1Internal migration and spatial distribution are key indicators for the assessment of integration of new immigrants into their host country (Lieberson and Waters, 1988; Massey, 1981; 1985; Massey and Mullan, 1984). Classical assimilation theories, as well as more recent empirical studies, suggest that new immigrants initially tend to settle in close geographic proximity to others of the same nativity or ethnic group (co-ethnics). As time in the new country elapses, and with the acquisition of new skills, improved information on opportunities in alternative destinations, and increased human capital, they will migrate, thus lowering the level of group concentration (Dunlevy, 1980; Fang and Brown, 1999; Funkhouser, 2000; Gordon, 1964). New immigrants will then have higher rates of migration than will the native born (Bartal and Koch, 1991; Rogers, Henning, and Little, 1995). Other studies on social networks have shown that the propensity for subsequent resettlement among new immigrants is lower if they live in communities with compatriots; nativity concentration reduces the cost of international moves by providing a sense of communality and group belonging, and, more specifically, by assisting in job and housing searches and other adjustments to the new country (Bartal, 1989; Gurak and Caces, 1992; Gurak and Kritz, 2000; Kobrin and Speare, 1983; Kritz and Nogle, 1994; Massey and Denton, 1987; Newbold, 1996; Sammuel, 1988). In other cases, high levels of internal migration among the foreign-born population were aimed at reinforcing the spatial concentration of the nativity groups (Belanger and Rogers, 1994; Neuman and Tienda, 1994; Nogel, 1997; Rogers and Henning, 1999; Trovato, 1988). Nevertheless, immigrant groups vary in the importance they attach to nativity concentration implying different patterns of resettlement and integration (Bartal and Koch, 1991; Kritz and Nogel, 1994; Moore and Rosenberg, 1995; Tilly, 1990).
2Immigrant groups also differ in their human capital which may, in turn, be an important determinant of migration processes. This includes, among other individual characteristics, age, sex, marital status and socio-economic stratification (Goldscheider, 1971; Long, 1988; Ritchey, 1976; White and Mueser, 1994). Migration stage is yet another factor which can affect the odds of migration; increasing length of residence in the new country strengthens human capital and ties to local communities thus reducing the likelihood of resettlement (Kritz and Nogel, 1994; Speare, Kobrin and Kingkade, 1982). Other underlying forces, such as language fluency (Chiswick and Miller, 2001) or structural affinities at the group level, including age composition (Belanger and Rogers, 1994) and percentage of self-employment (Gurak and Kritz, 2000), have also been suggested as possible determinants of the internal migration decisions and adjustment of immigrants.
3The many immigrant groups are variously dispersed throughout the country, and they are thus exposed to different social and economic contexts in their areas of initial settlement. Spatial inequalities in employment opportunities and income, in governmental intervention of settlement assistance, as well as in other non-monetary conditions such as crime rates and climate, accumulate to shape the individual’s quality-of-life. As in other voluntary processes involving social and economic change, these factors determine the necessary threshold of desirability and feasibility of migration (Cadwallader, 1992; Clark and Hunter, 1992; Heaton, Clifford and Fuguitt, 1981; Michalos, 1997).
4Much of the concern with the effect of nativity concentration on internal migration patterns among foreign-born populations has focused on the United States (e.g. Bartal and Koch, 1991; Fang and Brown, 1999; Kritz and Nogle, 1994; Nogle, 1997). Given the importance of separation between church and state as an official American ideology and policy on the one hand, and the increasing emphasis on cultural pluralism and the maintenance of group distinctiveness on the other, differences among immigrants by religion, ethnicity or nativity (country of birth) are important to their social and cultural behaviors including identifiable communal services, formal and informal institutions, as well as enclave economies of business and goods production (Glazer and Moynihan, 1963; Yancey, Ericksen and Juliani, 1976). In Canada, which has also often been a research site of internal migration patterns among the foreign-born population (e.g. Newbold, 1996; Nogle, 1994; Trovato, 1988), two major languages and two cultures form the heterogeneous and multicultural host society with the added desirability of elevating minority immigrant groups as a third element. Israel is yet another unique case of a country which has absorbed a large number of immigrants from diverse socio-cultural backgrounds. The gathering of Jews from around the world to their own country is a core ideal of nation-building in the Zionist agenda and the state of Israel (Goldscheider, 2002). The formal immigration policy first appeared in the Declaration of Independence in 1948, was later anchored by the Law of Return in 1950, and proclaims the right of every Jew to come and settle in Israel. That only Jews are eligible for citizenship upon immigration to Israel suggests that they share the religious identity and historical heritage of the majority population.
5The collective identity of Israel as a Jewish state is manifest in every-day life and in governmental institutions and services. This includes, inter alia, weekends and holidays which accord with the Jewish calendar; a homogenous national Hebrew school system, divided between the general and religious sectors, both of which, under governmental supervision, have curricula which emphasize Jewish themes; dispersion of synagogues and other religious services throughout the country; and an accepted status quo of selling all foods in chain supermarkets and private groceries in compliance with Jewish dietary laws (Kashrut). From the ethnic-specific-goods perspective, foreign languages both in print and in the electronic media are similarly available throughout the country. It should also be noted that the political system in Israel is based on a single voting district of the entire country; the size of the various parties’ delegations in the parliament, which further selects the prime minister and the government, adheres to the distribution of the overall vote (Arian, 2004). Thus, nativity concentration has no role in gaining political power for immigrants’ parties.
6Similarly, Jewish ethnicity broadly defined on Ashkenazic (European-American origin) and Sephardic (Asian-African origin) lines, or by specific countries of origin, is not a source of Jewish-national identity; rather, Israel is (Goldscheider, 2002). Nor is ethnic origin significant for any kind of communal services or social and economic opportunities. Since Israel is a small country, any attempts to preserve and continue ethnic culture and identity through family gathering and kin are independent of residential concentration. Overall, differences of ethnicity or nativity are not bases for separatism in Jewish daily life, socio-economic services, or opportunities for Jews in Israel.
- 1 Only Shaginian introduced an explicit variable of nativity concentration focusing on recent immigra (...)
7The characteristics described above turn Israel into a unique and interesting case study of the issue in question. Very few studies have been concerned with the role of ethnic and social ties in determining the internal migration patterns of immigrants in Israel and they often assessed this dimension only indirectly (Beenstock, 1997; 1999; Hasson, 1998; Shaginian, 2000).1 The present endeavor is built on strong ties between these factors, as well as on the broadening scope of numerous immigrant groups by specific country or area of origin. More specifically, I address the following questions: 1) How do human capital and area context of residence determine the internal migration of immigrants to Israel? 2) How does nativity concentration shape these patterns? 3) How do these determinants of migration vary across immigrant groups? 4) How do immigrants’ migration patterns differ from those of native-born Israelis? To the greatest possible extent, I have adopted the models developed in earlier studies for other immigrant receiving countries (mainly in Kritz and Nogle, 1994 for the United States). This will allow testing of the robustness of these approaches as well as expansion of our insights into the specifics of immigrant adjustment in diverse settings (Chiswick and Miller, 2001).
8The distance between any two cultures, namely the hosting and the absorbed, in the major socio-cultural component — be it racial, religious or ethnic belonging (and associated dimensions such as language) — affects the course of immigrants’ integration (Ward and Kennedy, 1992). The general and consistent findings demonstrate that greater cultural differences imply “the need for greater cultural shedding and cultural learning, and perhaps large differences trigger negative inter-group attitudes, and induce greater culture conflict leading to poorer adaptation” (Berry, 1997: 23). Under such circumstances, immigrants will attempt to segregate from the social mainstream to maintain high levels of cohesion including residential concentration.
9In Israel, the major source of distinction between sub-populations is their religious belonging. The Jewish identity of the immigrants provides an important feeling of mutual commitment and cohesion with the dominant society, thus substantially moderating cultural distance. Both the “absorption policy” which reflects the formal attitude of the establishment, and the “absorption climate” which is the informal reaction of the public at large, show appreciation and a welcoming attitude toward the new immigrants. Yet, shortly after the beginning of the large wave of Soviet immigrants in 1990, some negative stereotypes and prejudices against them began to evolve. This growing antagonism was mainly expressed by Israelis from low socio-economic status who were struggling towards upward mobility and perceived the new immigrants as competitors (Leshem and Sicron, 2004). This may act to preserve social and cultural surroundings. Such negative attitudes were not observed towards the smaller and relatively wealthy immigrant groups from America, Western Europe, or South Africa.
10We propose that the social and psychological desire to live within a familiar cultural milieu of co-ethnics is short-lived and limited to the initial stages of settlement in the new country. Nativity concentration will have a small, if at all statistically significant, effect on the internal migration of immigrants, and it will mainly characterize immigrants from the various republics of the former USSR.
11Immigrants vary according to their motivations to move to the new country. They are spread along a continuum between reactive and proactive, with the former responding to push factors at origin and the latter being facilitated by economic or ideological pull factors at destination (Lee, 1966; Richmond, 1993). People who migrate mainly due to high push motivations have more psychological adaptation problems (Berry, 1997) and are thus anticipated to maintain for a longer time their original daily interactions and their preferred residential concentration. Empirical support for this model can be found, inter alia, in Nogle (1994) and Mesch (2002).
12Given the data at hand, these effects can not be measured directly. Yet, we can roughly distinguish between immigrants from North America, Latin America and West Europe, on the one hand, who were motivated by religious and national factors (Dashefsky et al., 1992; DellaPergola, 1986), and East European immigrants many of whom responded to unfavorable political and economic conditions in their countries of origin (Leshem and Sicron, 2004). Furthermore, the former groups often had prior knowledge and direct acquaintance with Israeli society including relatively high language proficiency; after a short-term adaptation, such characteristics are likely to avoid ethnic residential enclaves. We anticipate divergent patterns of relationships with nativity concentration having no meaningful effect on relocation among American and West European immigrants, but a significantly negative effect among immigrants from the former USSR and other Soviet satellite countries.
13Institutional intervention is significant for patterns of interregional migration and residential location choices (Cadwallader, 1992). Perhaps more than any other economic incentive, housing policy through regional bias of housing supply and related benefits such as high quality, low prices and favorable mortgages, is strongly inherent to a given area. Responses to housing availability vary among different sub-groups with the strongest effect being on people with low socio-economic status and other disadvantages who prefer housing security to labor market opportunities (Clark and Whiteman, 1983). New immigrants may often fall into these categories and serve as a target population for governmental policies aimed at determining spatial population distribution whether to develop desolated and sparsely populated areas (King, 1993), to eliminate ethnic segregation (White, 1993), or to achieve political and security goals (Clark, 1983; Lipshitz, 1991; 1998).
14Research findings show that massive government building within the national periphery, accompanied by financial incentives, was important among recent Soviet immigrants to Israel in their decision-making as to place of residence (Shaginian, 2000). Dwelling considerations outweighed employment opportunities and access to jobs that best suit the professional skills and educational level of the immigrants (Gotlibovski, 1995; Hasson, 1998). A few studies that were concerned with the living arrangements of American and West European immigrants (Avruch, 1981; Dashefsky et al., 1992) emphasize the importance of a neighborhood that is socially suitable, the quality of the local educational system, and the size of apartments as determinants of residential choice. Thus, many American immigrants reject public housing and look for apartments on the private market, i.e. places where established Israelis tend to predominate. These apartments are often expensive and inversely related to the level of mortgage to which they are entitled. Single persons or young couples often receive assistance from their parents abroad.
- 2 In the 1950s and 1960s, many of the immigrants to Israel, especially from countries in North Africa (...)
15Shortly after the stage of initial settlement, immigrants search for employment opportunities in the open economy (ethnic enclave economies in Israel are not very developed), and for housing. Given the smallness of the country, many of the peripheries can be considered as outer fringes of metropolitan areas, and the developed public transportation enables convenient and easy commuting. Thus, we anticipate a strong positive relationship between housing opportunities, which were improved in the early 1990s by governmental intervention including financial assistance, and resettlement patterns. In the spatial trade-off involving inexpensive housing on the one hand, and high employment opportunities and salaries on the other, the former will have greater significance. This will be mostly salient among immigrants from the Former USSR and other East European countries who arrived in Israel with a relatively little capital. 2
16Data for this study are drawn from the 1995 Israel Census of Housing and Population (20 per cent “geographic version” file). The sample is restricted to the foreign-born aged 25 through 64 who arrived in Israel between 1970 and 1990 from Europe, America, Oceania and South Africa; these immigrants comprise 85.3% of the total foreign-born population who entered the country during this period. I focus on household heads who were not enrolled in school in 1995, rather than multiple adults, thereby eliminating the potential bias for interdependence of migration behavior (Kritz and Nogle, 1994).
17Applying these criteria, I generated a sample of 17,642 immigrants. They were aggregated into 15 countries, or regions, of origin: three from the Americas (United States, Argentina, other Latin America), three from Western Europe (France, United Kingdom, other Western Europe), and nine from Eastern Europe and the Former USSR (Russia, Ukraine, Moldova, Belarus, Baltic States, other Former USSR in Europe, Georgia, other Former USSR in Asia, and Romania); the remaining foreign-born persons were merged into a sixteen group. Group sizes range from 396 persons from the United Kingdom to 3,037 from Russia. Although we would have preferred a more detailed aggregation of all immigrants by country of origin, which would have decreased the internal heterogeneity among a few immigrant sub-groups, the relatively small sample data at hand did not permit this.
- 3 ‑Israel is officially divided into six districts. Each district is further divided into smaller ge (...)
- 4 I am aware of other models which are appropriate for testing hypotheses with these data. Perhaps mo (...)
18The dependent variable distinguishes between immigrants who stayed in their beginning-of-period place of residence, those who moved within the same sub-district, and those who departed to another sub-district over the five-year interval examined.3 Sub-districts divide the country into relatively coherent areas in regard to their physical structure; at the same time, they are large enough that migration between them constitutes a meaningful geographic change. The categorical character of the dependent variable with three possible values is appropriate for multinomial logistic regression.4
19Nativity concentration was calculated by allocating migrants back to their beginning-of-period area of residence and measuring the percentage of immigrants of a given group in each sub-district. All persons of a given group who lived in a specific sub-district have the same concentration value. Nativity concentration is treated as a continuous variable.
20Explanatory individual characteristics included in our model are sex, age, education, marital status and employment status. Sex was set equal to 1 if the person was female, with males being the reference category. Age is treated as a continuous variable. To account for the particularly high levels of education among western immigrants to Israel (Israel Central Bureau of Statistics, 2000), I included education as a continuous variable: a code of 1 indicates less than high school education; 2 indicates high school graduation; 3 indicates 1-3 years of college; 4 indicates baccalaureate degree; and 5 indicates advanced academic degrees. Marital status distinguishes between married persons who served as the reference category, and two dummy variables of singles and ex-married (divorced and widowed). Also added as a control sociodemographic variable is the nature of employment, with self-employment coded 1 and those working for others as the reference group.
21I employed three measures to evaluate the effect of contextual conditions on migration within and between sub-districts: per capita income, unemployment rate, and rate of new housing construction. The data on income refer to the condition in the sub-district of residence at the beginning of the five-year migration interval. In order to reduce the possible effect of intertemporal fluctuations in rates of unemployment, I use the average of the mean total unemployment rate for each sub-district for the five-year interval (1990-1995). New housing construction is the deviation from the national average rate of each sub-district in the number of houses whose construction was completed from 1990 to 1995 per 1,000 inhabitants. Data for the contextual measures were drawn from various official sources (i.e. Central Bureau of Statistics, National Insurance Institute, and Ministry of Interior); they are introduced as continuous variables, and were attached to individual records.
22In addition to nativity concentration, another immigration factor which is included in the model is time in Israel. It reflects the time elapsed since immigration to the beginning of period (i.e. 1990) and distinguishes between duration of up to one year, 1 to 2 years, 3 to 10 years, and 11 or more years in Israel. Time in Israel is measured as a categorical variable with the longest duration as the reference category.
- 5 In this study native Israelis refer to people of European-American origin.
23Table 1 shows that the migration rate for foreign-born household heads age 25-64 is significantly higher than for native Israelis (50.9% and 38.7%, respectively).5 This is true for most of the immigrant groups separately but particularly for immigrants from Belarus, Ukraine and “other Former USSR in Asia”. Three groups exhibit lower migration rates than do natives including “other West Europe”, Georgia and Romania.
Table 1: Levels of five-year migration and group concentration for foreign-born heads of household age 25-64, Israel 1995
24Both immigrants and natives are more likely to move to another sub-district than to move within their sub-district. Generally, immigrants in 14 of the groups are more likely than natives to move to another sub-district. The three groups with the overall highest level of migration, i.e. Belarus, Ukraine and “other Former USSR in Asia” are also the most likely to move to another sub-district. Immigrants from “other West Europe” and Romania who had an overall lower level of migration than natives are less likely to migrate within as well as between sub-districts; however, immigrants from Georgia have lower rates of migration than natives within the same sub-district but tend to move more to other sub-districts. Similarly, the overall higher rates for immigrants from the United Kingdom and other Foreign-born than natives is attributed to their inclination to move to other sub-districts although they are less likely than natives to make a move within the same sub-district.
25Differences are revealed between immigrant groups in their levels of concentration by sub-district. There are seven sub-districts which contain the heaviest concentration of immigrant groups. The highest proportion is for immigrants from the United States among whom 37% live in the sub-district of Jerusalem; this sub-district is likewise a preferable locality for a relatively high rate of French (23.5%) and British (22.2%) immigrants. A high concentration rate also characterizes immigrants from Georgia, with 29.2% living at the southern sub-district of Ashqelon. Argentinians and immigrants from other parts of Latin America have very low concentration rates which are approximately one-third those of the most concentrated group; other Former USSR immigrants, both from European and Asian republics, also have very low levels of concentration.
26Table 2 presents the means of individual and immigration characteristics for each of the immigrant groups. Variations in life cycle and socioeconomic characteristics, as well as in duration in the host country, are likely to affect mobility patterns and residential preferences. In general, immigrants from America and West Europe are younger than immigrants from East Europe and the Asian republics of the FUSSR. The sex composition of household heads shows a relatively high percentage of females among groups from the United States and West Europe, reflecting a stronger tendency to immigrate of women in their early adult ages (Goldscheider, 1974: 364), as well as among Russian immigrants which is likely attributed to a widespread phenomenon of one parent families among this group (Sicron, 1998).
Table 2: Sociodemographic and immigration characteristics of foreign-born heads of household age 25-64, Israel 1995a
a) Data are means of covariates. Age is in years; sex=1 for female; education is a scale of the mean level: less than completed high school=1, high school graduation=2, some college=3, baccalaureate=4, advanced academic degree=5; type of employment=1 if self-employed.
27The educational level of immigrant groups depicts significant variations with the mean scores ranging from 1.75 for immigrants from Georgia to considerably higher levels among immigrants from the United States (3.84), Romania (3.58) or Belarus (3.55). The proportion of currently married persons, which is largely affected by the age structure of each immigrant group, is especially high for immigrants from Moldova (86.5%) and Belarus (83.5%), while it is considerably lower for French (71.4%) and “other Latin America” (71.9%). Less affected by the age composition is the rate of ex-married (most of whom are divorced) with very high rates among the Russian and Ukrainian immigrants. Low levels of entrepreneurial activity, of between 8% and 13%, were found among groups from the Former USSR; however, immigrants from the United States, the Latin America countries and the United Kingdom own private businesses at rates above 20%.
28The distribution of time in Israel across immigrant groups shows that a majority of American and west European immigrants have an experience of more than 11 years in Israel; another significant proportion among these groups have lived in Israel between 3 to 10 years. This pattern of time of arrival in Israel also characterizes Romanians who, despite their origin in a communist satellite country were permitted, in relatively large numbers, to leave Romania in the 1970s and early 1980s. By contrast, the waves of Soviet immigration to Israel are divided rather dichotomously, albeit unevenly, between the earliest period (most of whom arrived between 1975 and 1979) and the latest period following perestroika.
29Table 3 presents the results of multinomial logistic regression for the evaluation of the effect of different sets of factors on migration within and between sub-districts, relative to no migration, for foreign-born heads of household in Israel. The first model considers the role of human capital (sex, age, education, marital status, type of employment and an interaction variable of sex by education) and area economic context (unemployment, income, housing construction) as determinants of migration. The second model incorporates time in Israel and nativity concentration. In order to gain insights into inter-group variations in migration patterns, the third model introduces 15 dummy variables, each of which represents a different nativity group, with other West Europe as the reference group. The relationships between the independent variables and migration are presented as odds ratios (exp[b]) which express the relative odds of the event (migration) occurring. A “pseudo R2” value (Nagelkerke R2) for the multinomial logistic analysis is also shown to illustrate the explanatory power of each model.
Table 3: Multinomial logistic regressions (odds ratios) of migration within sub-districts and between sub-districts on individual characteristics area context variables, immigration factors, and nativity group for foreign-born Israelis heads of household, 1990-1995a
a) Omitted categories are: male; married; employee; 11-20 years in Israel; immigrant from other West Europe. *P<.05; **P<.01; ***P<.001
30Most of the human capital variables in Model 1 exert a statistically significant effect on migration. As expected, sex and age each decreases the likelihood of migration both within and between sub-districts. High educational attainment enhances the probability of making an internal move: for each point increase in level of education, the odds of making a local move within the sub-district were 23% more, and for migration between sub-districts, 29% more. The findings also indicate that marital status produces a significant effect on migration with singles being less likely than married persons to move within a sub-district while the ex-married (most of whom are divorced) have higher odds of migrating within as well as between sub-districts with a respective odds ratio of 1.19 and 1.36. Self-employment, in turn, deters both types of migration. An interaction effect of education by sex was found significant only for migration within sub-districts suggesting that females with higher levels of education are more likely to move locally than are males with similar educational attainment.
31Unemployment increases the likelihood of migration within a sub-district, whereas for each point increase in the unemployment rate, the odds of migration were 12% higher. No significant effect of unemployment was proven on migration between sub-districts. This latter finding on the macro level largely coincides with the insignificant influence of a micro level variable of labor market status (employed vs. unemployed) on migration between four major regions (Beenstock, 1999); as well as with the higher frequency of origin from localities of low unemployment level than from areas of high unemployment level (Shaginian, 2000). Under conditions of lack of employment, immigrants do not hurry to change their area of residence; given the relatively small size of Israel, they prefer to commute to areas with better job opportunities which are likely to be situated in large metropolitan areas. Migration within a sub-district with a high unemployment level may reflect the diminution of economic resources on the part of the local inhabitants and the need to move to a lower standard of housing (which among immigrants may often be rented) while endeavoring to maintain a familiar residential environment and close proximity to relatives and compatriots.
32The odds ratio for per capita income suggests that a large change in this independent variable led to only a small change in the probability of moving between sub-districts. By contrast, housing construction was found to be a most salient and meaningful determinant of migration between sub-districts; immigrants who live in sub-districts with above average rates of new housing construction are less likely to move than immigrants living in sub-districts where the rate of new housing construction is below the national average. Taken together, the individual characteristics and area context of residence were relatively effective in explaining the variation in migration behavior with R2 of 15.3 per cent.
33Model 2 assesses the importance of immigration factors as determinants of internal migration among the foreign-born. It shows that recent arrivals are significantly more inclined to move both within and between sub-districts. Within the first year in Israel, immigrants are six times more likely to migrate within sub-districts as are immigrants with 11-20 years in Israel; the parallel odd of migration between sub-districts is ten times more. The likelihood of migration gradually decreases as time in Israel grows, although it remains statistically significant for migration between sub-districts also among those who are in the country for 3-10 years. Nativity concentration very slightly increases local migration within sub-districts. The interpretation is that immigrants either resettle closer to co-natives within the limited boundaries of the sub-district which has a high concentration level, or that the large size of immigrant population in the area promotes local residential moves derived by macro-level considerations while ensure the maintenance of close proximity to co-natives. By contrast, nativity concentration has a negative and statistically significant relation to migration between sub-districts. The odds ratio of 0.96 indicates that heads of household who in 1990 resided in sub-districts with a higher presence of their compatriots (percentage group size), were less likely to move to another sub-district than were those living in sub-districts with smaller concentrations of co-natives (for each point increase in nativity concentration, the odds of migrating were 4% less).
34This model denoted twice as much variance than the previous model with 30.2% of the variation in internal migration between 1990 and 1995 being explained by means of the independent variables, many of which turned out to have a significant statistical effect. Yet, it should be noted that separate models (not shown here) with and without different sets of data, point to the paramount role of duration in Israel in the increasing explanatory power of model 3 and a very slight contribution of the ethnic concentration variable: if duration in the country is removed, the explanatory power of the model diminishes to 16.1%, while the removal of the variable ethnic concentration would have resulted in only a slight decline to 29.7%.
35Model 3 incorporates each of the nativity groups in order to evaluate differences in migration patterns. Using immigrants from “other West Europe” as the reference group (due to their rather similar sociodemographic profile to that of natives), the odds ratios show significant differences in migration within sub-districts for 12 of the 15 groups; all groups are more likely than “other West Europe” to move within their sub-district. Especially high levels of migration within sub-districts were found among immigrants from Belarus (odds of 2.18), from Moldova (odds of 1.89) and from Ukraine (odds of 1.73). Six of the immigrant groups have a significantly different propensity than “other West European” to move between sub-districts; these groups include immigrants from the United States, Argentina, other Latin America, France, Belarus and the Baltic States. All of these groups are more likely than “other West Europe” to migrate with odds ratios ranging from 1.48 for Americans to as high as 2.03 for Argentinians. The insertion of the different nativity groups did not meaningfully improve the explained variance over that of the previous model.
36Immigrant groups to Israel come from different social and cultural backgrounds. This involves different levels and types of heterogeneity between areas of origin and the area of destination. The immigrants are motivated by unique sets of push and pull factors, and may have come as individuals or in mass migration. All this may affect patterns of integration into the host society, raising the possibility that each immigrant group considers individual sociodemographic characteristics differently, and factors regarding the economic context or nativity concentration are seen from varying points of view in deciding on internal migration. Thus, I estimated separate models for each of the 16 foreign-born groups; for each group, the full model was estimated retaining all the independent variables, where the dependent variable distinguishes between no migration (the referent category), migration within sub-districts and migration between sub-districts.
Table 4: Multimonial logistic regressions of migration within and between sub-districts on individual characteristics, area context of residence, and immigration factors for foreign-born heads of household in Israel, 1990-1995: Summary table of statistically significant* odds ratios
- 6 These four groups constitute about half of the total sample which can explain the significant effec (...)
37The separate models show that nativity concentration has no statistically significant effect on migration within sub-districts for nine of the groups, and when significant (United States, Argentina, other Latin America, France, United Kingdom, Ukraine and other Former USSR in Europe) it increases the odds of migration (Table 4). Nativity concentration has a statistically significant effect on migration between sub-districts in four of the 16 immigrant groups; it decreases the odds of migration between sub-districts among Russians, Ukrainians, Moldovians and Romanians, all of whose provenance is East Europe.6 One unit of change in the percentage of nativity concentration decreases the odds of migration by 5%-9%. At the same time Short duration in Israel has a significant positive effect on the odds of both types of migration, but particularly on movements between sub-districts. Duration of less than one year increases the odds of migration between sub-districts for 12 of the groups; it remains statistically significant for 1-2 years among nine of the groups albeit with a smaller odds ratio. Note that in comparison to the full model, a reduced model without the variable of ethnic concentration (not shown here) only slightly diminished the percentage of variance being explained: from 31.8% to 30.8% for Russia, from 35.5% to 34.8% for Ukraine, from 36.5% to 35.9% for Moldova, and from 19.1% to 18.6% for Romania. If ethnic concentration is maintained in the model but the other immigration variable of time in Israel is removed, the explanatory power of the model is substantially hurt and decreases: to 20.3%, 17.4%, 23.0% and 18.6%, respectively.
38Area unemployment rates have a significant positive effect on the odds of migration between sub-districts for five of the groups including immigrants from the United States, Argentina, France, United Kingdom, and other Foreign-Born. Of these, the first three groups also want residential opportunities with above average housing construction and this decreases the odds of migration to another sub-district. Housing construction has a very strong negative effect on migration between sub-districts among the immigrants from Russia, Ukraine, Moldova and Belarus as well.
39Among the human capital variables, age consistently decreases migration throughout all the groups both within and between sub-districts. Both types of migration are positively associated with education; the education effect is especially salient for immigrants from Russia, Ukraine and Romania. Single and ex-married persons, relative to married persons, are more likely to migrate to another sub-district. This effect tends to be significant for immigrants from the United States and from a few republics of the Former Soviet Union (e.g. Russia, Ukraine and other Former USSR in Europe), but very weak and statistically insignificant for immigrants from Latin America and West Europe. Somewhat surprisingly, the effects of sex and self-employment on the odds of migration are mixed and often insignificant; these findings are in contradiction to the strong negative effect of sex and self-employment for the total foreign-population. While this requires further analysis, at this stage we can point out that the significant effect of self-employment on migration between sub-districts for the total foreign-born was caused largely by Argentinians and immigrants from other Latin America; this effect is strongly negative on the migration of both of these two groups which differ significantly in their levels of self-employment as compared with other immigrant groups. Noted that the addition of interaction terms of sex and nativity and of self-employment and nativity (not shown) proved to be insignificant.
40Finally, I compared the behavior of each immigrant group with that of individuals who were born and raised in Israel. To that end, I estimated two multinomial regression models: a zero-order model which introduces only the nativity group as a dummy variable with natives as the reference population; and an adjusted model which controls for group differences in all the covariates used in the analysis. Thus, I prepared two separate sets of 16 multinomial regressions each of which contains all the immigrant groups. Since “time in Israel” is not comparable for immigrants and natives, the latter were assigned the longest duration of 11 to 20 years. For nativity concentration of Israeli-born heads of household I adopted an approach similar to that for immigrants by measuring the percentage of natives in each sub-district.
41The zero-order models suggest that with the exception of two origin groups, from the United States and other Latin America, all the immigrants have significantly different propensities to migrate than do native Israelis (Table 5). Immigrants from Argentina, France, Russia, Ukraine, Moldova, Belarus, Baltic States, other Former USSR in Europe and other Former USSR in Asia are more likely than natives to migrate both within and between sub-districts. Immigrants from other West Europe are less likely than natives to migrate within as well as between sub-districts; immigrants from Georgia and Romania also exhibit lower propensities to move between sub-districts than do native Israelis. Controlling for the other covariates produced substantial reductions in the differences between immigrants and natives in migration patterns (e.g. Russia, Belarus, other Former USSR in Europe) or even their total disappearance (e.g. France, Baltic States).
42Nevertheless the human capital, area context and immigration covariates do not fully account for the zero-order differences between immigrants and natives; the adjusted odds point to the maintenance of significant differences from natives in migration within sub-districts for nine groups (three of which have lower odds of migration and six higher odds) and in migration between sub-districts for five groups (three of which have lower odds and two have higher odds). The inference is that there are other factors, not introduced in our models, which are important determinants of internal migration behaviors of a few large immigrant groups in Israel. These factors may operate differently for each of these groups either to reduce or to increase their likelihood of relocation.
43This paper has explored the effects of nativity concentration on the internal migration patterns of the foreign-born in Israel and how these patterns differ from those of native Israelis. It was written with the intent to contribute to the growing literature on this topic which has focused on other countries which receive large waves of immigration including the United States and Canada. However, Israel is a unique case since its formal immigration policy grants the right to settle in the country and citizenship only to Jews; thus they share the religious identity and historical heritage of the majority population. Furthermore, ethnic-specific-goods are readily available throughout the state; the political system is based on a single voting district; and the smallness of the country makes maintenance of family and other kinship ties possible, independent of residential concentration.
44Largely in accordance with our first two working hypotheses, the group homogeneity hypothesis and the immigration motivation’s hypothesis, the results of multinomial regression analysis show that all other things being equal (human capital, economic context and immigration factors), nativity concentration deters migration between sub-districts and operates to hold people within the same sub-district even when they wish to move because of individual needs or economic opportunities. The contribution of the ethnic concentration variable to the overall explanatory power of the model was nevertheless small; other factors, including duration in Israel, were more important in resettlement patterns and migratory behavior of immigrants. The statistically significant and negative effect of ethnic concentration on migration between sub-districts in the integrative model is attributed to the nature of relationships among immigrants from Russia, Ukraine, Moldova and Romania as revealed by the estimation of separate models for specific country or area of origin. These groups originated in East Europe, mainly responding to push motivations, and at times were also faced with some rejecting attitudes from certain segments of the veteran Israeli population.
45The third hypothesis, of the institutional approach, was only partly confirmed by our findings. Housing construction above the national average substantially deters migration to another sub-district. This economic incentive is, however, important not only for Soviet immigrants of presumably low economic status; statistically significant and large odds ratios were also found for relatively wealthy immigrant groups including from the United States, Argentina and France. Data not shown here suggest that the preferred areas with high rates of housing construction differ among the immigrant groups: Jewish-Soviet immigrants are mainly moving to less developed towns in southern sub-districts; a large proportion of American and French immigrants are relocating to Tel Aviv (Israel’s primary core) and its nearby peripheries which are characterized by high socioeconomic well-being. Thus, rather than being concerned with economic incentives, they are mainly motivated by the life styles associated with these areas of settlement.
46Some significant variations exist between immigrant groups in their initial geographic dispersal and in rates of migration as well as according to key individual affinities. These sociodemographic profiles also differ from those of the native-born population. For several immigrant groups it was found that similarity with native Israelis in these characteristics would imply marked convergence in migration patterns. By contrast, other immigrants, including those from Argentina, Russia, Belarus, other Former USSR in Europe, and Georgia, exhibited unique propensities to migrate both within and between sub-districts. The spatial integration of these groups, which include substantial numbers of immigrants who recently arrived in Israel, is to a great extent determined by factors other than those considered in this study.
47With all due caution, and given the objective limitations to evaluation of these findings from a comparative perspective, it seems that the importance of social capital such as geographic proximity to people of the same nativity is weaker in Israel than in other countries which absorb large numbers of immigrants such as the United States. Important characteristics shared by immigrants and the host society, alongside wide acceptance by, and willingness of, both governmental agencies and the public at large to assist in the initial stages of settlement, enhance the sense of security in the new and unfamiliar environment and reduce the need for residential segregation. At the same time, the physical conditions of a small country, which ease the maintenance of ties in primary groups of family and friends and allow access to specific ethnic activities, moderate the rhythm of social and cultural assimilation of immigrants. This largely coincides with the somewhat ambivalent approach among Soviet immigrants who attempt to achieve integration as a secular group while preserving the legitimacy of their cultural uniqueness (Lissak, 1995). The equilibrium between these contradicting inclinations varies according to the specific immigrant group; those groups which attach importance to high territorial density, such as the Russians and Ukrainians, wish to preserve their original culture without compromise rather than to maintain a few selected cultural components while waiving others. These being the largest immigrant groups, they have more resources than other groups. Furthermore, they come from metropolitan areas which are the major cultural, political and scientific centers of the former Soviet Union and adhere to their cultures of origin which they see as being superior to that of the Israel’s host society.
- 7 Indices of concentration are expressed here by coefficients of variation, i.e. the ratio of a distr (...)
48The strong effect of housing construction provides an effective tool for a government wishing to shape the residential patterns of the new immigrants. However, this institutional approach presents a paradox: on the one hand, it may attract and pull people to certain areas, often in the metropolitan or national peripheries, thus contributing to the overall dispersion of the population; but on the other hand, this geographic dispersion will again create a concentration because large numbers of the new immigrants will relocate to the same destination. An index of concentration shows that despite the short duration, between 1990 and 1995 most foreign-born groups have increased their spatial concentration with immigrants from Belarus and a few other European republics experiencing the most substantial change.7 When a group is spatially isolated and segregated, its cultural integration process and entrance into primary group relationships with the majority society will be very slow. These policy considerations should receive even greater attention in regard to later arrivals from the former Soviet Union among whom a growing proportion are people with weak religious and national-Zionist identification; more than their earlier counterparts, they will presumably seek high socio-cultural spatial segregation.
49This research was supported by a Grant from The Israel Foundation Trustees (Project AG44/2002); and by a Grant from The Harvey L. Silbert Center for Israel Studies, The Hebrew University of Jerusalem. Data analyses presented here were managed by Dalia Sagi whose assistance in the project was invaluable. I gratefully acknowledge the help of Benny Anderman in computer programming; of Israel Eynot for his statistical advise; and of Judith Even for editorial assistance. Final polish of this manuscript was done during the author’s sabbatical year, as a Visiting Scholar, at the Center for Demography and Ecology, The University of Wisconsin-Madison. Responsibility for the contents of this article rests solely with the author.