Jump to ContentJump to Main Navigation
Shifting Ethnic Boundaries and Inequality in IsraelOr, How the Polish Peddler Became a German Intellectual$

Aziza Khazzoom

Print publication date: 2008

Print ISBN-13: 9780804756976

Published to Stanford Scholarship Online: June 2013

DOI: 10.11126/stanford/9780804756976.001.0001

Show Summary Details
Page of

PRINTED FROM STANFORD SCHOLARSHIP ONLINE (www.stanford.universitypressscholarship.com). (c) Copyright Stanford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in SSO for personal use (for details see www.stanford.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 16 January 2019

Residential Segregation and Economic Isolation: The Moroccan Paradox

Residential Segregation and Economic Isolation: The Moroccan Paradox

(p.162) Chapter Eight Residential Segregation and Economic Isolation: The Moroccan Paradox
Shifting Ethnic Boundaries and Inequality in Israel
Stanford University Press

Abstract and Keywords

This chapter examines how dichotomization occurred, by considering the impact of residential location on labor market outcomes. It considers the Moroccan paradox—that Moroccans who were relegated to single industry, low opportunity areas (development towns) had better returns to education and better overall attainment than Moroccans who lived elsewhere. It shows complex relationships between ethnicity and attainment in the first encounter with the labor market. It considers when educational attainments were distributed to the immigrants' children through the national school system. It finds a dynamic that caused Iraqi attainment levels to drop to those of other Mizrahim: in the schools, Iraqi boys experienced ethnic discrimination, such that they obtained no returns to their fathers' occupational attainments.

Keywords:   Moroccan paradox, residential location, labor market outcomes, occupational attainment, national school system, Mizrahim, ethnic discrimination

In a recent Israeli movie named Go to the End of the World and Turn Left, two families have been unceremoniously dumped into an isolated Israeli settlement deep in the southern desert. One family is Moroccan but speaks French at home, another is Indian but speaks English at home. These two families must mingle with the more numerous “Berber” Jews, who exemplify the eastern stereotypes from Chapter 6 in everything from their inability to control bodily functions to their crass manners, dress, and posture, to the religious head coverings of some older women. In fact, they are so lacking in decorum that when they join the Anglophile's cricket team and play against visiting British diplomats, they cause the game to degenerate into a disorganized brawl. That brawl, and the embarrassment of the Europhile Mizrahim in front of non-Jewish European onlookers, are reminiscent of Segev's (1986) description of the first Independence Day parade (Chapter 6), in which I argued that Ashkenazi Jews mourned their inability to uphold western ideals.

But in the main, the movie chronicles the sense of isolation, betrayal, and depression of the Europeanized immigrants, as they cope with their Oriental neighbors, their marginalization from Israeli urban centers and cultures (Tel Aviv, says the Francophile Moroccan mother, is Israel's Paris), and with the downward occupational mobility caused by their placement in the desert town. By the end of the film, there is evidence of accommodation and even dichotomization. The daughter of the Francophile family questions her mother's edict that she set herself apart from the Berbers and wonders if her mother has not exaggerated the number of family members who attended the French Sorbonne. As the daughter begins to get friendly (p.163) with a Berber suitor and finds that she must stay in the town rather than join the army, one gets the sense that distinctions between Europeanized and non-Europeanized Mizrahim will, through residence in the town, attenuate. This lessening of differences between different Mizrahim is a basic dynamic of dichotomization.

The movie was set in one of Israel's development towns, new settlements on the geographic borders of Israel that were established in order to populate the outlying areas to which most immigrants did not want to move. The movie was written by the daughter of the Francophile Moroccan protagonist and contains two claims that reflect Moroccan Jewish collective memory and recent academic research. First, it claims that Mizrahim were indiscriminately relegated to these isolated Oriental spaces. This has been a source of debate for decades, but as I noted in Chapter 5, I found strong evidence in favor of the movie's claim in earlier work (Khazzoom 2005b).1 Second, the movie claims that relegation to these spaces resulted in downward occupational mobility. Academic research has largely agreed and has added that because Mizrahim were more likely to live in the towns, the towns themselves are one cause of occupational gaps between Mizrahim and Ashkenazim. That claim is the focus of this chapter. As I will show, the story is not simple.

Hypotheses for the Relationships Among Settlement Type, Moroccan Attainment, and Dichotomization

The Debate over Development Towns in Israel

In the movie, downward mobility occurs because the only jobs available in the town are in its bottle-making factory. Even the position of factory manager is not available to Mizrahim, Europeanized or not, because it is taken by an Ashkenazi, who, in his demeanor and sexism, is reminiscent of the Ostjude stereotype.2 Though a labor strike led by the Francophile Moroccan father is eventually successful, the Mizrahim only win higher salaries; the managerial position is still occupied by the Ashkenazi, and the only form of true upward mobility is finishing high school and leaving town.3 It is an option that is taken by the Indian daughter.

This occupational story might well have been written by Shlomo Swirski (1989), who, in combination with Deborah Bernstein, launched academic (p.164) research into the state's role in creating ethnic inequality (Bernstein and Swirski 1982). For Swirski, development towns were tools in the proletarianization of Mizrahim. Like the town in the movie, most development towns had only one industry that provided most jobs in the town and offered residents low skill, low status work that left little room for occupational mobility (see also Spilerman and Habib 1976).4 Swirski argued that in placing Mizrahim in the towns, the state ensured that they had no choice but to work in the factories. Moreover, he argued that to establish the factories, the state gave loans to Ashkenazim who in reality were not any more prepared to manage them than the Mizrahim who worked there. This recalls the movie's contrast between the classless Ashkenazi factory manager and the more cultured and sophisticated, but only temporarily present, poet-teacher from Tel Aviv. Swirski's dynamic also leads directly to dichotomization; in positing that Mizrahim were indiscriminately proletarianized he suggests that class differences between them were reduced, and in positing that equally unprepared Ashkenazim obtained managerial jobs, he suggests that class differences between Ashkenazim and Mizrahim were created.

In contrast to Swirski, early academics and state planners portrayed the establishment of the towns as a necessary, and in fact beneficent, framework for integrating new immigrants into the society. They argued that in the early days Israel was inundated by large numbers of unskilled and uneducated immigrants, particularly those who arrived from the underdeveloped economies of the Arab world. The immigrants were choking city economies, services, and residences, and, given their low skills, were having difficulty finding work. Locating them in areas in which they had ready employment, and, more importantly, job training, sheltered them from the more competitive economies of the cities and central area of the country. The low skilled industries were an advantage, these scholars argued, because it took little time to train immigrants for productive work, thus they could be given jobs right away. Moving these immigrants to the towns also benefited the state by providing Israel with a working industry and productive infrastructure and by populating the militarily important outlying territories (Altman and Rosenbaum 1973; Amiram and Shachar 1969; Berler and Shaked 1966; Comay and Kirschenbaum 1973; Spiegel 1967).

Today, opinion is divided between followers of Swirski (1989)—who continue to see the towns as state tools in the generation of ethnic inequality (p.165) (Yiftachel 1998, 2000; Tzfadia 2000)—and a modified version of the earlier pro-state arguments. Recent work in this latter vein has dropped the assertion of state beneficence, and in fact often casts the state as cynically manipulating immigrants. In this conception, the state needed to populate these outlying areas, so it sent “weak” immigrants—mainly those with low human capital or large families—to the towns. Since these groups had fewer options and were normally dependent on state support, they were better targets. For this group of researchers, the concentration of Mizrahim into these areas of low opportunity contributed significantly to the development of ethnic gaps in resources across the country as a whole (Spilerman and Habib 1976; Semyonov and Tyree 1981; Lewin-Epstein et al. 1995; Lipshitz 1995; but see Adler et al. 2001).

The empirical implications of these debates center around how much of a disadvantage living in the towns is expected to be for the arriving immigrants and whether there is an ethnic difference in the effect of living in the towns. Swirski's argument that the towns were vehicles for proletarianization of Mizrahim implies that residence should have lowered Mizrahi attainment relative to the attainment of Mizrahim elsewhere. Similarly, his argument that they offered unqualified Ashkenazim managerial positions implies that town residence increased Ashkenazi attainment. The other two positions, in positing that weak immigrants were sent to the towns, imply that those who were placed in the towns were unlikely to be hurt by the low opportunity there, at least in the first generation, since they would not have been able to obtain better positions in any case. Even further, the earlier argument that the towns' single industries protected desperate immigrants implies that town residence should actually increase the attainment of Mizrahim, especially relative to those in the unprotected markets of the large cities.

Background and Variables Used

I focus here on urban settlements in Israel and divide them into three types: Israel's three cities (Tel Aviv, Haifa, and Jerusalem), “regular” towns, and development towns.5 These three types of settlement differed from one another on the following criteria: proximity to economic centers, ethnicity and immigrant status of residents, and size. Regarding proximity, Israel's social and economic centers were largely in the three cities, primarily Tel (p.166) Aviv and Haifa, and regular towns tended to be closer to one of the cities than development towns (though there are important exceptions). Regarding ethnicity and immigrant status, in 1961 regular towns tended to have higher veteran populations and lower Mizrahi populations (Appendix 19). And regarding size, regular towns had overall larger populations, although there is also significant variation in size within both town types (see Appendix 18). In addition to these three settlement types, one can distinguish among three rural types: development towns, cooperative farms including kibbutzim and moshavim, and other rural areas. However, as noted, after some initial comparisons, I focus on the three urban types.

Settlement type is important in the analysis of Israeli society because it affected access to labor markets and therefore occupational opportunity. Because they were economic centers, the cities offered the most white collar “empty spaces” for new immigrants to fill, followed by regular towns and then development towns (Appendix 21). Among all Jewish men in the labor force in Israel in 1961, 35% of those in the cities had white collar occupations versus 27% of those in regular towns and 16% in urban development towns. For new immigrants alone, the numbers are 27%, 22%, and 14%, respectively. In 1961, these differences in occupational opportunities were correlated with the immigrant and ethnic makeup of a settlement; thus, in towns with more veterans the overall prestige of jobs available to residents was higher, and in towns where the new immigrant population was proportionally more Ashkenazi, the overall prestige of jobs available to residents was higher.

It is because of these combined differences in opportunity and demographics that scholars turned their attention to the effect of the towns on ethnic gaps in attainment in Israel. However, no earlier analyses have used country of origin as their variable, and when country of origin is used, new patterns emerge.

Effect of Development Town Residence on Prestige in Israel

Figure 8.1 graphs the gap in Israeli occupational prestige points between new immigrants who lived in the cities on the one hand and those who lived in development or regular towns on the other.6 One can see that without (p.167)

Residential Segregation and Economic Isolation: The Moroccan Paradox

Figure 8.1. Differences in expected Israeli prestige by location, for the six largest countries of origin, after controlling for human capital, region of the country, and year of arrival. Urban development towns and regular towns compared to the three cities.

(p.168) controls for human capital (Figure 8.1a), development town residence has a negative effect on the attainment of most country groups—it is associated with anywhere from 3 to 8 fewer Israeli prestige points—and has no apparent effect on that of Moroccans (the slightly positive effect is not statistically significant). However, for all countries of origin, controlling for human capital makes the coefficients for development and regular towns more positive, so that for most groups, type of urban settlement does not have a statistically significant effect on Israeli occupational prestige. This occurs because for all countries of origin, human capital was lower in development towns than elsewhere. Thus, although development towns offered fewer high status jobs, for the most part their residents didn't have the qualifications for higher status jobs anyway, and the net effect is neither loss nor gain.

Even this initial finding undermines Swirski's arguments, at least as they pertain to 1961. Rather than the towns generating downward mobility for Mizrahim and upward mobility for Ashkenazim, as he predicted, they didn't actually matter much at all. Only Yemenites—few of whom lived in the towns—obtained lower status in development towns than elsewhere, and Moroccans—who were the largest group in the towns and who were Mizrahi—obtained higher status in them. Moreover, not only does Figure 8.1 contradict Swirski's expectations, but it also appears to support his detractors'. The fact that human capital was lower in the towns than elsewhere is consistent with the argument that the state put weak immigrants in the towns, and the finding that controlling for human capital eliminates the effect of town residence for most groups and makes it positive for Moroccans tends to support arguments that the towns protected weaker immigrants by offering them jobs that were appropriate for their prior skills. Of course, one might still hypothesize that the towns generated ethnic inequality, but one would have to rely on dynamics that entered the picture after the settlement of the immigrants. For example, one might expect that in towns with so many weak immigrants and so little occupational opportunity, schools would be of low quality and that the psychological and practical effects of a poor educational system and limited occupational structure would reduce the next generation's chances for upward mobility.

This hypothesis regarding the next generation may well be correct, but the story of immigrant attainment is not that easily summed up. Note, in Figure 8.1, that controlling for human capital had relatively little effect on (p.169) the development town coefficient for Moroccans compared to other countries. This is because for Moroccans human capital was not much lower in the towns than outside. As I showed in Khazzoom (2005b), the state essentially sent two groups to the towns: Ashkenazim with low human capital, and all Mizrahim. The result is that the towns initially contained a fair number of Mizrahim—from all countries—with high human capital. Once placed, Asians and Ashkenazim were more likely to leave the towns than Moroccans, and Moroccans were more likely to move in than other groups (Adler suggests they were engaging in family unification; personal communication 2004). Thus by 1961, the towns had a relatively large number of Moroccan men with higher attainments; in fact, for all intents and purposes, the only development town residents to report western language primacy were Moroccan. While it may be true for other groups that the jobs available in the towns were appropriate to the immigrants' backgrounds, this was not necessarily as true for Moroccans.

Then why was Moroccan attainment increased by living in the towns? This is where prior work on development towns seems to have missed the critical dynamic. The people who benefited most from living in the towns were not the lower classes at all, but well-educated Moroccans, particularly if they were young. This is shown in Figure 8.2, which charts returns to education for Moroccan men in the three urban settlement types, for different ages at arrival, and with and without western language primacy (see equations in Appendix 23).7,8 As one can see from this figure, residential location mattered little for men with less than eight years of education.9 Among those with more than eight years, on the other hand, younger men were better off in the towns, especially if their primary language was Arabic rather than French.

For example, for twenty-year-old Arabic speakers with at least ten years of education, development towns provided more opportunity than even the three cities; they provided between 4 and 10 more Israeli prestige points than regular towns and about 5 points more than in the three cities. For older (forty-year-old) educated Arabic speakers, development towns were still better than regular towns, but the cities were best; development towns provided no more than 5 extra Israeli prestige points relative to regular towns, while the three cities provided up to 10 points more.

Separate regressions (not shown) confirm what Figure 8.2 implies, namely that Moroccan men had higher chances of obtaining white collar (p.170)

Residential Segregation and Economic Isolation: The Moroccan Paradox

Figure 8.2. Effect of residential location and age on Israeli prestige. For men with secular educations who do and do not speak western languages. Prestige abroad held constant at sample median (24.6), year of arrival set to 1956, region of the country set to north (with the exception of the three cities).

jobs outside of regular towns, net education and whether or not they had white collar jobs before immigration.10 They also confirm that for young men between the ages of twenty and thirty who did not report western language primacy and who had at least eight years of education, Israeli prestige was higher in development towns than in the three cities and that this was not true for similar men with western language primacy (Appendix 25).11 Clearly, development towns were associated neither with downward mobility for their residents nor with protection of weak Mizrahi immigrants, nor even with upward mobility for Ashkenazim. Rather, the most important role of development towns was to offer spaces in which younger, (p.171) well-educated Moroccans who did not report western language primacy could have relatively high chances at white collar jobs (protecting, perhaps, the weakest of the strong Mizrahim?).

So what is to be made of this dynamic? Answers are found not in prior work on development towns but in work on the segregation of Palestinians in Israel (who were relegated to more peripheral areas than Mizrahim), as well as U.S. work on segregation, labor market queues, and ethnic enclaves (Massey and Denton 1993; Wilson 1987, 1996; Lieberson 1980; Portes and Manning 1986). Their work is useful for this chapter because, unlike the work of Swirski and his detractors, they explain how segregation can sometimes have a positive effect on minority attainment, particularly the attainment of the middle class.

Racial Segregation and Immigrant Enclaves in the United States and Palestinian Segregation in Israel

The U.S. literature often treats enforced segregation and voluntary concentration of minorities as different dynamics, with the first disadvantageous to minorities and the second advantageous. However, Lewin-Epstein and Semyonov (1994) use this U.S. literature and the case of geographically marginalized Palestinians to argue that whenever ethnic minorities are concentrated, the same set of advantages and disadvantages apply. The disadvantage is that the higher the minority concentration, the lower the quality of the labor market. The advantage, according to their work and other work, is that concentration sets in motion four dynamics that give minorities better access to whatever jobs do exist. Those dynamics are as follows:

  1. 1. Networking, where minorities hear about jobs from friends who are from their home country or ethnic/racial group

  2. 2. Ethnic economies, where firms are owned by minorities and therefore more likely to hire minorities

  3. 3. Availability of clients for minority doctors, lawyers, and other professionals, who are often avoided by majority groups

  4. 4. Relief from labor market queues, where the reduced number of majority job seekers means that minorities have better chances at obtaining high status jobs than they would in areas with high majority concentrations

(p.172) Thus, wherever minorities are concentrated, there are fewer high status jobs, but minorities also have better access to whatever is there.

Both Wilson (1987, 1996) and Massey and Denton (1993), who examine the consequences of the segregation of African Americans, acknowledge that Harlem went through a golden age in which it supported a middle class, because whites by and large did not patronize African American professionals but African Americans did. However, in the main, both works argue that because of the involuntary nature and intensity of the segregation of African Americans, concentration could not be beneficial. In the Israeli case of development towns, segregation was also intense—some development towns were close to 100% Mizrahi—and involuntary—as I discussed in Chapter 3. Yet even under these conditions, the minority that was concentrated did better in its areas of concentration than in integrated centers.

Does a Queuing or Networking Dynamic Explain the Moroccan Results?

Lewin-Epstein and Semyonov (1994) would posit that as concentration increases, labor market quality would exert a downward effect on Moroccan attainment at the same time that the concentration itself would exert an upward effect. This can be modeled statistically by returning to the equation from Figure 8.1—which predicted attainment for all Moroccan immigrants and added terms for settlement type—and adding in terms that measure the ethnic concentration and labor market quality of an individual's town.

I measured labor market quality as average Israeli prestige of all men in the labor force in 1961. There are two possible ways to measure ethnic concentration—proportion Moroccan and proportion Mizrahi—and the contrast between them can indirectly assess whether the development towns' boost was due to a queuing dynamic or to a dynamic that is network based. In a networking dynamic, individuals who experience some sense of commonality—because they know each other, know people in common, come from the same town, etc.—hire each other or pass on information to each other about available jobs. As such, a networking dynamic would be manifested by a significant effect of the numbers of Moroccans who were in the towns. In contrast, in a queuing dynamic, it is not the high number of Moroccans that generates the higher attainment, but rather the low number (p.173) of Ashkenazim (or high number of other Mizrahim) (for the U.S. parallel, see Lieberson 1980). Thus, a positive effect for proportion Moroccan appears to indicate a networking dynamic, and a positive effect for proportion Mizrahi appears to indicate a queuing dynamic.

The results for this analysis (for urban areas and rural development towns) are in Table 8.1.12 As with the equations for Figure 8.1, these regress Israeli prestige on human capital, year of arrival, settlement type, and region. In the first equation, average Israeli prestige of a town is added. Once this has been done, Moroccan attainment in the three cities and regular towns are similar to each other and 3 to 6 prestige points lower than in development towns (the difference between urban and rural development towns is not statistically significant).13 In the second equation, proportion Mizrahi new immigrant and proportion veteran are added so that the comparison category is proportion Ashkenazi new immigrant. These variables reduce the coefficient for urban development town residence from a statistically significant 3 prestige points to an insignificant half point, and reduce the effect of rural town residence from a significant 6 points to an insignificant 2. This indicates that the prestige “boost” associated with development town residence is due to the higher proportions of Mizrahim in the towns. In the final equation, proportion Moroccan is added. It does not appear to affect the equation in any substantial way; thus the Moroccan attainment boost in development towns appears to be due to a queuing effect more than a networking one.

The Effect of Ethnic Concentration on Iraqi Attainment in High Opportunity Areas

If the same regression as Table 8.1 is estimated for the rest of the six largest countries of origin, the attainment of Iraqis also appears to be affected by the Mizrahi concentration of a town. In this case, however, the source of the dynamic is not a settlement type but a single city: Ramat Gan. Controlling for residence in Ramat Gan causes the effect of Mizrahi concentration on Iraqi attainment to be halved and reduced to insignificance. Controlling for other towns in which Iraqis were concentrated, such as Petah Tikva or Holon, does not affect the Mizrahi concentration term. Thus Ramat Gan is to Iraqis what development towns are to Moroccans, namely an area in (p.174)

TABLE 8.1 Regressions predicting prestige of individual's Israeli occupation, using human capital, settlement type, average prestige of Israeli occupation in town, and population distribution in town








Stand. Coeff.
















































































Tel Aviv







Urban development town







Rural development town







Regular town

Omitted category

Northern region







Southern region







Central region







City regions

Omitted category

Average Israeli prestige of town








Proportion Mizrahi new immigrant






Proportion veteran





Proportion Ashkenazi new immigrant

Omitted category

Proportion Moroccan







AG: Age at arrival; ED: Education; PA: Prestige abroad; YR: Year of arrival

(*) Significant to .05 level (2-tailed)

+ Significant to .10 level (2-tailed)

(p.175) which Iraqis were concentrated and also seemed to do particularly well. Moreover, in both cases, it was the more educated who benefited, through increased access to white collar jobs.

The case of Iraqis in Ramat Gan also turns out to be a useful comparison to the case of Moroccans in development towns, because it highlights one downward pressure on Moroccan attainment in the towns. Table 8.2 presents equations estimating the chances of obtaining a white collar occupation in Israel for men who had at least ten years of a secular education in development towns and in Ramat Gan (there were not enough cases to sort on western language). In development towns, where Moroccans did best, they still had a significantly lower chance of obtaining white collar jobs than other groups, net education, age, and having held a white collar position prior to immigration (Yemenites provided too few cases to be included in the equation). This is consistent with a queuing dynamic, in which Moroccans obtain higher prestige occupations not because they successfully competed with other groups but because there were too few members of other groups to occupy all valued positions. In Ramat Gan, in contrast, Iraqi men were more likely than other groups to obtain white collar jobs, and Iraqis had this advantage only in Ramat Gan (in Figure 5.1, this is manifested in the rise of the Iraqi line above Romanians and Poles at higher educational levels).14

Why might this occur? There is not enough research on the towns or on Ramat Gan to be certain, and future qualitative work might focus on the comparison. However, the dynamic brings to mind Portes and Stepick's work on Miami (1993). The Miami case was similar to the Ramat Gan case in that the human capital of the Latino (Cuban) immigrants concentrated there was relatively high and in that there, too, the minorities who were concentrated outstripped the attainment of majority members—even, argues Portes, creating a revitalization of the local economy. The dynamics to which Portes attributes this outcome are many and include physical capital—which immigrants could lend to each other to establish businesses—and networking.

The first explanation does not appear to apply to the Iraqi Ramat Gan case. It is of course true that Iraqi immigrants who moved to Ramat Gan often had physical capital (Khazzoom, in progress), and it may be that some used this capital to create businesses. It is also true that more Iraqis were self-employed in Ramat Gan than in other regular towns (7.2% versus 2.4%). (p.176)

TABLE 8.2 Logistic regression of chances of obtaining a white collar occupation in Israel on education, age, white collar incumbency, and country of origin (men with at least ten years of secular education)
















Had white collar occupation abroad






















Not enough cases



















Omitted category

Omitted category








However, by and large the set of gatekeepers from whom Ramat Gan Iraqis obtained jobs didn't differ from the set from whom Iraqis in other regular towns obtained jobs; in both cases, gatekeepers were mostly Ashkenazi and mostly veteran.15 Moreover, in both Ramat Gan and other regular towns about 90% of Iraqis were salaried (89.9% and 95.6%, respectively) and therefore were dependent on these mostly Ashkenazi gatekeepers for their occupational status. Na'im's story from the previous chapter is of course in line with these numbers, as he obtained his white collar job from a veteran Ashkenazi gatekeeper who noted his western behavior. The census contains no networking variables, so it is not possible to evaluate Portes' second dynamic. However, it is of note that the educational attainment of Ramat Gan Iraqis was particularly high: 9.1 years of education, on average, compared to 6.5 years in other regular towns and 6.2 in neighboring Tel Aviv. This does provide for networks that that would tend to lead toward high status jobs.

Thus a reasonable working hypothesis—which can be examined with other data—is that the Iraqi edge in Ramat Gan was due to the access Iraqis had to each other, i.e., to a network of immigrants who could pass on the kind of information necessary to obtain high status jobs, even from employers from a different ethnic group. Regarding development towns, then, the suggestion of the Iraqi case is that although concentration in the towns (p.177) did benefit Moroccans, that benefit was also limited by the lower human capital of the individuals who lived in the towns. This is suggested by the evidence that in both the case of Miami and Ramat Gan, high human and physical capital enabled minorities to turn concentration into a strong tool for interethnic competition for resources, producing advantages that are larger than those obtained by Moroccans in this sample.

So was there Discrimination?

The answer to the discrimination question is clearer in this chapter than in Chapter 7. First, as Khazzoom (2005b) made clear, there was discrimination in development town placement. Even Mizrahim who were of high human capital had a slightly higher chance of being placed in a development town than Ashkenazim of low human capital. Moreover, even though Moroccans with high educations did better in the development towns than they would have in other urban areas, that occurred only because there were fewer Ashkenazim around, and Moroccans still experienced a disadvantage relative to those Ashkenazim who were in the towns. Thus, despite the development town boost, the overall implication of this chapter and the preceding one is that Moroccans experienced discrimination, both inside and outside of the towns.

Implications for Dichotomization and the Formation of Ethnic Inequality

As with Chapter 7, examination of a social dynamic—in this case segregation—makes dichotomization less, rather than more, expected, though in this case the push against dichotomization had little to do with gatekeeper preferences. By placing Mizrahim from all countries at equal rates in isolated towns with lower opportunity, gatekeepers set in motion a process of dichotomization. But as it happened, Moroccans were more likely than other groups to immigrate during the years of high development town settlement, and Asians were more likely than Africans to leave the towns. As a result, the towns became centers of North African, and especially Moroccan, concentration, not Mizrahi concentration. Thus any effect of development towns on attainment would tend to disproportionately affect Africans over (p.178) Asians, undermining dichotomization. Moreover, in 1961, with the towns providing white collar jobs to Arabic-speaking well-educated Moroccans, and the cities providing jobs to French-speaking well-educated Moroccans, not only was dichotomization undermined, but so was any form of ethnic inequality in access to middle-class positions.

What was the Effect of Time?

As noted, one advantage of the Israeli case is that we can compare ethnic inequality at its formation with its later permutations. Adler et al.'s (2001) analysis, reviewed above, shows that development towns are now a disadvantage for their residents, in that residents' matriculation rates are lower, net background, than those of residents of other areas in Israel. Moreover, since Adler et al. found no ethnic differences in this dynamic, development towns seem to have shifted from having a positive effect on Moroccan attainment to having a negative one. Thus, time has changed the effect of segregation on Moroccan attainment, turning from positive to negative.

Though this shift has never been empirically examined, its sources would seem clear. Segregation is associated with lower labor market quality. Although causality is hard to prove, I did show that even in 1961, a town's proportion of new Mizrahi was associated with its labor market quality, even after the average human capital of a town had been taken into account (Appendix 20). Logic would suggest that gaps between labor market quality in segregated and nonsegregated areas would increase with time because high paying industries prefer areas that feel more “central” and, in the Israeli case, more modern and western. As the gap in labor market quality grows, the increased access to high status jobs in segregated areas becomes less of an advantage relative to the lower availability of such jobs. This provides a dynamic through which segregation itself lowers opportunity, over time.

In addition, Massey and Denton (1993) offer a second mechanism for how conditions in segregated areas decline over time, and this appears to apply to the Israeli case. Because minorities, be they African Americans or Israeli Moroccans, tend to lose their jobs first, segregated areas tend to experience higher rates of job loss during economic downturns. Individuals who have lost their jobs may stop taking care of their homes, giving the neighborhood a general feel of neglect. Both the middle-class and the lower-class (p.179) working population may move out of the neighborhood in search of better opportunities, robbing those left behind of social capital and practical resources (Sanchez-Jankowski 1997). In all these senses, concentration of Mizrahim into development towns would tend to increase ethnic inequality over time; however, by depressing Moroccan attainment it would also tend to undermine dichotomization. Again, the comparison between Iraqis in Ramat Gan and Moroccans in development towns appears important; because Iraqis were concentrated in a central area, in both a geographic and a cultural sense, the same decline over time should not occur.

Implications for the United States and other Areas

In this chapter, I used U.S. research to make sense of the Moroccan paradox, but what utility might the Moroccan paradox have for U.S. research? There are several answers. First, Israel provides an important test case for the assertion that democratic states can be active in producing segregation (Khazzoom 2005b). As noted, the data clearly indicate that gatekeepers placed Mizrahim in the towns at greater rates than Ashkenazim (see Figure 5.4). This is another instance in which Israel is interesting internationally because it can be caught engaging in direct ethnic discrimination at a time when such discrimination is theoretically expected by those who believe that “race matters.”16

Second, the Israeli case is interesting because the apparent reasons for segregation parallel Massey and Denton's (1993) contentions about the United States. They argue that blacks were segregated in large part because they were despised, that is, for Massey and Denton as well, the nonmaterial aspects of race and ethnicity are posited as having independent causal effects on the development of racial/ethnic inequality. The Israeli case is therefore another instance in which nonmaterial factors appear to account for moves to geographically marginalize a minority population.

Third, the Israeli case sits uncomfortably next to Massey and Denton's argument that societies interested in racial justice should focus first and foremost on desegregation. On the one hand, their contention that segregation is the direct expression of the devalued status of African Americans in U.S. society is echoed in the Israeli case. The sense of cultural difference and personal pain that such marginalization can generate pervades the (p.180) Go to the End of the World film with which I began this chapter, as well as the statements and activities of political organizations, from the Black Panthers of the 1970s to today's party Shas. On the other hand, in the Israel case the disadvantages of segregation are not in the segregation itself, but in the tendency of gaps in opportunity to increase over time. If the gains of segregation are always present and the losses are associated with degradation of labor markets, then in the continuing debate over whether to integrate (Massey and Denton) or economically develop (Wilson) already segregated areas, the answers appear at minimum complex.

This is in many respects a chicken/egg quandary, since both segregation and its negatives effects are rooted in racial/ethnic preference and both desegregation and economic development of segregated areas are hard to engineer as long as racial/ethnic preference remains in place. However, Israeli experiments with attracting wealthier Ashkenazi residents to the towns may be instructive. Several studies found that when the overall economic health of the town is the measure of success, this strategy brings more benefits than those focusing on attracting new industries to the towns. In particular, the new Ashkenazi residents do tend to spend their money in the town, and the overall quality of the schools increases. However, one of the more complete studies also found that the new Ashkenazi elite of the towns began to monopolize the political structure of the city, pushing out older Mizrahi residents (Ayalon et al. 1993). It may be that when the economic opportunities of local Mizrahi residents are considered, a more conservative strategy of job creation is more successful than suburbanization strategies. In addition, these programs designed to attract wealthier residents have not been the main trend in Israeli desegregation programs. It is easier to attract new Russian immigrants and other, less wealthy immigrants to the towns by offering them housing that is cheaper than that in the center. As a result, those who desegregate the towns tend to be lower in human capital themselves. Because I showed in this chapter that the average human capital of the town was also associated with lower market quality, and because one can imagine that this relationship is causal, attracting Ashkenazim with lower human capital may tend to eliminate the benefits of segregation for Moroccans without significantly altering the losses in labor market opportunity.

(p.181) Conclusions

On the surface, the finding that Moroccan attainment was boosted by residence in development towns contradicts the Moroccan assertion of discrimination by isolation. It suggests that rather than being a cause of ethnic inequality, Moroccan relegation to development towns was actually to their benefit. By reducing the number of Ashkenazim higher up in the queue, development town residence appears to have given Moroccans some relief from the discrimination they experienced in other areas and opened up opportunities for better educated men who did not have French primacy to obtain high status jobs. But the experience of the Iraqis in Ramat Gan also suggests that Moroccans in the towns were held back by the low human and physical capital that concentrated there. Moreover, we know that over time development town residence became associated with a lower tendency to matriculate, suggesting that over time their effect was negative, probably because of the increasing gaps in opportunity between segregated peripheral areas and less segregated central areas. Finally, I suggested that the gains made by segregated Moroccans in 1961 have complex implications for whether “development town development” should focus on desegregation or economic vitalization.

To the extent that development towns increased or decreased Moroccan attainment and not that of other groups, they functioned to undermine both ethnic inequality and dichotomization. This is not because gatekeepers didn't dichotomize or discriminate, but because of a complex set of historical accidents. Though all Mizrahim were placed in the towns at higher rates than all Ashkenazim, migration patterns resulted in the towns becoming Moroccan spaces. Since Moroccans with higher human capital but no ability to prove westernness had better access to high status jobs in the towns, ethnic inequality was mitigated, as was dichotomization. Moreover, with the towns becoming Moroccan spaces, any association between life chances and living in the periphery is likely to undermine dichotomization, as it will affect Moroccans differently from other Mizrahi groups.

Finally, it is worth pointing out that neither the theories addressed in this chapter nor the chapter itself attended to one of the central complaints from the movie: the Europhilic families experienced not only downward class mobility, but, from their standpoints, downward cultural mobility. (p.182) This omission on my part may seem odd, since one of the primary contributions of this book is to argue for the salience of the concept of “eastern” in shaping resource distribution in Israel, yet in this case the focus on resource distribution appears to hide rather than highlight processes of easternization of Mizrahi immigrants. The simple fact is that it is unclear how this cultural component could be examined quantitatively. My contribution in this book is to argue that Mizrahim were marginalized because they were seen as eastern, and that both residential and occupational marginalization are part of a single process of controlling the influence of the east in Israel. However, while this chapter's analysis implies a clear research project on the interactions of peripheral residence, ethnic concentration, and the distribution of resources over time, it is unclear what the parallel project would be about the distribution—or perhaps preservation—of easternness over time.


(1.) Taking the ten countries that produced the most post-1952 immigrants (when development town placement rose), I found that country-group probabilities of being placed in the town clustered into the binary configuration, with the probabilities for immigrants from Mizrahi countries higher than those for immigrants from Ashkenazi countries.

(2.) In the film, the school teacher stands in stark contrast to the plant manager; the teacher is an idealistic young man from Tel Aviv, who, in his composure and love of poetry, exemplifies the western image. In contrast to the manager's sexual harassment, and in compliance with Israeli support for miscegenation as the solution to the ethnic gap, the school teacher falls in love with the Francophile daughter and wants to marry her.

(3.) The western-appearing Ashkenazi, having finished his tour of duty, returns to Tel Aviv.

(4.) The towns were initially conceptualized as way stations for agricultural produce. When that didn't work, factories were established instead.

(5.) This study follows other Israeli work (Lewin-Epstein and Semyonov 1994) in conceptualizing Israel's segregation issue as one of settlement type. To the extent that researchers care about segregation because it removes minorities from “where the action is,” it was the settlement of people into development towns, not patterns of dispersion among neighborhoods within municipalities, that most isolated them from major currents in the developing society.

(6.) The equation on which this graph is based is in Appendix 22, and it compares attainment in all six settlement types, for each country of origin separately. As the equation in Appendix 22 shows, and as I explained earlier, Moroccans also did well in rural areas—in fact, better than in urban areas. This is an interesting result that needs to be examined separately. Because it was impossible to calculate ethnic concentration for most rural areas from the 20% sample of the census, and because of the political importance of the issue of Moroccan attainment in development towns, urban areas and development towns are the focus of this chapter. Nevertheless, it should be noted that the rural effect for Moroccans is larger than the development town effect and should be studied fully at a later date.

(7.) Some coefficients in the equation are not statistically significant. However, those coefficients are also small, so that Figure 8.2 does an accurate job of portraying the important contrasts.

(8.) This figure follows my standard practice of drawing lines only when at least fifty cases have the relevant combination of characteristics. However, I do not restrict the range to the middle 90% of the education distribution. Rather, I take each line to sixteen years of education, since the important dynamics are among the more educated, and since for all combinations represented there were at least some cases in the twelve- to sixteen-year range. For nearly all groups represented in Figure 8.2, the middle 90% of the distribution ranged from zero to twelve years. The main exception is young residents of the three cities who spoke western languages; their range was four to thirteen years.

(9.) The exceptions are young men who spoke Arabic and who had particularly low attainment in the three cities (and may well have been the group upon which state planners focused; see Segev's [1986] discussion of the Gelblum articles). Statistical significance for this dynamic is borderline, but an analysis of exact occupations of older and younger men in the cities suggests a pattern that makes sense. While older Moroccan men in Israeli cities tended to become janitors or salespeople (nonproprietors), younger men tended to go into blue collar occupations. This included both occupations that were not common for Diaspora Jews, such as longshoreman or construction worker, and those that were common, such as shoemaker. These blue collar occupational categories contain many with low prestige, such as unskilled construction worker. Thus, statistically, the prestige loss for younger men occurs because of the low scores of the unskilled jobs. Among older men, the prestige of the sales positions balances the lower prestige of the janitorial positions. Another apparent exception is that forty-year-old men who spoke western languages may have done somewhat better in the cities; however, few older men lived in the three cities. Attainment in rural towns is modeled in Appendix 22 but not represented in Figure 8.2. The effect of age on returns to education in rural development towns is statistically significant but has a large standard error. The estimate is that in rural development towns younger men at all levels of education obtained between 2 and 4 extra prestige points relative to urban development towns. However, probably the best way to read dynamics in rural development towns is to note that they are substantially similar to those in urban development towns, in that benefits accrued primarily to the young and more educated.

(10.) Note that development towns look good even in comparison to the cities because although the cities had twice as many white collar jobs to offer, Moroccan chances of obtaining one were not higher than in development towns.

(11.) The coefficient for development towns for non-western-language speakers is statistically significant only to the 0.10 level. However, the number of cases was quite small, increasing standard errors. Under these conditions, 0.10 might be considered enough. For those reporting western language primacy, the effect of (p.289) living in a development town is as large as the coefficient for those not reporting primacy, but this coefficient is not significant to any standard level.

(12.) I included only those settlements for which ethnic concentration and average Israeli prestige could be calculated (this includes all regular towns, the three cities, and nearly all urban and rural development towns, but includes no other rural areas). In addition, because ethnic concentration has little meaning in the cities, I included only Tel Aviv and left residents of Haifa and Jerusalem out of the sample.

(13.) Earlier equations suggested that Moroccans did better in the three cities than in regular towns, and this equation suggests that difference was because of labor market quality.

(14.) The Iraqi coefficient is significant only to the .07 level. But given the small number of cases, anecdotal evidence regarding Iraqis in Ramat Gan, and a series of other regressions that also suggest that Iraqis obtained particularly high educational attainments in Ramat Gan, it is reasonable to accept this level of significance.

(15.) Overall, 59.9% of employers and managers in Ramat Gan were veteran Ashkenazim, compared to a very similar 61% in other regular towns; 28.3% of employers in Ramat Gan were new immigrant Ashkenazim, as opposed to 29.2% in other regular towns; 3.0% and 2.7% of employers, respectively, were veteran Mizrahim, and 8.7% and 7.0% were new immigrant Mizrahim.

(16.) As I noted in Khazzoom (2005b):

In U.S. work, researchers differ on the extent to which they posit the state as instigating segregation. Massey and Denton (1993) arguably posit whites rather than governments as the primary actors; for them it was largely white real estate agents, supported by banks and threats of white violence, who created both the inner city and the system of redlining that was later picked up by the state. Gotham (2000), on the other hand, posits those real estate agents as one arm of the state. And Grannis (1998) implicates the state more directly when he argues that street patterns determine neighborhoods, and that state-employed urban planners determine street patterns…. Israel is interesting because the normal limitations on state intervention were momentarily suspended. Most sociologists posit that democratic modern industrialized societies, including both the U.S. and Israel, prefer the state to discriminate indirectly, for example by passing laws that benefit some groups more than others but that are not explicitly directed toward any group (Parkin 1979; see also Massey and Denton on government programs after the Depression). Even in Israel, direct government intervention in residential placement lasted only a brief moment…. Therefore that moment in which immigrants were placed (or allowed to escape from placement) by state workers provides a unique chance to ask the question: in a case in (p.290) which a state had unusual direct control over residential patterns, did it generate segregation?