even slave labour. When we imagine the importance of the ideal of the small freeholder in the American West of the nineteenth century and the experience of slavery as a negative foil, we can start to comprehend the threat this group of immigrants presented to European settlers in California (Faist 1991).
Given all these inconsistencies in enforcing complete external and (p. 169 ) internal control in immigration states, but nevertheless overall effective external and internal control measures that do not threaten the integral security and welfare functions of immigration states, it is hard to escape the conclusion that the appearance of control is one of the principal goals of restrictive policies. This provides the fertile ground for the gardeners of meta-politics—populist intermediary organizations such as right-wing anti-immigrant parties. Overall, restrictive policies are not simply the result of hard-to-measure infrastructural limits to the intake capacity of immigration states. Since the transformation of welfare states, the end of the Cold War and the musings in public debates over a ‘clash of civilizations’, the welfarization and securitization of international migration have quickened. As a consequence, the meta-politics of immigration has created its own dynamics: the more restrictive policies have become, especially towards asylum seekers and undocumented immigrants, the more this has been seen to be proof of the illegitimacy of members of these immigrant categories to be admitted. As a consequence of these ratcheting effects, immigration legislation and practice has become more restrictive, partly counteracting chain migration.
Summing it all up: The S-Shaped Migration Curve
If we represent the inverted U-curve in a cumulative manner, we get an S-shaped curve (Figure 6.4). All steps of the processes so far can be visualized in the S-shaped migration curve. This general third-order curve indicates that there are accelerative and decelerative tendencies inherent in migration processes. The pattern of this curve depends upon factors that arise from the very process of migration itself, thus
When the process of migration diffusion runs for a sufficiently long time, the frequency of migration almost always follows an S-shaped curve. After pioneer migrants have sowed the seeds, the number of desired exits tends to increase quite rapidly, in fact almost exponentially. There follows an approximately linear increase, until the increase finally slows down and is barely perceptible. These are characteristics of the cumulative number of exits; the number of migrants per time unit—the rate of migration—is the derivative of the cumulative curve, and ordinarily increases to a maximum before decreasing considerably (see also Rogers 1983: 243–5 on general third-order diffusion in social and political life). The total number of migrants grows in a logistic fashion, whereas the temporal pattern of migrant flows follows a bell, i.e. inverted U-shaped curve.
Three phases can be distinguished. The first phase is dominated by the accelerative phase and the problem of start-up costs. The adoption rate is low at the beginning but accelerates as the network expands, and the costs and risks of migration decline due to the increased opportunities of transferability and deployment of various forms of social capital. The second phase, the middle phase, is the period of positive interdependence and growth. As the system develops its own momentum, migration becomes a widely imitated behaviour, and diffusion occurs. The third and last phase is dominated by optimization and self-limiting processes. Immigration flows diminish, as the system becomes saturated. It stops when all potential migrants who wished to move abroad have done so. A turning point is reached. Those with the strongest intentions have already migrated, and those who have not yet done so had weak desires. As a corollary, the composition of migrant flows becomes more heterogeneous over time because migrant networks increasingly encompass all segments of the population of origin. The social diversity, in most cases low at the initial stages of migration, usually increases strongly during the intermediate stages.
However, most international South–North migration flows have not reached their endogenously induced turning points. Instead, public policies of the countries of immigration have curbed flows significantly. As an analysis of two important state functions—welfare and security—suggests restrictive immigration is not simply an outgrowth of immigration reaching a maximum beyond which infrastructural problems arise and the adaptive capacity of the countries of settlement is (p. 171 ) exhausted. Even more obviously, intermediary organizations such as political parties, populist and nativist groups tie immigration and immigrant adaptation to serious political and policy problems in a globalizing world. While the politics of immigration does not exhaust itself in symbolic politics—collective agents trying to define the terms and images that serve above all strategic and tactical purposes of organizational competition—immigration control and immigrant adaptation serve as a meta-issue: intermediary organizations refer to them as causes of manifold problems in a context of persistent structural unemployment, welfare state retrenchment, and a perceived loss of external and internal security because of growing cross-border traffic (Faist 1994).
The Turkish–German Example: The Unfinished Yet Terminated Migration Curve
Processes of selection and diffusion in the Turkish–German case are of particular interest because the international migration curve has not yet run its full potential course. None the less, mass migration had been practically terminated by the mid-1970s through restrictive admission policies on the German side. Therefore, this case yields unusual insights in how endogenous and exogenous factors have intertwined in shaping the migration curve.
From 1961 to 1974, about 648,000 workers migrated to Germany by passing through German recruitment offices in Turkey—in total, more than 865,000 workers went to Germany. The German authorities estimate that about two-thirds of the legal, newly recruited Turks came under official auspices (1961–71). The balance came mostly as tourists who later legalized their status (Paine 1974: 59–60). In later stages of the migration process, during the 1970s and 1980s, family reunification and marriage migration became prominent, classic instances of how potential and actual migrants moved along social and symbolic ties. In yet a third period, family reunification and marriage migration continued at lower levels, while the number of refugees increased substantially after the military coup in Turkey in 1980. From the early 1980s until the late 1990s, asylum seekers from Turkey ranked year after year second or third in terms of total numbers applying for certified refugee status in Germany (see Figures 3.4 and 3.5).
(p. 172 ) The Selective and Diffusion Functions of Turkish Migrant Networks: Hemşeri Ties
All available evidence implies that urban migrants who dominated the very first phases of international mgration (1961–3) could rely on access to recruitment offices, while later-coming rural migrants heavily depended on kinship and communal ties. In the cases of rural–international migration, these important migrant networks did not arise ex nihilo but can be seen as continuations of existing village or neighbourhood networks, based on specific ties. In the villages, these networks lowered transaction costs in controlling the behaviour of villagers regarding the cultivation of land, communal tasks, and proved instrumental to arrange marriages. Generalized reciprocity and diffuse solidarity found in hemşeri ties undergirded chain migration.
Hemşerilik—Landsmannschaften in the old German sense of regional connotation, such as Saxons, Hessians, and Swabians—refers to networks of solidary social and symbolic ties that is based on communal or regional ties. Like kinship, it can be thought of as a primordial tie (Dubetsky 1973: 345), albeit in a symbolic way. The interpretation of the range of hemşeri ties varies among the people using it. Depending on context and usage, the term denotes fellow-villagers or people from the same region, viz. province.
Chain migration within Turkey, mostly rural–urban migrations, has always progressed within hemşeri ties. In the urban destinations of these rural migrants, the gecekondus—literally ‘built overnight’—hemşeri networks provide accommodation and all sorts of material and psychological support to the migrants (Karpat 1976). In a country such as Turkey, where an extensive welfare state has been absent, hemşeri networks also offered welfare, mutual help, and support in the big cities—classical expressions of reciprocity and solidarity.
Most Turkish migrants considered friendship and hemşerilik on the same level. Interestingly, hemşeri, neighbourhood, and migration networks merge. The apogean case is when people of rural origin tend to marry hemşeris. In international migration, most migrants followed the footsteps of hemşeri pioneer migrants and depended on them: in a representative sample of recruited contract workers in the late 1960s, approximately two-thirds of the sample had either a relative or a friend in Germany before migrating. They received virtually all information from those already working in Germany. Only through the social capital benefit of information could the potential migrants compare their wages with those received for the same job in Germany and not to alternative jobs in Turkey (Aker 1972: 75–99). The reference being wage (p. 173 ) levels in Germany may explain why two-thirds of the respondent contract workers indicated that insufficient income was the cause of their emigration. However, their average income was more than that of 70% of the Turkish population. Wage differentials between Germany and Turkey were 2.1–5.6 to 1 in the period 1970–1, depending on the specific occupational group. Also, the proportion of the unemployed in this population was negligible. The unemployment rate among labour migrants prior to emigration was only 3%-–4%, compared to more than one-third of the total working age population in Turkey (Gitmez 1984). Unemployment was not the main motive. Evidence from another survey among Turkish migrants in Berlin reinforces the image of the crucial role of pioneer migrants. Approximately 28% of labour migrants in Germany found their housing with the help of relatives and friends; another quarter used brokers in travel and interpreters' offices. 62% of Turkish migrants in a mid-1970s sample reported that their relatives, hemşeris and friends found jobs for them (Akpnar and Mertens 1977: 77). In other surveys, conducted in the same period, more than 40% of Turks disclosed that they found their apartment through a relative or hemşeri (Kleff 1984: 193). Beyond the selective and adaptive functions, hemşeri ties underwent a transformation in the city, be it in Turkish gecekondus or in German immigrant colonies. These ties lost their grip on people's behaviour and turned into more instrumental devices, both in Turkish (Ayata 1989; see also Magnarella 1972) and German urban contexts (Çağlar 1994). All in all, hemşeri ties are important for migration decision-making and timing, the selection at point of origin, the occupational and residential clustering of migrants in the immigration contexts, and finally, the internal organization of migrants' groups and their relations with immigration society in general. Hemşeri or similar networks foster international migration beyond narrow kinship groups.
Threshold Effects: Pioneers and Ties
Most of the first migrants destined for Germany left with little reliable information, yet subsequent migrants had much better knowledge of conditions abroad. Among the first migrants, the recruited persons were mainly men, usually 20 to 40 years old, with at least a primary school education. They possessed relatively high occupational skills when compared to the average skill level in Turkey. Many of them had worked as civil servants such as primary schoolteachers, but also as skilled industrial workers, shopkeepers, and some had been self-employed (Abadan-Unat 1964). They hailed from the economically more developed regions of the country (Keyder and Aksu-Koç 1988: 20). In this (p. 174 ) period, the proportion of urban-based migrants was much higher than that of rural migrants, more than 80% compared to 17% from villages. Migrant networks could barely be found.
Soon, loopholes emerged, increasing chances for network recruitment. The German government abolished the rotation principle, so that workers could stay more than one year in Germany, and fashion closer links to German employers and co-workers. Moreover, responding to an increasing demand by German employers, the German labour administration allowed the direct nomination of workers by German employers via Turkish workers. The selection committees in Turkey did not draw these workers from the official waiting lists. Instead, pioneer migrants—acting as both brokers in the villages of origin during vacation and as gatekeepers in German companies—nominated these workers, mostly activating their hemşeri ties if they came from villages in rural areas or were stepwise migrants who had earlier migrated from villages to big Turkish cities. These processes stimulated kinship migration and the migration of fellow-villagers and neighbourhood friends a few years after migration had started, in the mid-1960s. This meant that most contract workers interviewed immediately before departure were step-migrants (Paine 1974: 76, 187–8). They had migrated at least once before in their life but within Turkey (Krane 1975: 161).
In the case of rural emigration, international migration constituted an integral part of stepwise migration, first from rural Turkey to bigger cities and then abroad. Especially in the 1950s migration in Turkey occurred on the internal plane, seasonal pendular migration dominating the scene (Stirling 1974: 209–10). Many migrants left their villages and later on, after having consolidated economically, fetched their family members and settled in the rapidly growing cities (Akkayan 1979: 224). Some evidence suggests that an exploration phase preceded this early phase of massive seasonal migration (Struck 1988: 217). Once they had settled in towns and cities, the exploratory movers, viz. pioneer migrants, procured jobs, accommodation, and other daily necessities for the followers (Karpat 1976: 53–5).
The Formation of Networks: Official Recruitment and Nomination of Workers
Migrants and refugees from Turkey took five major routes entering and settling in Germany and in all of them resources inherent in social and symbolic ties proved important for migration: official recruitment, pioneer migrants giving names to German employers, illegal entry and regularization of status, family reunification and marriage migration, (p. 175 ) and asylum. In each of these cases broker-centred mobilization of social capital into networks proved essential, albeit in different ways.
Depending on the social and political structure of the emigration region within Turkey and the type of migration, migrant networks emerging during official recruitment and nomination took clientelist structures. Clientelism means that one partner is clearly superior to the other in his capacity to grant goods and services. This implies reciprocal rights and duties between patron and client. It is a clear instance of tightly controlled and centralized social capital. Clientelism is most useful where the formal institutional structure of society is weak and unable to deliver a sufficiently steady supply of goods and services. In south-eastern Anatolia, where Kurdish-speaking groups concentrate, relatively isolated groups who have poor access to power, prestige, or wealth establish a fertile ground for clientelism. The patron, often a wealthy and powerful ağa, provides economic aid and protection against both the legal and illegal executions of authority and the client in turn shows respect and loyalty to his patron (Kudat-Sertel 1972).
Emerging clientelist structures can be seen when patrons helped new arrivals to acquire residence and work permits, housing, and help with the tax authorities. We have seen earlier that the higher the stocks of social capital, the lower the transaction costs. Also, lower transaction costs usually favour the reproduction of existing organizational models in the immigration country. Established stocks of social capital and concomitant forms of groups and organization generally have the advantage over innovative solutions in the short term because transaction costs of devising, perfecting, and installing new solutions coming up during migration and post-migration processes exceed the costs of maintaining old ways. One interpretation would be that Turkish workers in Germany who were confronted with language barriers and the problem of getting access to housing and work, would use patrons—such as Turkish businessmen—as intermediaries (Kleff 1984: 244–60).
There are historical precedents: in the Italian padrone system, the powerholders had tasks such as financing migration, providing employment, and numerous other services which isolated new arrivals from the American society and kept them dependent. A padrone might have acted as a banker, landlord, foreman, scribe, interpreter, legal adviser, or ward boss. His clients bought continuing protection from a public figure who was somewhat subject to community pressure and dependent on its good will (Eisenstadt and Roniger 1984: 271). In Turkish gecekondus, political party activists have helped the new rural-born migrants to settle down, to deal with the authorities, to find employment, and even to marry. A particularly powerful weapon in the (p. 176 ) hands of the party controlling the municipal government has been to tolerate, or even legalize, the illegal squatter houses of the new urban migrants, or to demolish them by strictly implementing the laws. However, unlike the Italian padrones in the USA around the turn of the century or strongholders in Turkish gecekondus, the Turkish equivalents in Germany did not occupy key positions in the local German establishment. These transplanted networks were thus under constant pressure for readaptation in the country of destination.
Within these clientelist structures, the first two patterns mentioned above—official recruitment and nomination of workers—proved essential for the formation of migrant networks.
(1) Official Recruitment
During the peak of migration, in the late 1960s and early 1970s, the destination side, with the help of Turkish authorities, recruited most workers ‘anonymously’. The waiting lists got longer in the mid-1960s as a result of relative frustration—it is estimated that more than one million workers had waiting times up to ten years in 1970 (Paine 1974: 67). The point of departure was the original policy criteria of the German government. German employers looking for Turkish workers applied to the German Employment Service (Bundesanstalt für Arbeit, in short: BA) for a specific number and type of worker. When no privileged workers—German citizens and EU nationals—could be found, the BA transmitted the employers' requests to the Turkish Employment Service (TES). TES maintained lists of workers wishing to go abroad for BA representatives in cities such as Istanbul, Izmir, and Ankara. At this stage ties to brokers did not yet prove instrumental and migrant networks were not widespread. This changed when TES altered as a reaction to the expansion of the lists from 500,000 in 1965 when 55,000 migrants were sent abroad, to over 1 million in 1970 when the TES sent 130,000 abroad.
Specifically, the TES decided to prefer applicants from the less developed regions of the country, members of Village Development Co-operatives, and persons from officially designated disaster areas. To achieve this end, they divided Turkey into three socio-economic development categories—developed, developing, and less developed. Persons from less developed regions formally had priority on the waiting lists. In addition, a rotation system supposedly assured that each year another province received priority. Natural disaster areas received special quota (Abadan-Unat 1974: 369–70). Turks interested in going abroad needed a residence certificate to register with TES. A local leader, a muhtar, could issue such a certificate, provided that the person (p. 177 ) had maintained residence for at least six months. Turks with relatives and friends in priority areas sometimes moved to the area just to obtain a second residence permit and thus to augment their chance of being among the chosen few. In these cases the muhtars served as brokers. Exchange-based obligations and reciprocal ties of all kinds proved instrumental. None the less, with this turbo chaser, migrant networks evolved in rural areas, based on kinship and communal ties. Because of the late start, workers from the poorest regions in Turkey, especially from villages and small towns, were least likely to have their family with them abroad when Germany stopped recruitment in late 1973 (Wilpert 1992: 179). They had the greatest potential for future entry into Germany because many family members still resided in Turkey. As a result of these policies, the percentage of unskilled workers with minimum education rose sharply, and emigration covered rural areas as well, such as central and south-east Anatolia (Gökdere 1978: 178).
(2) Nomination of Workers
When the German government abolished the rotation principle and allowed nomination of workers by German employers, networks surged. Instead of simply requesting a certain number of Turkish workers, an employer could name or nominate a particular Turkish worker to fill a vacant slot. Since these nominated workers obtained priority from the TES to emigrate, potential migrants asked their relatives already abroad to locate vacant jobs and persuade employers to nominate them. Thus, not only brokers in Turkey but also gatekeepers in Germany, mostly pioneer migrants, proved essential in providing help for later migrants. In the course of these transactions migrant networks evolved. In these cases mostly ties of reciprocity and solidarity were important. It is estimated that about one-third of the legally employed Turkish migrants, who came to Germany between 1967 and 1971, arrived through these channels (Gitmez 1989).
Chain Migration as a Diffusion Process: Tourism and Regularization
The route of entry via tourism, illegal stay, and regularization depended a lot on full-fledged migrant networks. The longer migration proceeded, forms other than recruitment via waiting lists and official referral of gatekeepers also evolved—tourism, illegal stay in Germany and regularization of the status in Turkey, and subsequent return to Germany. This avenue required substantial information, embedded in reciprocal and solidary ties of cooperative hemşeri networks. It worked in the following way: potential Turkish migrants obtained tourist passports, (p. 178 ) emigrated to seek employment, and then regularized their status or worked without authorization. They needed relatives or fellow-villagers to protect their illegal stay in Germany after their tourist visas had expired. When kin members helped them they depended on specific reciprocity or focused solidarity; when fellow-villagers extended a helping hand, they partook in generalized reciprocity or diffuse solidarity. The German Länder with the highest concentration of and demand for foreign labour—Hesse, North Rhine-Westphalia, and Baden-Württemberg—enacted legislation which permitted illegal alien Turks to return to Turkey, get proper work and residence documents, and then re-enter Germany. The number and share of such unofficial workers in the 1960s and 1970s is not known by the nature of the process, but one study reported that over 40% of a sample of migrants abroad in the mid-1970s had originally entered as tourists. Observers believe that about 20% of all Turkish migrants until the late 1970s went illegally, and that many of them eventually regularized their status (Martin 1991: 29).
The convertibility of social capital into other resources serves to explain why especially those migrants engaged in migrant networks who had access to only few resources in terms of formal criteria such as education, professional prestige, and money. Gendered variances illustrate this tendency.
Distinguishing along gender lines, we encounter three forms of migration: male, female, and joint migration in couples. In a mid-1980s representative sample drawn from contract workers in the Federal Republic, we find the following composition: male pioneer migrants (76.4%), female pioneer migrants (13.1%), and family migration (10.4%). Most family migrants actually came before the termination of recruitment in 1973 (Özel and Nauck 1987). In all these cases, extensive kinship networks played a crucial role. In most instances, the availability of a tight and extensive kinship system favoured male pioneer migrants. Moreover, the majority of male pioneer migrants came from very religious families, whereas most female pioneer migrants and joint couples came from nuclear and incomplete families. Interestingly, kinship lineage made a difference: if the kinship system of the wife was operative, we find more often female pioneer migrants, and vice versa. Each of these three types had its own climax: men in 1964–5 and 1968–9, women and couples in 1972–3.
Female pioneer migrants came overproportionately from urban centres in less developed provinces, while there is no stable correlation between modernity of province and male pioneer migrants. Most joint migrating couples departed from the urban metropoles. Furthermore, (p. 179 ) this data not only suggests that there was a continuous and substantial movement of pioneer migrant women, often overlooked. We also learn that most pioneer migrants, either men or women, were unmarried. This is an expected result when considered in the broader life course of persons: for young adults, social and symbolic ties with kinship groups tend to be somewhat weaker before the establishment of a family than thereafter.
As the end of recruitment approached, the pattern of diffusion had changed considerably because of the long waiting lists: conservative Turkish men in rural areas accepted migration of women and joint migration to a much higher degree than before (Abadan-Unat 1985b: 208). However, the numerical impact of this pattern should not be overestimated. Out of all female pioneer migrants, only a minority of women engaged in migration along this pattern.
Diffusion processes arose because social capital could substitute economic capital and individual human resources once pioneer migrants were established; this differed among various categories of movers. Pioneer migrants could act as brokers and gatekeepers and lower the amount of economic and human capital necessary for other potential migrants. In her study of two Anatolian villages Engelbrektsson (1978; see Chapter 5) found that from one of the villages—Alihan—migration mostly occurred after the end of recruitment: only one migrant left Alihan to go abroad during the official recruitment phase before 1973; all others came with the help of migrant gatekeepers. In 11 out of 13 cases, later migrants relied on support by pioneer migrants, hemşeri ties. For example, pioneer migrants provided the travel money necessary. Other mechanisms chosen included family reunification and marriage arrangements with Swedish women, or obtaining an employer's certification (ibid. 191–8, 221). At first sight, this somewhat contradicts the correct observation that mostly middle-income strata migrated. Yet, as the migration processes and networks matured, this allowed even those without the necessary economic means to migrate. Again, this shows that village and hemşeri network ties unfold dynamics not simply determined by larger structures and micro-motives. Here, fellow-villagers both abroad and at home were willing and felt a moral duty to grant practical assistance to those who opted for emigration. In this context, ordered by norms of reciprocity and traditions of solidarity and mutual assistance, personal lack of finances did not necessarily pose a great obstacle for emigration (ibid. 287).
A totally different pattern emerges among emigrants from the other village studied, Yeniköy: international migration did not diffuse. Pioneer migrants exclusively supported members of the their own kinship group. Therefore, non-members of the (p. 180 ) kinship group could not possibly convert the benefits from reciprocity and solidarity into economic capital. Not surprisingly, the number of migrants from non-kinship groups remained small (ibid. 109). The crucial difference in hemşeri ties was that the Alihan migrants considered their group identity abroad primarily in terms of their being members of the same Turkish village, while the Yeniköy migrants in Sweden reckoned their collective identity primarily in terms of being members of the same kinship unit (ibid. 194). The generalized reciprocity and communal solidarity of Alihan movers and stayers starkly contrasted to the specific reciprocity and family or household solidarity of their Yeniköy counterparts.
The diffusion of migration had implications for the pattern of domestic–international migration, too. Stepwise migration increased the likelihood of international migration. In the first years of migration, representative surveys indicate through a comparison of place of birth and place of residence before departure that cities which received most internal migration were also the ones overrepresented in international migration. From 1963 until 1973, the majority of international migrants came from the cities of Istanbul, Ankara, and Izmir (Leopold 1978: 150).
Once pioneer migrants had fostered intensive contacts, step-migration proved unnecessary. Nationwide data on Turkey confirms this claim: in the second half of the 1960s, the proportion of those who directly migrated from rural areas to urban places in Germany without intermediary steps in Turkey increased (Paine 1974: 75–6, 86). However, the proportion of those who directly moved from rural areas to Germany was lower than the proportion who engaged in step-migration before moving to Germany. This can be attributed to the termination of recruitment in 1973.
All of the evidence adduced points to occupational and residential clustering of Turkish immigrants in Germany. Since the evidence is well known by now, suffice it to say that these were the end points of migrant and migration networks. When pioneer migrants helped latecomers to gain admission to Germany, the brokers also catered for jobs. This is true for housing, too, hence the high degree of residential clustering in certain regions and cities in Germany (Jones 1990: 27), and within cities on the neighbourhood level (Esser, Hill, and van Oepen 1983; Wulf 1985).
Cumulative Mobility: The Climax of Endogenous Processes
The principle of cumulative mobility worked as follows in the Turkish case (1961–74): domestic and international migration originated in profound transformations of Turkish industry and agriculture. In the (p. 181 ) rural areas land was unequally distributed with vast differences between great landholders, on the one hand, and inefficient tenants and small holders engaging in subsistence economy, on the other. The land cultivated became smaller and smaller. Moreover, the number of landless peasants who had to work as labourers increased. Only great holdings had the means to engage in large-scale mechanization, starting in the late 1940s. They could increase productivity, purchasing mechanical power and yields. When the mechanization of agriculture picked up in Turkey during the 1950s, this process displaced peasants and agricultural workers. In addition, expanding great holders bought the land of tenants. All these processes speeded up the concentration of land. Domestic migration to big cities followed; built upon networks developed during seasonal migration since the 1940s (for a case study, see Struck 1984). Many rural villagers responded by migrating to big cities, processes that had gained momentum during the 1950s. Subsequently, in cities such as Istanbul and Izmir, rapid urbanization in the 1950s proceeded much faster than industrialization. In the industrial sectors import substitution ruled supreme and new industries concentrated in Western Turkey around Istanbul and Izmir. Yet its absorption capacity for the rural population proved to be much too low—in spite of a rapidly expanding service sector. In other parts of Turkey, mainly in the south-east, no industry developed (Leopold 1978: 154–63).
In the early 1960s when recruitment from Germany started, no new jobs were available in agriculture and employment elsewhere was scarce. Among those especially hard hit were small holders who were often highly indebted. Only when network diffusion reached deep into villages, did international chain migration emerge directly out of rural villages on a large scale. Before, primarily industrial workers and stepwise migrants from urban areas had participated. Increasingly, the income of those who neither engaged in internal nor international migration came to depend upon extra-village occupation and other sources, like the remittances of migrants.
For the regions of origin diverse effects can be observed, especially in the rural countryside. In the Turkish case, most of these feedback effects worked to increase domestic and international migration. First, in small villages, some return migrants invested in mechanization and used more intensive production methods so that other villagers finally gave up farming (Steinbach 1993: 522). By migrating relatively sooner than others in their system, pioneer migrants and early migrants achieved windfall profits, thereby tending to widen the socio-economic gap between these earlier adopting categories versus laggards. Thus, the early movers got richer, and the later migrants' economic profit was (p. 182 ) comparatively smaller. As migration became widespread, more households were able to invest in capital-intensive production methods and purchased tractors and farm machinery, partly because actual and potential migrants were reluctant to do farm work in the traditional manner. This prompted rapid agricultural mechanization in Turkey during the 1960s and 1970s (Martin 1991: 30). Nevertheless, many people withdrew from cultivation. In this way, migration exacerbated the falling demand for agricultural labour and accelerated the shift to commercial agriculture. In turn, it also increased permanent departure for Turkish and German cities. Yet these processes can only be observed in those cases when brokers of all kinds—pioneer migrants and, later on, kin and hemşeri in Germany—not only accumulated stocks of economic capital but also engaged in utilizing old ties back to their home communities and forged new ones in Germany to employers, bureaucrats, and landlords. The higher the density and strength of hemşeri ties, the higher the rate of emigration from particular communities and the higher the degree of residential clustering.
Most studies on the effects of remittances and return migration on the emigration communities agree that migrant income allowed households to improve their material standards of living dramatically but that earnings were spent in relatively unproductive ways. Remittances and savings generally went to current consumption rather than investment, such as the purchase and repair of homes and the acquisition of consumer goods. The little investment that occurred went to small commercial activities that generated little employment (Gitmez 1989). Nevertheless, migrants from rural areas did indeed buy farmland when they were able to do so. However, return migrants who had acquired land, often held it fallow. This was because some of them preferred to continue migrating or they retained this land as a source of security and prestige in the community. When farmland thus became even scarcer for those potential migrants who had remained, even more moved abroad.
Self-feeding effects operated relatively independent of economic factors. The example of successful migrants sometimes encouraged others to migrate as well. It created a sort of bandwagon effect, an example of positive feedback causation. It also increased in areas where farmland did not get scarcer. The interception and termination of recruitment then accelerated the flow of dependents from Turkey into Germany. The latter was a clear effect of the maturing Turkish–German migration system: network variables became an increasingly important reason for the attraction of immigrants, while the impact of economic and political factors in Germany declined (Waldorf 1996).
(p. 183 ) No Turkish-German Culture of Migration
In the Turkish–German case no veritable culture of migration emerged. Because of effective German admission policies that did not stop but curbed flows, exogenous opportunities for lifestyle migration did not exist for the hoi polloi. None the less, family reunification speeded up in later stages of the migration process, with an additional push after recruitment ended.
When family reunification had largely run its course and supply was exhausted in the mid-1980s, marriage migration became relatively more important. Often, this type of migration necessitated more generalized and diffuse ties. First-generation immigrants tried to arrange for the marriage of their off spring, their daughters in particular. Indeed, as late as the early 1990s, many first-generation Turkish immigrants all over Europe still preferred to marry their children to kin over non-kin. Among other things, they expected a relative to behave somewhat more loyal than an unrelated son- or daughter-in-law. Also, kin in rural areas of Turkey often anticipated their extended and remote family members to search for brides and bridegrooms within the wide kinship group first. A refusal on the part of immigrants in Europe was taken as evidence that they do no longer value kinship ties with home. In a study of a Dutch city in the late 1980s, for example, it turned out that half of the brides and bridegrooms of the second generation came from the kinship group in Turkey (Böcker 1994).
Understandably, the pressure of kin in Turkey to obtain a daughter-in-law in Europe increased after recruitment ended. In order to live and work in Europe, marrying the daughter of a migrant was one of the most attractive avenues for young men in their twenties and thirties, the major group of potential migrants. And this was not surprising because the demand for migration did not decrease during the 1980s. If at all, it probably increased in the 1980s as compared to the decade before because structural conditions did not change, this age cohort has numerically become stronger, and migrant networks still existed. However, transaction costs have gone up: over time, some second-generation women have grown more reluctant to be married to remote kin from Turkey. They increasingly want to choose their marriage partner on their own (Schmidt-Koddenberg 1989). Moreover, transnationally operative kinship and hemşeri ties grew weaker over the years.
While migration for family reunification can be interpreted as a straight continuation of earlier trends of transnational reciprocity, and marriage migration from Turkey to Germany as a result of continued symbolic and social ties to Turkey, the migration of asylum (p. 184 ) seekers bears similarities to these two kinds of migration but is yet different. It is similar in that certain political conditions, such as Turkey's unacknowledged civil war and Germany's comparatively liberal asylum laws, have to be seen in the context of mature migrant networks. Many of the asylum seekers already had relatives and friends in Germany. Among the two million Turkish immigrants in Germany in the 1980s, a fourth came from south-eastern Anatolia and can be considered from an ethnic point of view as Kurds. Network mechanisms provided access for activists who were persecuted, targets and victims of persecution, and those who used Germany's asylum regulations to improve their economic fortunes.
Decelerating Migration Flows: Effective Termination of Mass Migration
We can only speculate about the magnitude of migration flows from Turkey to Germany had they not been interrupted by the termination of recruitment and tightened asylum regulations. Based on all other known cases, it is safe to say that in the absence of restraints, migration would have continued on an even higher level than before. It could have been expected to continue snowballing because no endogenously derived turning point can be discerned. Turkish internal migration has still been substantial all through the 1990s. In the late 1980s, the country's population increased by more than 1 million annually, but the urban population grew about 1.5 million each year. This indicates that each year at least 400,000 rural Turks migrated annually to cities from 1980 until 1995. These internal migrants included both families and young men, the majority of which could not be absorbed into the industrial and service sectors (İçduygu 1998: 28). Given the existence of structural conditions conducive to migration, the relevant political and trade ties to Germany, and the preferred status as an associated country to the EU, but above all the mature and self-feeding nature of international migration observed in the Turkish case, it seems safe to speculate that international migration would have increased on an even larger scale than it did during the 1980s and 1990s. However, with the exception of refugee migration, all flows boiled down to a trickle when compared to Turkish-German flows from 1961 until 1973.
The three forms of migration have varying prospects. During the 1980s and 1990s, asylum migration has been of about the same size as family reunification. While it would be presumptuous to predict the future course of asylum migration, family reunification has tended to fall for obvious reasons. The same is true for marriage migration, albeit (p. 185 ) for somewhat more interesting causes. We have seen that many first-generation parents prefer arranged marriages for their children. However, this preference has slowly changed over the past years. Certainly, the children themselves are nowadays more likely to look for a spouse in Germany. Parent–children relations and thus reciprocal kinship ties have changed considerably from the early migrants to their children about 40 years after the first Turkish immigrants arrived in Germany.
In sum, despite the fact that the Turkish–German migration curve has not yet run its endogenously-patterned course, a decelerating period has been reached before the S-shaped curve had run its course and reached a decelerating stage under its own energy, mainly because of restrictionist policies on the German side (Figure 6.5).
German Immigration Policies: Effective Termination of Labour Immigration
The Turkish case and much of contemporary South–North migration differs from the earlier transatlantic labour migration system in the nineteenth and early twentieth century that brought millions of Europeans to the Americas and Australia. Transatlantic mass migration was more migrant-friendly than the international migrations after World War One. The former took place in a world that allowed relatively free migration because white settler colonies looked for population to settle land deemed readily available and to power the production lines of emerging industries. However, World War One and the period after signalled a growing trend of protectionist food, trade, and labour policies. In many countries in Western Europe and in North America welfare state measures called for tighter protection of labour markets. Nowadays, labour recruitment and refugee reception, on the one hand, (p. 186 ) and the stop of recruitment and curtailment of refugee inflows, on the other, occur on the basis of occupied land and a high level of social rights of the resident population when compared to the late nineteenth century. The scarcity of resources, potential and actual competition between the newcomers and the settled population, and meta-politics have contributed to a politicization of immigration. The main function of meta-politics has been to strengthen restrictive immigration policies. The impetus for restrictive immigration policies and thus the deceleration of international migration flows caused by exogenous macro-factors can be traced in three realms of state integration in welfare and security: functional cost-benefit calculations, moral questions of justice, and expressive considerations of solidarity and identity.
Viewed in the functional realm of societal integration, the German case vividly illustrates changing cost-benefit perceptions. The end of contract worker recruitment in 1973 marked a decisive change in that labour migrants thereafter had to use other entry channels—including family reunion, political asylum, illegal immigration and, since the late 1980s, increasingly seasonal and posted worker arrangements. Because some contract workers stayed on and others joined, they developed preferences and expectations which in turn made some of them remain in Germany for good. The characteristics relevant for social policies have come to resemble more and more those of German citizens. For example, the percentage of women and young people among the immigrant population has been rising, the labour force participation rate of men has decreased but has increased in the case of women, and employment in services and self-employment has been slowly rising. As the demographic, work and public services use patterns of former labour migrants and their children have adapted to those of German citizens, cost-benefit perceptions have been changing on the part of the latter.
Yet most empirical scholarship on costs-benefits associated with immigrant integration is inconclusive—making it a fertile field for meta-politics. Studies have consistently pointed out the overall beneficial aspects of immigrant employment for the native German population, such as increased mobility of German blue-collar workers into white-collar employment as a result of guestworker-recruitment and employment in the 1960s and 1970s (Heckmann 1981: 166). Numerous studies have confirmed that foreigners are net contributors to social insurance schemes such as pension funds. Moreover, some results suggest that during the late 1980s and early 1990s non-citizens have received much less in tax money than they contributed (Barabas 1992). A position favourable to immigration emphasizes the economic-demographic benefits immigrants provide for the host country. In regard to social (p. 187 ) security, it could be argued that declining birth rates in advanced industrial countries necessitate new immigration in order to pay for the pensions of retirees. Two arguments recur again and again: first, immigrants are a net benefit for public budgets because they pay more in taxes than they get in return in the form of services and social insurance. A typical calculation goes like this: In 1989 immigrant workers paid 12.8 billion Marks (8%) into pension funds, while immigrants only received pensions amounting to 3.7 billion (2%). Thus, immigrants pay about the amount that corresponds to their share in the total population, while they benefit much less (Rehfeld 1991). Second, contract workers and, later, ethnic Germans have stimulated the economy, especially in buying consumer goods, because they tend to be younger than the German population. A plausible conclusion derived from such arguments is that developed countries such as Germany need not fewer but perhaps even more immigrants (Hof 1993).
Costs to German taxpayers have been much more visible in the case of asylum seekers than contract workers. But even here it is hard to find conclusive evidence on costs and benefits. Research on labour market participation of recognized refugees and asylum seekers is too scarce to allow firm conclusions. Most of the debate focused on means-tested social assistance (Sozialhilfe) granted to asylum seekers. At first sight, if we look at total numbers, the percentage of non-citizens compared to German citizens receiving public assistance looks impressive. Although foreigners made up only 8.3% of the total population in 1991, their percentage of recipients of one type of means-tested public assistance, Aid to Subsistence (Hilfe zum Lebensunterhalt), was close to 15%. By comparison, about 5% of the German population received this type of assistance (Statistisches Bundesamt 1993: 513). Most of the non-citizens drawing assistance were asylum seekers. An unprecedented number of asylum seekers entered Germany during the late 1980s and early 1990s; in 1992, the number had reached a climax with 438,000 before it started receding. Therefore, the financial burden for the local communities who are responsible for housing and other services to asylum seekers rose drastically. Impressive as these numbers look at first sight, they exaggerate the dependence of non-citizens on welfare. Most of the non-citizen recipients were asylum seekers and since the mid-1990s ethnic Germans, not former contract workers. Also, in the early 1990s the government drastically cut benefits that accrued to asylum seekers. Since 1993 asylum seekers do not even have access to regular means-tested public assistance. They are only entitled to a limited set of social assistance benefits, such as a reduced form of Aid to Subsistence (p. 188 ) according to the Asylbewerberleistungsgesetz (Deutsches Ausländerrecht 1998: 263–9).
The second realm pertains to moral questions of social justice. An immigrant can be entitled to full social rights (denizen) without being a full member of the polity (citizen). This effectively means a decoupling of rights and duties, on the one hand, and nationality, on the other. This situation raises conflicts in the ethical and moral dimensions of welfare state integration. First, the dominant perception is that although there are universal human rights, social rights are ultimately tied to membership in the national polity. The increasing presence of non-citizens and their access to social rights thus raises the question of welfare state solidarity, of who is to be included. For example, in nationalist-populist views the foremost duty of the welfare state is to guarantee the decent livelihood of its own citizens before rights or claims are granted to non-members or those who aspire to become members. In effect, immigration and its consequences then become a question of social justice. In this view, migrants will be allowed to stay in the host country for only a short period of time. Accordingly, immigrants could claim only the most basic civil rights and certainly no political rights. Second, this immediately brings up another question: Whose rights are recognized?
The third realm concerns expressive considerations—collective representations—and involves the question of solidarity with whom? Debates over the effects of immigration on wages, unemployment, social security, education, and social services have increasingly been cast in ‘we’ vs. ‘them’ distinctions which fall along ethnic lines, albeit to varying degrees. There is a recurring argument that it is ‘we’, the taxpayers who pay for ‘them’, the immigrants of developing countries that are allegedly flooding welfare states in the West. Perceptions of increasing costs involved in recognizing social rights of newcomers tend to be associated with certain immigrant groups and not others. In particular, refugees from non-European developing countries seem to be seen as contributing to rising costs. ‘We’ vs. ‘them’ distinctions have long been a feature of the German national state and discrimination based on race a main feature of twentieth-century politics—of course, most explicitly so during the Nazi regime. However, political conflicts in the post-war welfare state were not demarcated along ethnic lines. Current debates thus have an expressive dimension that has been missing in post-war German welfare state politics.
What the national welfare state offers to its members, is not only a bundle of rights. The national welfare state also offers a sense of belonging. This belonging is located at the level of the nation-state. (p. 189 ) While immigrants may seek to benefit and usually do at least partly so from individual rights of citizenship, they may be excluded from this second meaning of citizenship. The main point is that the degree of perceived social distance in terms of nationality, religion, or ethnicity does serve as a criterion for informally evaluating the claims of immigrants for social rights and provisions. It is evident that in this sense ethical considerations have started to play an important role in welfare state politics once cost-benefit perceptions changed in the course of labour migrant employment, settlement, and the arrival of asylum seekers.
There is a probably changing rank ordering of national or ethnic groups which compete for admission, work, and settlement in the receiving welfare states. The most important question concerns the changing perceptions of social distance towards immigrant groups. They may be reinforced or weakened over time. They heavily depend upon the perception of costs and benefits of incorporating a particular group. Of particular importance are processes of the reproduction of social distance. In some cases, for example, social distance towards non-European labour migrants has been reproduced since the 1960s (e.g. Turks); in the case of workers from the southern European periphery (e.g. Italians), high degrees of social distance have been reduced (D. Fuchs, Gerhards, and Roller 1993). One of the main questions for empirical analysis is how perceptions of social distance change depending on different immigration policies. The reproduction of the rank ordering could depend upon the functions the receiving countries ascribe to migrants and the evolving ethnic and national solidarities of immigrant and immigration communities. For example, workers performing temporary work may come from countries or belong to groups that are placed very low on a scale of ethnic-national preference. At the opposite end of the spectrum are policies that give explicit preference to groups whose collective characteristics are seen, at least initially, as similar to the dominant group in the immigration country. This is the case with respect to immigrants considered to be permanent, like ethnic Germans from Eastern Europe.
In addition to these endogenous effects arising from immigration and integration, the changing scale of welfare state redistribution has contributed to a politicization of immigrant social rights. In welfare states citizens are not only integrated through political rights but also through social rights and benefits. Social rights entail claims to scarce resources such as social security, education, and social services. Political inclusion into the welfare state depends on the extent to which the distributional activities of the welfare state are considered to be legitimate and are (p. 190 ) accepted by the populace (Bommes and Halfmann 1998). In times of sluggish economic growth or recession, distributional struggles about scarce redistributional resources intensify. Immigration can serve as a rallying point around which the issues of redistribution and political inclusion in the welfare state crystallize. Indeed, the experience of the settlement of labour migrants and refugees, and very high levels of immigration give more opportunities for political actors to voice dissatisfaction and demands with welfare state politics and policies because the effects of immigration on natives in areas such as work, housing, social services, education, and health care are far from clear-cut.
The security function of states and its ramifications for restrictive immigration policies has recently nicely converged in the ‘Fortress Europe’ metaphor. It summarizes the security fears of politicians, policy-makers, and the populace. Generally, there is a widespread argument that border controls in the EU before abolishing many frontiers were effective. But in a borderless internal market Europe will be swamped by migrants unless tough measures are taken. In this view, functionally, the link between immigration and illegal activities such as crime—including smuggling of human cargo—is a prime question (for evidence on the activities of the transnational mafia in exploiting unauthorized migrants, see Thamm and Freiberg 1998). In the normative realm, the possibility that immigrants—and here especially asylum seekers and unauthorized immigrants—exploit social services and welfare is seen as a grave threat. Finally, in the expressive realm, the discussion focuses on the link between immigration and instability caused by xenophobia and racism which presumably is on the rise because of high numbers of immigrants. In the absence of direct borders with the South, the EU borders—especially towards the Mediterranean—have become important for German policy.
The loss of border controls between selected EU Member States, for example through the Schengen agreement, has led law enforcement authorities to partially suspend their traditional reliance on the filter function of border controls. As compensation, EU authorities determined to introduce a series of compensatory measures capable of minimizing the risk of a growing ‘internal security gap’. Among other things, we can observe a plethoric development of working groups in the EU and on the European Council. For instance, the so-called Trevi group originally only covered terrorism. Later on it added concern over illegal immigration. This change has been symptomatic for the linkage between diverse aspects of a security problem internal to the EU: drugs, terrorism, and migration connect to each other. However, it is doubtful whether these many uncoordinated working groups have contributed to (p. 191 ) effectively coordinated policing of borders and internal territories of Member States.
In the absence of a common European migration policy, national and supranational institutions have engaged mostly in ad hoc security strategies. There has been no transfer of authority to regulate labour migration to the supranational level. Intergovernmentalism continues to be the dominant mode of coordination, as Member States have remained the major players in this game (Collinson 1994). Nevertheless, supranational institutions have made some efforts at a common policy, which can be perceived as an important agenda-setting. The European Community measures regarding immigration are included in each of the three ‘pillars’ of the Maastricht Treaty (1991). The first pillar, the treaty provisions concerning Community matters, include an article (100c) stating that the Council of Ministers ‘shall determine the third countries whose nationals must be in possession of a visa when crossing the external borders of the member states’. The second pillar (foreign and security issues) touches, albeit indirectly, upon migration questions because the document relates security issues to the spatial movement of persons. The third pillar (intergovernmental cooperation in justice and home affairs) lists the areas that ‘member states shall regard…as matters of common interest’. These matters include asylum policy, rules governing the crossing of external borders, immigration policy; policy regarding nationals of third countries, more specifically conditions of entry and movement, residence, family reunion and access to employment; and unauthorized immigration. The development of a common migration policy in Europe concentrates on three main policy lines: (1) action on the causes of migration pressure; (2) action on controlling immigration flows; and (3) action to strengthen the integration policies for legal immigrants (EC Communication 1994: lb–lc). Overall, then, policy-makers on the nation-state level have accepted the free movement of labour and services under the umbrella of enhanced economic liberalism and eased the barriers of migration to EU citizen-workers. However, nation-states have not lost their sovereignty concerning the conditions of admission and employment of third country people.
To the extent that within the EU nation-state borders have diminished in importance because of agreements such as Schengen, the emphasis on internal controls has grown. For example, police forces have significantly stepped up capacities for internal control, such as increased checks in border regions: German border security police forces can now patrol a larger section of territory near the borders. Overall, there has been a slow shift from border and community policing to (p. 192 ) policing using more sophisticated criminal intelligence services. Other internal controls include measures such as intensified inter-agency cooperation and the exchange of information and the introduction of identity cards for non-citizen and citizen workers.
As to normative issues of justice and crime, no hard data exists to show that migrants threaten the lawful fabric of immigration states. Indeed, the two main categories of migrants charged with crime— asylum seekers and illegal immigrants—are unlikely candidates. Indeed, it is more feasible to assume that illegal immigrants who seek illicit employment and/or residence are less likely to commit crimes, so that they can avoid contact with the police (Bovenkerk 1992: 89, cited in den Boer 1995: 102). However, we would expect that members of these two categories are more likely to engage in offences only foreigners can commit, such as labouring without working permits, or working without a residence permit. And this is indeed what can be gleaned from the official statistics in Germany in the 1970s and 1980s. While crime rates of non-citizens have gone up comparing German citizens and foreigners between 1970 and 1989, there is a remarkable trend. More than 80% of the crimes committed by foreigners concerned asylum and residence permit offences. Certainly, German citizens could not commit those crimes. Moreover, when controlling for socio-economic status, the crime rate of foreigners is not higher than among Germans (Piehler 1991).
Concerning the expressive realm of security, there are two recurrent arguments in public debates, the control-adaptation thesis and the control-racism proposition. First, one justification for stricter border controls has emerged in the debate over resident extracommunitari in the EU. This policy proposition—forcefully argued not only by nation-state governments but also by the EU Commission and the European Parliament—is that strict oversight of new immigration is necessary in order to advance and secure the integration of these (foreign) immigrant populations and avoid the formation of ethnic ghettos. in short, the proponents claim that a successful immigrant adaptation, viz. integration policy, depends on strict entry control. Adaptation can be most easily defined in terms of labour market integration, using statistical indicators such as unemployment rates, income, and labour force participation. Full adaptation, then, would be said to have occurred if immigrants were to have rates of labour force participation, unemployment levels, and average incomes close to those of natives. According to these indicators, however, it appears that immigrant populations in the EU are disadvantaged, albeit with significant cross-country variations and differences between various immigrant groups (Werner 1994).
(p. 193 ) Second, there is the claim of a causal link between tight border and internal controls, on the one hand, and the absence or low rate of racist and xenophobic attacks, on the other hand. However, this link cannot be established empirically. For example, in the early 1990s, racist violence in Britain was at levels similar to Germany, although the UK had a much tighter regime of immigration control (Thränhardt 1993: 338).
All these selected examples suggest that international migration does not pose a grave security threat to the immigration states. Nevertheless, the issues are real in the sense of average increases to be expected in any kind of broder-crossing transactions—but they also become exaggerated in the meta-politics linking international migration to a host of domestic problems such as rising crime rates and a wider array of transnational exchanges crossing nation-state borders.
Conclusion: International Migration as Metamorphosis
Once begun, migration eventually develops an infrastructure that enables movement on a mass basis. Migrants transfer local assets internationally by means of networks of cooperation. Over time, the number of social and symbolic ties between areas of origin and destination grows, establishing migrant networks. All of them progressively reduce the risks and costs of international movement. People from the same kinship group and community are enmeshed in a web of reciprocal obligations upon which new migrants draw to enter and find shelter and work in the immigration society.
The application of the principle of positive feedback cumulative causation on the meso level has increased our understanding of the first puzzle of simultaneously and relatively little international migration yet massive chain migration within certain regions. Increased opportunities for migration abroad immediately raise expectations among potential migrants. But a change in the level of information, control, and access to other resources is achieved more slowly, and its effects on the other factors are delayed, so that there is a lag in the whole process of cumulation. Of special importance for these metamorphic developments is generalized reciprocity and diffuse solidarity. They are responsible for the rapid diffusion effects in modern international migration. Regarding future empirical studies, this evidence calls for a closer consideration of ties which reach beyond ties of narrow kinship groups.
These ties and corresponding networks eventually crystallize as new transnational linkages. It is a sort of ratcheting effect because one of the prerequisites for migration is usually pre-existing linkages between (p. 194 ) emigration and immigration states. Migrant networks emerge, maintained and sustained by ongoing processes of return whereby migrants and sometimes even refugees move back and forth between countries of origin and destination. Even settled migrants remigrate back to the emigration countries. While nation-states are not pusillanimous when it comes to the regulation of migration, the very policies that are meant to restrict additional migration often turn out to stimulate new forms of migration, although they effectively curb older patterns. All of this suggests that it is useful to analytically separate the endogenous dynamics of migration from macro-level structures and micro-level desires.
Most often, the endogenous dynamics of migration flows do not reach their internally induced turning point. They are primarily affected by welfare and security policies in the immigration countries. The welfare and security functions of the state do not only form a backdrop for restrictive policies. They also tie in with the symbolic politics of migration and immigrant adaptation which are amenable to immigration as a meta-issue.