Introduction

Immigrant labor market outcomes have long been a topic of considerable interest to researchers and policy makers alike. Nativity-based labor market disadvantages are well-documented in major immigrant-receiving countries including the U.S. (e.g. Beckhusen, Florax, Poot, & Waldorf, 2013; Tesfai, 2017) and Israel (e.g. Rebhun, 2010; Stier & Levanon, 2003). Researchers hypothesize that some of this disadvantage is due to immigrants’ difficulty assimilating to the labor market when they are not selected based on skills applicable to the host-country labor market (Cobb-Clark, 2003; Constant & Zimmerman, 2005). Public opinion and policy debates in the U.S. reflect this hypothesis. Americans strongly prefer high-skilled immigrants (Hainmueller & Hiscox, 2010) and—as a result—Congress has consistently put forward bills to further prioritize skill selection (Krogstad & Gonzalez-Barrera, 2018). Yet it is unclear if this type of immigration policy would actually benefit the United States. Few studies empirically test the impact of immigrant selection policy by comparing labor market outcomes of immigrants from a single origin in multiple destinations (Lewin-Epstein, Semyonov, Kogan, & Wanner, 2003). Furthermore, the literature addressing how immigrants’ race (specifically blackness) may impact the utility of immigrant selection policy is even smaller. Consequently, it is unclear whether selection policies lead to immigrant labor market success, even for immigrants who may experience anti-black racial discrimination in the host-country. This paper fills this gap by determining Ethiopian immigrants’ labor force participation (LFP), occupational status, and self-employment in the U.S. and Israel—countries with and without immigrant selection policies respectively.

Examining the impact of selection policy on one foreign-born group’s labor market success across place requires that the host-countries being compared have (1) similar labor market institutions and (2) an immigrant group that is relatively homogenous across place. As in Cohen and Haberfeld’s (2007) work, we base our comparison on the assumption that (Ethiopian) immigrants’ skills are about equally transferable to the American and Israeli labor markets. Both countries have similar economies/labor market structures (Larom & Lifshitz, 2018) in that they are considered free market economies, are members of the Organization for Economic Cooperation and Development (OECD), and both countries experienced consistent GDP increases over the past 25 years (OECD, 2016a).

Not only are the U.S. and Israeli labor markets similar, but also the timing and motivations for Ethiopian migration in both places are comparable. This provides a natural experiment enabling us to test the utility of immigrant selection policy. Ethiopian Jews began entering Israel en masse after their rabbinical recognition as Jews by the state of Israel in the 1980s (Chehata, 2012). Initially entering through three large-scale transfer missions, they migrated to Israel due to drought and famines in Ethiopia caused by the combination of food crises and political chaos during the Derg regime (Keller, 1992). Large-scale Ethiopian migration to the U.S. also began in the 1980s (Kent, 2007) due to famine and political instability (Chacko, 2003). Like Ethiopian Jews in Israel, Ethiopians in the United States primarily originate from Gonder—a battleground region that was destabilized by a series of military campaigns (Getahun, 2005). Accordingly, a large proportion of the Ethiopian populations in both countries have similar cultural characteristics, and many share family ties (Kaplan, 2010). The Ethiopian population in both countries has increased substantially since 1980. Ethiopian Jews have the right of return to Israel and are granted citizenship upon arrival and today, nearly all of Ethiopia’s Jewish population has migrated to Israel (i.e. there has been no selection of the Ethiopian population in Israel) (Bar-Haim & Semyonov, 2015; Shuval, 1998). In the U.S., the (predominantly Christian and Muslim) Ethiopian immigrant population has grown from 10,000 in 1980 to 178,000 in 2013 (Migration Policy Institute, 2014).

Ethiopian immigrants’ substantive population size and timing of migration in both the U.S. and Israel provide a unique opportunity to investigate the impact of immigrant selection policy on racial minority (black) immigrants’ labor market incorporation. There is a great deal of evidence showing that black immigrants experience considerable racial discrimination in host-country labor markets that negatively affects their employment patterns and likelihood of labor force participation. In the United States, employers prefer white applicants to black applicants generally (Moss & Tilly, 2003), but exhibit greater preference for black applicants when hiring for low-skilled jobs (Stewart & Perlow, 2001). Similar hiring patterns are observed in Israel (Blum, 2009; Epstein & Siniver, 2012). Limited employment options can impact whether immigrants seek employment (i.e. are active in the labor market) and the type of work they pursue. Poor labor market expectations due to discrimination may lead to individuals eventually leaving the labor force (the discouraged worker effect) (Dagsvik, Kornstad, & Skjerpen, 2016). Alternatively, some immigrants turn to self-employment as a strategy for economic success when they experience blocked mobility in the wage labor market (van Tubergen, 2005).

Discrimination experiences in the labor market are likely to vary by gender. Ethiopian men and women occupy different societal positions (Alonso-Villar, Gradin, & del Río, 2013) because race is constructed with gendered meanings (Browne & Misra, 2003). Specifically, negative stereotypes about racial minorities have more similarities with stereotypes of men than women of the same background (Ghavami & Peplau, 2012). Taken together, these findings point to a need for studies examining whether immigrant selection policy has a differential impact on the labor market outcomes of men and women.

Whereas immigrant selection policy is meant to improve immigrants’ success in the host-country, immigration to a racially stratified society may mean that racial boundaries impede upward mobility (Alba & Nee, 2003) for racial minority immigrants like Ethiopians. In this study, we compare Ethiopian immigrants’ integration in the labor market, using the outcomes of LFP, self-employment and occupational attainment, in the United States and Israel to that of the dominant and marginalized groups in each place as well as to that of white immigrants to answer the following questions: (1) Does the presence of an immigrant selection policy lead to more positive labor market outcomes among the foreign-born? (2) Does immigrant selection policy have the same impact on black immigrants’ labor market outcomes as it does for white immigrants? and (3) Does the impact of immigrant selection policy vary by gender?

Background

Immigrant selection policies are expected to improve immigrant labor market outcomes in the host country by selecting more highly skilled immigrants. Regardless of immigration history, most countries have policies in place regulating the size and composition of immigration flows (González & Miles-Touya, 2014). Since 1965, U.S. immigration policy has primarily focused on family reunification with a minority of admissions based on occupational skills (Duleep & Regets, 2014). While family-based policies do not select based on human capital, immigrants in the U.S. have similar levels of educational attainment as those in countries with a skills-based selection policy after controlling for country of origin (Kaushal & Lu, 2015). Like other immigrants, most Ethiopians entered the U.S. on family-based visas (Migration Policy Institute, 2014). Although family-based immigration suggests a rather lax level of human capital selection, researchers find that the U.S. attracts highly educated immigrants at the same rate as countries with skills-based immigration systems (Antecol, Cobb-Clark, & Trejo, 2003, Kaushal & Lu, 2015). This may be a result of the obligations involved in a family-based visa. Those entering on family-based visas must have a sponsoring relative in the U.S. who agrees to provide their relative with “any support necessary to maintain him or her at an income that is at least 125% of the Federal Poverty Guidelines for his or her household size... [if the sponsoring relative does] not provide sufficient support to the person... that person may sue [the sponsoring relative] for this support” (U.S. Citizenship & Immigration Services, 2019). This requirement means that family-based immigration policy selects immigrants based on socioeconomic characteristics in two ways. First, it selects immigrants who have financially stable relatives in the United States, thereby providing those entering the country with a springboard into a settled household with resources available to help them enter the labor market. Second, because the immigrants’ own economic resources can (and are) used in determining whether to grant a family-based visa, the policy indirectly prioritizes those with high-skills and/or wealth.

Although a large proportion of Israel’s population is foreign-born (Semyonov, Raijman, & Maskileyson, 2015), it is not a traditional immigrant host-country because of its immigration system. Immigration to Israel is restricted through the Law of Return which grants every Jew the automatic right to move to Israel and become a citizen (Kruger, 2005). Under this system, there is no selection based on any characteristic other than religion. While the Law of Return led to self-selection of some immigrant groups, this policy allowed the nearly complete movement of the Jewish Ethiopian population to Israel (Bar-Haim & Semyonov, 2015; Shuval, 1998)—a unique situation in which there was no selection within the prospective immigrant population.

Based on immigrant selection policy alone, we would expect Ethiopians in the U.S. to have more positive labor market outcomes relative to the native-born than their Israeli counterparts because family-based migration system selects Ethiopian immigrants based on socioeconomic characteristics. Yet, differences in immigrant economic success across countries may also occur due to variation in individual- and national-level characteristics (or host-country context). In labor market analyses, the most important individual-level characteristic considered is human capital and—among immigrants—migration characteristics such as time in the host-country also play a role. Aspects of the host-country context that are important in examinations of immigrants’ labor market outcomes include labor market institutions (Kaushal, Yao, Denier, Wang, & Trejo, 2016) and welfare regimes (Lewin-Epstein et al., 2003). Given the racial/ethnic stratification patterns in Israel and the U.S., we also expect racial discrimination patterns to play a role in Ethiopian immigrants’ labor market outcomes in both countries. In the sub-sections below, we describe each of these determinants of labor market outcomes.

Individual-Level Characteristics

Human capital theorists contend that labor market disparities between groups can be attributed to nativity/race differences in education and skill (Becker, 1985). Based on human capital theory, Ethiopian immigrants in Israel are not well placed for economic success. Most Ethiopians came to Israel from rural communities with low levels of education likely as a result of inadequacy in the provision of educational services (Ajala & Asres, 2008). While the high school completion rate of adults in Ethiopia is approximately 9% (The World Bank, 2020), educational attainment in Gonder (the region from which most Ethiopians in Israel and the United States originate) is much lower. A study in one area of Gonder finds that between 70 and 89% of adults with teenage children are illiterate (Eyasu, 2017). These statistics from Gonder largely match the characteristics of Ethiopian immigrants when they entered Israel: over 70% of Ethiopian immigrants arriving after 1990 were illiterate in Amharic (their home-country language) (Raijman, Semyonov, & Geffen, 2015) and did not have relevant skills for the Israeli labor market (Semyonov et al., 2015). With time in Israel, educational attainment increased and now 14% of Ethiopians have completed high school (compared to 56% of the general population (Habib, Halaban-Eliat, Shatz, & Almog, 2010)). Ethiopian immigrants’ educational attainment in the U.S. is much higher than that of Ethiopians in both Ethiopia and Israel providing further evidence for the U.S. family-based immigration system positively selecting Ethiopian immigrants. Ethiopian immigrants’ educational attainment is similar to that of the general U.S. population—approximately 30% have at least a college degree (Migration Policy Institute, 2014).

Education is a main stratifying factor in both labor markets. Therefore, inequalities in educational attainment are usually expressed in immigrants’ low LFP rates and lower likelihood of attaining high-status jobs (Semyonov, Raijman, & Maskileyson, 2016). Kanas, van Tubergen, and van der Lippe (2009) also find that lower educated immigrants are more likely to be self-employed. Yet labor market disparities remain significant after controlling for human capital characteristics in both the U.S. and Israel (Beckhusen et al., 2013; Offer, 2004), suggesting that human capital alone does not explain immigrants’ labor force disadvantages.

One reason that human capital does not adequately explain nativity-based disparities is that migration related characteristics also play a role in immigrants’ labor market experiences. Immigrants are initially disadvantaged by imperfect information about the labor market. Specifically, workers may have lower LFP rates or take low-status employment because they have inadequate information about the local labor market (Tani, 2012). Classic assimilation theory predicts that with time in the host-country, ethnic differences between the native- and foreign-born will decline and immigrants will experience socioeconomic advancement (Anderson & Massey, 2001). In large part this occurs because—with time in the host-country—immigrants gain information regarding their occupational options helping them find jobs that are commensurate to their qualifications. Yet while immigrants’ labor market outcomes improve over time (Beckhusen et al., 2013), they do not consistently reach parity with the native-born (Thomas, 2010).

Another explanation for the persistent immigrant disadvantage may be related to imperfect transferability of credentials. In both the U.S. (Adamuti-Trache, 2014) and Israel (Remennick, 2012), employers have difficulty evaluating foreign diplomas/work records. With few resources to aid in this process, immigrant skills are underutilized. These issues are not specific to the U.S. and Israel. Evidence from Canada (Aydede & Dar, 2016; Oreopoulos, 2011), Europe (Aleksynska & Tritah, 2013) and Australia (Clarke & Skuterud, 2016) all show that foreign credentials are discounted leading to immigrant labor market disadvantage.

Given the consistency in the impact of immigration characteristics (i.e. time in the host-country and imperfect transferability of skills obtained in the host-country) across place, disadvantages associated with migration is likely similar in the U.S. and Israel. At the same time, it may not be similar among men and women. For women, migration related characteristics may be related to gender norms in the home and host countries. Immigrant women from countries with more traditional gender roles are less likely to work after migration, especially if they were tied movers (He & Gerber, 2020; Rebhun, 2008). In both the home and host country, childcare is often considered women’s work. As a result, women are less likely to be in the labor force when they have young children, but the strength of this relationship varies by race/ethnicity (Omori, 2016; Stier & Levanon, 2003; Stier & Yaish, 2008): the presence of young children has the smallest impact on black women’s labor force participation (England, Garcia-Beaulieu, & Ross, 2004). The effect is also smaller if women live close to both their mothers and their mothers-in-law (Compton & Pollak, 2014). Because immigrant women are less likely to live close to their parents, they may also be less likely to be in the labor-force than the native-born.

National Level Characteristics

Labor Market Characteristics and Welfare Policy

To accurately compare the impact of immigrant selection policy on Ethiopians’ labor market success in Israel and the U.S., the two countries must have similar labor market institutions and welfare policies. The U.S. and Israel have very similar employment characteristics: men’s and women’s unemployment rates in Israel and the US have been nearly identical since 2007 (5% among men and 4% among women) (Larom & Lifshitz, 2018) and both countries have relatively high levels of foreign-born employment rates (OECD, 2016b). Research examining immigrant women’s labor force participation in the United States finds that immigrant women have much lower participation than the native-born (McManus & Johnson, 2020) and refugee women are less likely to participate in the labor force compared to foreign-born women who entered on other types of visas (Vijaya, 2020).

These employment and labor force participation patterns reflect larger similarities in the U.S. and Israeli labor markets. Israel and the U.S. now have nearly equivalent public expenditures on active labor market policies that are lower than all other major immigrant receiving countries, spending 0.2% and 0.1% of their 2011 GDP in Israel and the U.S. respectively (OECD, 2015). Furthermore, both Israel and the U.S. have legislation guaranteeing a minimum wage (Center on Budget & Policy Priorities, 2018; Lurie, 2015) and have similar tax policies: Israel’s negative income tax (introduced in 2007) is comparable to the U.S. earned income tax credit (Larom & Lifshitz, 2018).

Where the U.S. and Israel seem to differ are in their welfare policies. Immediately after the state of Israel was established, Israel adopted a European social security model (Adler & Avgar, 2019). Despite its status as a welfare state, Israel’s social expenditures per capita are consistently lower than the United States (Razin, 2018). In 2018, Israel spent less on social programs than all other major immigrant receiving countries (16% of their GDP) (OECD, 2020a). Because spending in the U.S. was quite similar (18.7% of the GDP) (OECD, 2020a), the United States and Israel can be grouped together in the Esping-Anderson typology of social policy families (Shalev, Gal, & Azary-Viesel, 2012). Since 2002, the welfare and benefit system drastically cut income support and unemployment benefits (Larom & Lifshitz, 2018). With these changes, Israeli unemployment benefits are more similar to those provided in the U.S.—0.28% of the Israeli GDP compared to 0.19% of the U.S. GDP (OECD, 2020b)—with the main difference being that in Israel, benefits are provided for the duration of unemployment (National Insurance Institute of Israel, 2019) while there is a time limit (that varies from state to state) in the U.S. (Office of Unemployment Insurance, 2018). These changes in welfare policy may explain why the United States and Israel have nearly identical Gini coefficients (0.377 and 0.375 respectively) and are ranked as the second and third most unequal OECD countries respectively (Gornick & Jäntti, 2014).

Racial Stratification

Due to the similarities between the United States and Israel in terms of the impact of immigration-related characteristics and host-country labor market and social policies, we argue that immigrants’ race, specifically blackness, may impact the utility of immigrant selection policy. For example, queuing theory states that when applicants apply for a job, employers rank order these individuals based on their evaluation of the applicants’ race/nativity (Lieberson, 1980). There is a great deal of evidence of race/ethnic discrimination in the Israeli and U.S. labor markets. Therefore, testing whether selection policies lead to immigrant labor market success, should account for whether immigrants experience anti-black racial discrimination in the host-country.

Israeli society has consistently been organized according to ethnic-national lines. However, rather than a color line, the primary divide is between Palestinians and Jews, with Palestinians occupying the subordinate position (Lewin-Epstein & Semyonov, 1992; Sa'di, 1995). Ethiopians formed the first black community in Israel and the discourse of race and racial inequalities in Israeli sociological and immigration studies emerged only after their arrival (Elias & Kemp, 2015). While not explicitly stated, race may have played a role in their integration process since they are the only immigrant group who were subject to high levels of institutional control (Offer, 2004). Specifically, while immigrants from the Former Soviet Union (FSU) entered Israel during the same period and were free to choose where to live and work (Fenster, 1998), Ethiopians were sent to absorption centers (housing/training/educational facilities) immediately upon arrival (Chehata, 2012; Offer, 2004). Although absorption centers helped Ethiopians gain skills, scholars note that these integration policies were patronizing due to their basis in ethnocentric cultural theories. For example, only men were sent to vocational training programs and if mothers went to work, they were accused of neglecting their children (Hertzog, 2001). Ethiopian culture was frequently described as “primitive” and the extended process of “absorption” left Ethiopians secluded from the rest of Israeli society (Kurman, Eshel, & Zehavi, 2005). Even after leaving absorption centers, re-settlement policies left Ethiopian immigrants highly segregated (Habib et al., 2010).

Despite Israeli settlement policies, Ethiopians have lower rates of LFP than other Jewish groups (Habib et al., 2010) and are also more likely to work in low-skilled or unskilled occupations (Offer, 2004). Possibly due to their exclusion from work training in absorption centers, Ethiopian women in Israel often work in jobs at the bottom of the socioeconomic ladder (Fanta-Vagenshtein & Anteby-Yemini, 2016). Ethiopians’ labor market characteristics persist after controlling for human capital; therefore, they have been attributed—at least in part—to institutional discrimination and everyday racism (Amit, 2012; Ben-Eliezer, 2004, 2008). In Israel, 53% of employers state that they prefer not to hire Ethiopians and 70% tend not to promote them (Blum, 2009).

Individuals who experience these kinds of significant barriers to employment may leave the labor force entirely because they do not expect to be hired (Dagsvik et al., 2016). Among those who remain in the labor force, low-status immigrants are commonly confined to the bottom rungs of the job ladder (Alba & Foner, 2015). Racial minority status also impacts Ethiopian self-employment in Israel. Ethiopian self-employment in Israel has long been seen as nearly non-existent (Kayam, 2014; Offer, 2004). Like Palestinians in Israel (Shavit & Yuchtman-Yaar, 2001), many Ethiopian immigrants may be unable to turn to self-employment because their community is quite poor and demands fewer goods and services than other Jews. Without patronage from other Ethiopians, potential entrepreneurs are more likely to suffer from discrimination because other Jewish groups may prefer to purchase goods/services from non-black Jews.

Ethiopian immigrants to the U.S., however, face a very different reality. Unlike in Israel, all Ethiopians do not receive re-settlement benefits. Out of the estimated 178,000 Ethiopian immigrants living in the U.S. in 2013, 48,600 were admitted as refugees (Migration Policy Institute, 2014)—the only immigrants that receives benefits immediately upon arrival in the U.S. Benefits vary across states, but the Office of Refugee Resettlement provides short-term cash and medical assistance to new arrivals, as well as services such as English classes and employment services during their first 8 months in the U.S. (Office of Refugee Resettlement, 2018). Despite the availability of these benefits to a subset of the Ethiopian population, they may not be sufficient to overcome obstacles in the U.S. labor market. Elo, Frankenberg, Gansey, and Thomas (2015) show that refugees from Africa actually have worse economic outcomes than those who entered on other types of visas: Sudanese and Somali immigrants (who primarily entered the United States as refugees) earn significantly less than other African immigrants.

Ethiopian immigrants in the U.S. also differ from Ethiopians in Israel because they joined a long-standing black population rather than comprising the entirety of the black population. Black individuals in the U.S. are surrounded by an impermeable color line (Sears, 2006) that some claim renders nativity unimportant (Waters & Eschbach, 1995). For black immigrants like Ethiopians, the fact that U.S.-born blacks face racial inequality and discrimination functions as a barrier to their full assimilation into the U.S. labor market (Anderson & Massey, 2001). Qualitative research in Washington D.C. (a major Ethiopian settlement area) shows that Ethiopian immigrants see themselves as a separate ethnic and racial category from U.S.-born blacks (Habecker, 2012). Yet despite their effort to maintain social distance from black Americans, Ethiopian immigrants experience similar labor market disadvantages as U.S.-born blacks in that they are unable to find jobs commensurate with their skills/experience (Habecker, 2012). These disadvantages seem stronger among African men than women: the effect of race on earnings is negligible among women, but significant among men (Nawyn & Park, 2019).

Restrictions in wage employment leads to higher rates of self-employment among the foreign-born: marginalized immigrants are more likely to use self-employment as a means to escape non-employment in the wage labor market (Blume, 2009). Ethiopian immigrants in the United States are no exception; they choose to become self-employed largely due to blocked mobility (Price & Chacko, 2009). Together, the research suggest that Ethiopians, like other black immigrants, are placed at the bottom of the U.S. racial hierarchy along with U.S.-born black individuals (Tesfai, 2017).

In this overview of Ethiopians’ individual-level characteristics and the policy environments in Israel and the U.S., two characteristics stand out as particularly salient in analyzing the impact of immigrant selection policy on Ethiopians’ labor market outcomes: human capital and race. The impact of these factors on labor market integration, in addition to that of immigrant selection policies, will be examined in our statistical models through various comparisons between Ethiopian immigrants and native-born groups as well as white immigrants. The overview also points to similarities between the US and Israel in labor market characteristics and welfare systems thus ruling out these factors as major explanations of the differences in immigrants’ integration across the two countries. Instead, an examination of whether immigrant selection policies in fact determines the level of the human capital of immigrants—which in turn impacts labor market integration—is required. Although the United States uses a family-based migration system, evidence from previous research shows that it serves to select immigrants with similar human capital characteristics as countries with merit-based visa programs. Thus, using descriptive statistics we first examine the educational attainment of Ethiopian immigrants in both countries to determine if this is indeed the case.

Our main research questions target the impact of immigrant selection policies on immigrants’ labor market integration. To examine the impact of selection policy we compare Ethiopian immigrants’ integration in two different countries, with and without selection policies. In particular, our models compare Ethiopian immigrants’ labor market outcomes with that of the native-born most dominant group in the labor market within each country. In the U.S., they are compared with native-born non-Hispanic whites, and in Israel they are compared with Ashkenazi Jews. In the absence of other major factor impacting immigrants’ integration, immigrant selection policies based on human capital would mean that Ethiopians in the U.S. would have better labor market outcomes relative to the native-born dominant group than Ethiopians in Israel. Conducting these within-country comparisons while controlling for educational differences between Ethiopians and these dominant groups could indicate whether factors other than human capital also impacts immigrants' integration.

We argue that, in addition to selection, the existence of a racially stratified society has significant implications for immigrants’ assimilation. Therefore, our models compare Ethiopian immigrant's outcome not only with the dominant group in each country, but also with marginalized groups and white immigrants. The presence of a racially stratified society implies that immigrants assimilate into particular races (Anderson & Massey, 2001). In the U.S., this means that Ethiopians’ are likely to assimilate into—and thus are also compared to—U.S.-born black individuals. Ethiopians in the U.S. likely have lower LFP rates and occupational status than even U.S.-born black individuals given the combination of their race and nativity. However, due to their marginalization in the wage labor market, Ethiopians in the U.S. may have higher rates of self-employment than the native-born. In Israel Ethiopian immigrants did not join an existing black community, and it is unclear whether their membership in the Jewish majority shields them from the most marginalized position in the Israeli stratification system, historically occupied by Palestinians. Therefore, Ethiopians’ labor market outcomes are compared not only to those of Ashkenazi Jews, the most dominant groups, but to Palestinians, the most marginalized group. In the absence of racial discrimination, we expect that Ethiopians in the U.S. would have better labor market outcomes relative to the native-born than Ethiopians in Israel.

The examination of the impact of blackness is strengthened by including a third comparison—white immigrants. To rule out that any disadvantages of Ethiopians in the labor market represent nativity effect only, Ethiopians are also compared with white immigrants from the FSU who arrived to the U.S. and Israel at the same period. In the absence of discrimination, we expect within country labor market differentials between Ethiopian and white immigrants to be mainly attributable to differences in human capital. Once we control for human capital characteristics, differences should be minor. Because our review indicates possible gender differences in the impact of immigrant selection policies and racial stratification on labor market outcomes, our comparisons are conducted for men and women separately.

Our comparisons use three different labor market outcomes because we expect differential effects of the examined factors on these outcomes. In particular, our review suggests that immigrants’ disadvantages in the labor market are evident mainly in lower LFP and occupational status. Therefore, our analyses will determine whether Ethiopian immigrants have lower LFP rates and lower occupational status than any other group in the host country. We also include comparisons in self-employment because we expect a different pattern of results on this outcome. Previous research suggests that racial discrimination has opposite effects on Ethiopian self-employment in the United States and Israel, occupying the lowest status position would result in the lowest self-employment rates in Israel (where a small community of black immigrants cannot support a large number of self-employed persons who mainly provide services to the community) and higher self-employment rates in the U.S. (where there is a more concentrated population to purchase goods and services).

Data and Methods

We utilize two national level datasets for the analysis. The U.S. dataset is comprised of the pooled 2012–2016 Annual Social and Economic Supplement of the Current Population Survey (Flood, King, Ruggles, & Robert Warren, 2017). Israeli data comes from the combined 2012–2016 Israeli Labor Force Survey (Central Bureau of Statistics, 2017). Using these datasets, we compare Ethiopian-born individuals to native-born dominant and marginalized groups. Similar labor market performances of Ethiopian immigrants and the most marginalized groups in each country may not be a result of anti-black discrimination. It is also possible that immigrants—regardless of race—experience discounting of their skills or are unaware of relevant labor market resources. Therefore, we include another comparison group: FSU immigrants. Immigrants from the Former Soviet Union began entering the U.S. and Israel in large numbers at roughly the same time as Ethiopian immigrants (Cohen & Kogan, 2007; Mehta & Elo, 2012) and—like Ethiopian immigrants—are a more selective group in the United States than Israel (Cohen & Haberfeld, 2007). By including this white immigrant group, we are more able to reliably test whether the impact of immigrant selection policy varies by race. Both samples exclude those employed in the armed forces and are limited to individuals in the working age population (age 25–65).

Our dependent variables are LFP, occupational status, and self-employment. LFP is defined as either working or actively looking for work. Occupational status is measured using the International Socio-Economic Index (ISEI) (Ganzeboom, De Graaf, & Treiman, 1992). Derived from the International Standard Classification of Occupations, the ISEI ranks occupations on a 1–100 scale based on the mean education and income of each job’s incumbents. High-status jobs—those that require high educational attainment and pay high wages—have higher scores on the ISEI. For the Israeli data, ISEI is calculated and provided by the Central Bureau of Statistics in the Labor Force Survey dataset. For the U.S., we utilize the 2010 U.S. census four-digit occupational classification code and build a crosswalk between this U.S. census definition and the 2008 International Standard Classification of Occupations.

The comparison of foreign-born Ethiopians’ LFP, occupational status, and self-employment to that of FSU immigrants and native-born groups is implemented using propensity score matching. Due to gender differences in immigrants’ labor market experience (Adsera & Chiswick, 2007; Stier & Levanon, 2003) and research showing that racial minority men and women have different experiences with labor market discrimination (Browne & Misra, 2003), we conduct separate analyses by gender. We also conduct comparisons of occupational status by skill-level because aggregate analyses may mask differences within skill-levels. We define two skill-levels based on educational attainment: low-skilled is defined as having a high school degree or lower and high-skilled as completing at least some college. Finally, we also examine occupational status differences by self-employment status because limited occupational options can lead to immigrants choosing self-employment.

The propensity score (PS) is defined as the probability of being an Ethiopian immigrant (vs. being a member of a native-born group/FSU immigrant) given background variables. The PS is estimated using a logistic regression model. Matching was applied using the teffects psmatch command in Stata 15.1 software. Using the estimated propensity score, each Ethiopian immigrant was matched with the most similar native-born/FSU immigrant individual(s). Once matched, we estimated the average treatment effect on the treated, that is, the effect of being a black immigrant on Ethiopian immigrants in each of our three outcomes.

The background variables we control for (the variables that predict the PS) include individual-level, household-level, and context variables that play a role in labor market experiences. Because family characteristics play a significant role in labor market experiences, we include marital status, and number of children below age 5. Women’s LFP is lower among those who are married (Azmat, Güell, & Manning, 2006), and having young children has a significant association with both men and women’s LFP (Aaronson et al., 2014). In comparisons with FSU immigrants, we also include years in the host-country. In addition, we control for survey year and regional variables: state and metro areas in the U.S. and district in Israel. Finally, in analyses of occupational status, we also control for employment status (full or part time). Although local language ability plays an important role in immigrants’ labor market success, direct questions about mastering the host-county’s language are not available in the data from either country, therefore local language ability is not included as an independent variable.

Previous studies also identify the importance of age, education and host-country labor market experience as factors affecting immigrants’ assimilation. We find, however, that accounting for these three variables at the same time creates a problem of comparability (or lack of overlap) between immigrant and native-born groups. For example, 40-year-old immigrants who earned a university degree in their country of origin and immigrated at age 35 have approximately 5 years of experience in the host country’s labor market. By contrast, 40-year-old native-born individuals with a university degree are expected to have about 15 years of experience in the labor market. We used the PS to examine the degree of comparability between immigrants and the native-born that results from controlling for these three variables at the same time; the distribution of the PS that is estimated using these three variables among Ethiopian immigrants and native-born black men in the U.S. is presented in Fig. 1. The lack of overlap between groups before matching is obvious (see boxplots on the left of Fig. 1): only outliers in the U.S.-born black population overlap with the body of Ethiopian men’s boxplot. Even when comparing Ethiopian immigrants to only the sub-sample of U.S.-born blacks who are most similar to Ethiopian immigrants, the lack of overlap between the two matched groups remains very high (see boxplots on the right of Fig. 1). This result points to a serious problem of comparability when age, education, and host-country labor market experience are included as matching variables; thus, we control for two of the three variables (age and education). Similarly, in models that compare Ethiopians to Palestinians in Israel we do not control for regional variables. Due to the high spatial segregation of Palestinians from the Jewish population, the likelihood of “finding” Ethiopian immigrants and Palestinians with similar socio-demographic profiles in the same regional unit is extremely low.

Fig. 1
figure 1

Distributions of estimated propensity scores for Ethiopian immigrant, U.S.-born black men, and matched U.S.-born black men, controlling for age, education and host-country work experience

Full size image

The implemented matching procedure avoids the pitfalls described above and succeeds in removing the initial imbalance in the raw data between Ethiopian immigrants and each of the comparison groups to create two balanced, matched samples. This result is demonstrated by comparisons of the distributions of estimated PS among Ethiopian immigrants and native-born comparison groups (before and after matching) and is presented in Fig. 2a and for the United States and 2b for Israel. In addition to these graphic demonstrations of comparability, we conducted a formal test of covariate imbalance using the measure of absolute standardized difference in means (Rosenbaum, 2010). The results of these analyses (not shown) show lower post-matching standardized differences, indicating a reduction in imbalance.

Fig. 2
figure 2

a Distributions of estimated propensity scores for Ethiopian immigrants, Native-born comparison groups and matched comparison groups, for men and women in the United States. b Distributions of estimated propensity scores for Ethiopian immigrants, Native-born comparison groups and matched comparison groups, for men and women in Israel

Full size image

Results

Descriptive Statistics

Table 1 presents descriptive statistics—including crude estimates of LFP rates, self-employment, and mean occupational status—by country and for men and women separately. We find very little difference between Ethiopian men and women’s mean age, marital status, and childbearing in the U.S. and Israel. Not only are Ethiopian immigrants in the U.S. and Israel similar to one another, they are also both younger than the native-born and FSU immigrants in their host-country, less likely to be married than the native-born majority, and more likely to have young children than nearly all other groups in their host-country. These differences between Ethiopian immigrants and the native-born highlight the importance of including family characteristics (marital status and number of children below age 5) in our matching procedures for both men and women. As expected given immigrant selection policy, the two groups of Ethiopians (and FSU immigrants) differ in educational attainment and age at migration. Ethiopian immigrants in the U.S. are more highly educated than those in Israel—men and women in the U.S. have 5 and 6 years more schooling respectively than their counterparts in Israel. Ethiopians in the U.S. also have similar levels of schooling to U.S.-born whites and higher educational attainment than U.S.-born blacks. Likely due to immigrant selection policies regulating entry, Ethiopian immigrants enter the U.S. at older ages (mean age of 28) than they do in Israel (age 20). Consequently, Ethiopians in Israel have been in the host-country longer than Ethiopians in the U.S.

Table 1 Descriptive statistics—individuals (age 25–65) in the U.S. and Israel, 2012–2016
Full size table

While there are few differences in Ethiopian immigrants’ individual and household characteristics across place, there are major difference in our labor market outcomes of interest. We find that Ethiopian immigrant men in the U.S. are more likely to be in the labor force than their Israeli counterparts—94% are in the labor force compared to 85% of Ethiopian men in Israel. Not only are Ethiopian men’s LFP rate in the U.S. higher relative to Ethiopian immigrants in Israel, it is also higher than both U.S.-born whites and blacks who have LFP rates of 84% and 73%, respectively. Ethiopian men in Israel, by contrast, have lower LFP rates relative to Ashkenazi Jews (though higher than Palestinians).

Despite high LFP, Ethiopian men in the U.S. work in lower status jobs than U.S.-born whites but the same status as U.S.-born blacks. They are also much more likely than U.S.-born white and black men to be self-employed suggesting occupational marginalization. Compared to Ethiopian men in the U.S., Ethiopian men in Israel work in lower status occupations and are much less likely to be self-employed. Their mean occupational status and self-employment rates are also lower than both Ashkenazi Jews and Palestinians. Mean occupational status and self-employment patterns among women both within and across countries are similar to those observed for men. The main gender difference is that Ethiopian women in the U.S. participate in the labor force at a rate (64%) that is not only lower than that of U.S. born white (72%) and black women (70%), but also lower than Ethiopian women in Israel.

PS Matched Analyses

Labor Force Participation

Focusing first on gaps in LFP among men, we find that Ethiopian men are significantly more likely to be in the labor force than native-born dominant and minority groups in both the U.S. and Israel. In both countries, the odds of LFP are between 2.6 and 4.2 times higher among Ethiopians than the native-born (Fig. 3). Ethiopian men also have higher odds of LFP than FSU immigrants in the US (odds ratio 5.8, p < 0.05) and Israel (odds ratio 2.5, p < 0.05).

Fig. 3
figure 3

Odds ratios and confidence intervals of labor force participation, Ethiopian immigrants vs. comparison groups, USA and Israel 2012–2016

Full size image

For women, the LFP advantage relative to the native-born holds for Ethiopians in Israel; they have higher odds of LFP than Ashkenazi women (odds ratio 3.1, p < 0.05) and Palestinian women (odds ratio 7.9, p < 0.05). They are also more likely to participate in the labor market than FSU immigrant women (odds ratio 1.9, p < 0.05). The same is not true in the United States. Ethiopian women are less likely to be in the labor force than native-born black women (odds ratio 0.51, p < .05) and their LFP is not significantly different from those of U.S.-born white or FSU immigrant women.

The difference in Ethiopian women’s labor force participation may be due to differences in migration characteristics. Ethiopian women in Israel have lived in the host-country longer than their counterparts in the U.S. and previous research shows that as time elapses in the host country, they are able to close the gap with native-born (Rebhun, 2010; Semyonov et al., 2015). In addition to living in the host country longer, Ethiopian women in Israel are also more likely to have family support available for child care. Unlike Ethiopians in the United States, Ethiopian women in Israel migrated with their entire extended family and are more likely to have parents or other family members available to help with child care.

Occupational Status

While Ethiopians’ high LFP relative to the native-born is an important aspect of work experience in the host-country, the types of jobs Ethiopians attain is an important indicator of their labor market incorporation. Table 2 presents mean differences in occupational status scores (Ethiopian score minus comparison group score) in the aggregate and by skill level. Focusing first on aggregate level analyses (leftmost column), we find that Ethiopian immigrants in both the U.S. and Israel work in lower status jobs than all other groups in their host-country. In the U.S., Ethiopian men seem to experience a unique disadvantage related to both nativity and race. Their mean occupational difference from U.S.-born blacks is − 7.5—approximately half the occupational difference from both U.S.-born whites and FSU immigrants. In Israel, Ethiopian men have the largest difference from the dominant Ashkenazi group (mean score difference = − 10.8) and the smallest difference from the most marginalized Palestinians (mean score difference = − 1.7). Unlike the United States, the mean occupational difference from FSU immigrants is closer to the marginalized group than the dominant group. Similar to men, we find that Ethiopian women also experience significant occupational disadvantage. The disadvantage is largest relative to the dominant group (mean occupational difference of − 14.8 in the U.S. and − 17.2 in Israel), but still significant relative to the marginalized groups and FSU immigrants in each country.

Table 2 Propensity score matching results measuring differences in occupational status between Ethiopian immigrants and native-born groups, 2012–2016
Full size table

Overall, analyses in the aggregate show a consistent Ethiopian occupational disadvantage. When analyses are conducted by skill, we find that Ethiopians’ occupational disadvantage characterizes both the high skilled and low-skilled in nearly all comparisons in both countries. Among men, the disadvantages among highly skilled Ethiopians are even larger than that among the low-skilled. For example, there is no significant difference in the mean occupational scores of low-skilled Ethiopians compared to low-skilled U.S.-born blacks and FSU immigrants. Among the highly skilled, by contrast, we find significant mean occupational differences of − 11.6 and − 12.1 relative to U.S.-born blacks and FSU immigrants respectively. While we also find larger occupational disadvantages among high-skilled Ethiopians in Israel, the differences by skill are much smaller than in the U.S.

We find a different pattern among women. Ethiopian women in both countries work in occupations of significantly lower status than all comparison groups. Occupational differences are smaller in Israel than in the United States for all skill levels. However, rather than larger gaps among the highly skilled, we find lower occupational differences between Ethiopians and all comparison groups except Palestinians.

Self-employment

The final labor market outcome we investigate is likelihood of self-employment (Fig. 4). Focusing first on the U.S. we find that Ethiopian men are approximately twice as likely as U.S.-born white men (OR 1.7, p < 0.05) and four times more likely than U.S.-born black men to be self-employed (OR 4.1, p < 0.01). In Israel, by contrast, Ethiopian men are less likely than all other groups to be self-employed (odds ratios ranging between 0.14 and 0.15, p < 0.01). Among women, however, we find that Ethiopian women regardless of host-country are less likely to be self-employed than the native-born (though the differences are larger in Israel).

Fig. 4
figure 4

Odds ratios and confidence intervals of being self-employed, Ethiopian immigrants vs. comparison groups, USA and Israel 2012–2016

Full size image

Although our analyses of occupational status show that Ethiopian men and women work in lower status occupations than the native-born and FSU immigrants, occupational status may vary by type of employment. Evidence in the U.S. shows that immigrants turn to self-employment when unable to find work in the wage-earning economy. To determine if some of the occupational status gaps observed in Table 2 are due to type of employment, we estimated mean differences in occupational status by self-employment status (Table 3). Due to the small proportion of women who are self-employed, differences in occupational status by type of employment are presented for men only. In the U.S., Ethiopian immigrants’ occupational status disadvantage relative to U.S.-born whites and blacks is larger among the self-employed than among wage earners while the opposite is true in comparisons to FSU immigrants. Possibly, due to low likelihood of self-employment in Israel, we find that, when self-employed, Ethiopians work in occupations of equivalent occupational status to native-born Ashkenazi Jews and are in significantly higher status jobs than Palestinians and FSU immigrants. Among wage earners, we find very similar patterns of occupational status across place. The mean difference between Ethiopian men and the native-born majority is approximately − 10 in Israel and the U.S. and the smallest disadvantage observed is relative to the most marginalized group in both countries.

Table 3 Mean difference in occupational status by type of employment among men in the U.S. and Israel, 2012–2016
Full size table

Discussion

Policy initiatives in Europe (Boffey, 2018) and the U.S. (Martin, 2013; Weiner, 2007) emphasize the need to select immigrants by human capital characteristics with the assumption that they will assimilate faster and contribute to the host-country economy. However, without conducting comparisons across place, it is unclear if integration is proceeding more successfully in countries with immigrant selection policies (Alba & Foner, 2015). By comparing Ethiopian immigrants in the U.S. and Israel, we conducted a systematic investigation of the ways in which immigrant selection policy impacts immigrants’ labor market outcomes. Although our identification of immigrant selection is based primarily on reviewing immigration policies and migration history, the findings based on our descriptive statistics show higher educational attainment (and consequently higher average occupational status) among Ethiopians in the U.S. relative to Ethiopians in Israel, thus supporting the notion of differential selection across the two places. Yet selecting based on human capital is not enough to fulfill the goal of immigrant selection policy. These policies are put in place to bring in highly skilled immigrants who will contribute to the host-country economy through their labor market activities. Our findings show that, despite having higher human capital characteristics in the United States, Ethiopian immigrants’ experiences in the labor market are quite similar in the U.S. and Israel suggesting that immigrant selection policy is ineffective.

Using PS matching, we find that—differential selection notwithstanding—Ethiopian immigrants in both Israel and the U.S. are more likely than the native-born and FSU immigrants to be in the labor force with the exception of Ethiopian women in the U.S. There, Ethiopian women are significantly less likely to be in the labor force than native-born black women. These patterns of women’s LFP in the U.S. are expected given black women’s long history in the U.S. labor force and consistently higher LFP rates compared to white women (Goldin, 1977; Toossi, 2013). In Israel, Ethiopian women’s high LFP results reflect both low LFP rates of Israeli women more broadly and Ethiopian women’s LFP improvement over time. Women’s LFP rates were consistently lower in Israel compared to the United States from the 1990s until the mid-2010s (Schein, 2020; Toossi, 2013). Despite increases over time, Arab women have much lower LFP rates than Jewish women in Israel (Kraus & Yonay, 2018). Ethiopians experienced low LFP rates compared to the native-born in the first decade after arrival but gradually, due to state-run employment programs (Dagan-Buzaglo, 2008), as well as increasing familiarity with the labor market, LFP rates increased to levels similar to the native-born (King, Fischman, & Wolde-Tsadick, 2012). Our results show that Ethiopian women’s LFP now surpasses that of the native-born.

At first glance, Ethiopian immigrants’ high LFP rates in both countries indicate positive labor market incorporation. However, because they occupy lower status jobs compared to any other group, our results actually indicate that Ethiopians experience a labor market disadvantage. In fact, we find that the position Ethiopian immigrants occupy in the labor market (indicated by occupational status) is lower than that occupied by even the most historically marginalized group (Palestinians in Israel and native-born blacks in the U.S.) and FSU immigrants. While the non-selective nature of Ethiopian immigration to Israel and their low level of human capital could partially explain their low performance in the labor market relative to other groups (Semyonov et al., 2015), if immigrant selection policy reduced differences between the native- and foreign-born, we would expect smaller differences between Ethiopians and the most marginalized groups in the U.S. and that is not the case. Ethiopian men’s occupational disadvantage is larger in the United States than in Israel and—despite shared racial background—Ethiopian immigrants experience more disadvantages in the labor market than even U.S.-born black men. These results are consistent with previous work showing that even well-educated Africans cannot find jobs commensurate with their skills (Creese & Wiebe, 2009; Habecker, 2012) and have lower labor market outcomes than U.S.-born black individuals (Corra & Borch, 2014; Kim, 2015; Tesfai, 2017). Ethiopian men’s occupational disadvantage can be attributed in part to their difficulty attaining high status occupations even when highly-skilled. There is a great deal of research detailing the disadvantages black jobseekers experience when applying for work (Moss & Tilly, 2003; Stewart & Perlow, 2001) and being considered for promotion (Maume Jr., 1999) in the U.S. Our results for men in both countries are in line with this research.

We may not see the same pattern of larger disadvantages among highly skilled women because Ethiopian women are more likely to be in competition with the low-skilled native-born workforce. In both the U.S. and Israel, Ethiopian women have the lowest educational attainment and—given these average educational characteristics—Ethiopian women with high skill levels may be seen as an anomaly. Arai, Bursell, and Nekby (2011) show that marginalized groups seen to be contradicting norms associated with their backgrounds do not experience the same labor market disadvantages as their peers. As a result, highly-skilled Ethiopian women in both countries would experience lower levels of occupational status disadvantage than their low-skilled counterparts.

The occupational status results indicate that Ethiopians experience discrimination in the U.S. and Israeli labor markets. Among immigrants, high rates of self-employment are another common indicator of discrimination because immigrants turn to self-employment when faced with discrimination in the wage employment sector (van Tubergen, 2005). In the U.S., our results show higher rates of self-employment among Ethiopians compared to the native-born. Despite dispersal across the U.S., Ethiopians are highly concentrated in a few major metropolitan areas (such as Washington D.C., Atlanta, and Seattle) where populations are large enough to sustain businesses even if they only serve co-ethnics. Unlike in the U.S.—and consistent with previous findings (Kayam, 2014; Offer, 2004)—our results show that Ethiopian self-employment in Israel is relatively low, likely due to Ethiopian immigrants’ residential patterns. Immigrant businesses and self-employment opportunities are most prevalent in ethnic enclaves. While Ethiopians in Israel are concentrated in specific neighborhoods (Offer, 2007), they do not form a large enough population in any place to sustain a community of businesses serving other Ethiopians. The social networks that pushed many Ethiopians towards these neighborhoods were not effective in increasing employment or community organization (Offer, 2007). Furthermore—due to their race—Ethiopians are unable to build businesses that trade between the elite and the most marginalized (i.e. they are unable to become middleman minorities) because they themselves are the most marginalized in Jewish society.

Together, our results suggest that rather than immigrant selection policy being the driver of labor market success, a combination of immigrant disadvantages and racial discrimination determines the extent of immigrant labor market incorporation. However, in arguing that race is an important factor determining black African immigrants’ labor market incorporation, one should consider a major difference between the two places—the presence of a same race native-born minority. The presence of a racially stratified society implies that different ethnic groups of immigrants assimilate into particular races (Anderson & Massey, 2001). In the U.S., our results suggest that—likely due to their race—Ethiopians experience the same barriers to labor market integration as the long-standing native-born black population. Yet, due to the disadvantages associated with their nativity status, Ethiopians occupy an even lower position than that of U.S.-born blacks.

The Israeli findings, however, present an interesting case. Similar to Ethiopians in the U.S., Ethiopian immigrants to Israel joined an ethnically stratified society, but one in which a same race (black) community did not exist until their arrival. Racial discrimination practices and ideology were present in the Israeli society since its establishment and were largely practiced against the Palestinians (Kraus & Yonay, 2018). Yet, as Shenhav and Yonah (2008) demonstrate, despite acknowledging the existence of methodical and systematic discrimination against these groups, even critical literature in Israeli social sciences rarely discusses the practices as issues of racism. Most likely the applications of these entrenched practices to the new arrivals of Ethiopians were immediate, but more severe due to their race. In a study of the conversion process required of many immigrants upon arrival to Israel, Goodman (2008), reveals that the racialization of black Ethiopians was different from of that of white FSU immigrants who arrived at the same period and this created hierarchies between the two groups. For example, the racialization of Ethiopians stressed Ethiopians as “lagging” and “holding simple, innocent and patriotic faith” compared to “modern, progressive” Russians (Goodman, 2008). Ethiopians arrived when Israeli society was transitioning from a relatively monolithic society into a more multicultural one—a transition that was expected to acknowledge diversity in Israel. However, Ben-Eliezer (2008) argues that, when young Ethiopians asked for recognition and public affirmation of their differences, they found themselves excluded and segregated. Israel’s transition into a multiculturalism did not diminish but increased cultural racism against Ethiopians, particularly its everyday non-institutional version (Ben-Eliezer, 2008).

While anti-black discrimination against Ethiopians in Israel could explain their lower position relative to Ashkenazi Jews, it is not clear why they occupy a position lower than Palestinians, who suffer from high levels of discrimination and are not members of the Jewish dominant group. In addition to anti-blackness towards Ethiopians, this finding could result from some combination of two factors: the relatively higher human capital among Palestinians and unique consequences of Palestinian spatial segregation. Palestinians comprise approximately 17% of the population of Israel (much larger than the Ethiopian community) and are highly segregated, living mostly in all-Palestinian localities that run separate local municipalities and school systems. The separate school system protects Palestinians from competition with more privileged groups and thus secures relatively high levels of educational achievements (Khattab, 2003; Shavit, 1990). Similarly, the combination of Palestinians’ spatial segregation and discrimination against Palestinian workers in the dominant labor market expedited the growth of an ethnic local economy. Consequently, a significant portion of Palestinian workers mange to avoid competition with more dominant groups and enjoy higher work status—especially those employed in the public sector (Kraus & Yonay, 2000; Lewin-Epstein & Semyonov, 1994). Although Ethiopians are segregated, they do not form “all Ethiopian” localities with independent local municipalities (Offer, 2007; Swirski & Swirski, 2002), therefore, they do not have the same local public sector opportunities as Palestinians.

Overall, despite immigrant selection policies in the U.S., Ethiopians’ labor market outcomes relative to the dominant group are similar to those of Ethiopians in Israel. That is, as black immigrants, Ethiopians occupy the lowest status position in both the U.S. and Israeli social hierarchies limiting their labor market success. As a result, our results indicate that immigrant selection policies do not even serve to mediate the impact of racial discrimination on immigrants’ labor market outcomes. However, there are differences by gender. Ethiopian men’s occupational disadvantages are larger in the United States than Israel when compared to both the U.S.-born and FSU immigrants. The occupational disadvantage relative to FSU immigrants was solely found among wage earners indicating that racial discrimination by employers is the main driver of Ethiopian men’s larger disadvantage in the United States. Ethiopian women’s occupational disadvantage, by contrast, are consistent across place which may reflect larger patterns of occupational segregation. In their study of occupational segregation in the United States Alonso-Villar, del Río, and Gradin (2012) found that racial disparities in occupational segregation are larger among men than among women. Similar patterns have been observed in Israel (Semyonov et al., 2015).

One factor that we could not control for in our analyses is language skills. Immigrant labor market disparities may be due—in part to—poor host-country language skills. Semyonov et al. (2015) find that the labor market disadvantages of Ethiopian and FSU immigrants can be fully attributed to differences in socioeconomic characteristics and low levels of Hebrew language proficiency. At first glance, it may seem that Ethiopians in Israel may have an easier time with the host-country language than FSU immigrants because both Amharic and Hebrew are Semitic languages. Yet despite the similarities between Hebrew and Amharic (and Ethiopians’ high rates of attendance in Hebrew language classes) Ethiopians have lower Hebrew proficiency than any other immigrant group (Raijman et al., 2015). The same study shows that FSU immigrants’ Hebrew proficiency is only slightly higher than that of Ethiopian immigrants: self-reported rates of Hebrew proficiency among FSU and Ethiopian immigrants is 2.6 and 1.9 respectively on a scale of 1 (no Hebrew at all) to 5 (very fluent).

In the absence of language classes provided by the U.S. government (as is the case in Israel), Ethiopian and FSU immigrants’ English ability should be comparable because the linguistic distance between English and Russian, and English and Amharic are quite similar (2.25 and 2.0 respectively) (Chiswick & Miller, 2005). That is indeed the case. Host-country language proficiency in the U.S. is higher for both groups than in Israel. Over half of Ethiopian and FSU immigrants both speak English at least very well with Ethiopian immigrants having higher English proficiency (Ruggles et al., 2019).

The host-country language proficiency of FSU and Ethiopian immigrants are likely similar in both countries, therefore the absence of a language variable in this analysis should have very little effect on our findings. If labor market disparities were attributable to poor language skills, we would see similar gaps for Ethiopian and FSU immigrants in both places, but we do not. These results suggest that—while language may play a role in Ethiopian immigrant labor market disparities—it does not fully account for their poor outcomes.

The results of our analyses have both policy and research implications. Immigration debates in the U.S. (Krogstad & Gonzalez-Barrera, 2018) and Europe (Secretary of State for the Home Department, 2018) currently revolve around selecting immigrants based on skill. However, our results show that for racial minority (black) immigrants, selection policy does not matter. Our results indicate that—rather than focusing on migration policy as the ideal way to improve immigrant outcomes—governments should instead focus on policy that better addresses racial discrimination in the labor market.