Introduction

Over the last 25 years, Spain has established itself as an immigrant-receiving country (CES, 2019: 12), with a significant increase in Latin American migration flows. Since 2005 Latin American migrants have outnumbered nationals from northern African countries and Eastern Europe (Ayuso & Pinyol, 2010; INE, various years). The greater influx of Latin Americans to Spain is explained by the tightening of immigration policy in the USA after 9.11.2001, as well as by the influence of an emerging and poorly regulated labour market, cultural and linguistic affinity and family networks (Bayona-i-Carrasco & Avila-Tàpies, 2020; Zufiaurre, 2006). During this period, the number of children of Latin American immigrants has also increased (INE, various years), posing significant educational challenges to accommodate this new reality (Cebolla-Boado & Garrido Medina, 2011: 607).

The ability of societies to increase social cohesion in the presence of large migration flows depends on their capacity to promote immigrants’ long-term integration through education. Education can help immigrants acquire the skills to participate in the labour market, contributing to the host country economy and the welfare arrangements. Education can also foster immigrants’ social and emotional well-being, motivate them to participate in social and civic life and help them to feel part of society (OECD, 2018, 3, 13, 18). Ultimately, the structural integration of second-generation immigrants is critically dependent on their educational attainment (Jonsson and Rudolphi, 2010). Whether or not they can integrate with conditions of greater equality with their native-born Spanish counterparts—especially in the labour market—will depend on how successful they are in school (Cachón, 2005; Parella, 2009; Portes et al., 2018: 150).

While previous research has identified several factors associated with the process of integration into the educational system1, international and Spanish studies tend to agree that family socioeconomic background is the factor with the greatest influence on the educational achievement of immigrant children—often referred to as the second generation (Álvarez-Sotomayor & Martínez-Cousinou, 2016; Anghel & Cabrales, 2010; Fernández et al., 2019; OECD, 2017, 2019; Praag et al., 2018; Zinovyeva et al., 2014). Indeed, in the case of Spain, it has recently been argued that immigrant children’s educational achievement is not associated with national origin but rather with family socioeconomic status (Aparicio & Portes, 2014: 199–202; Portes et al., 2018: 165; Haller & Portes, 2019; Fernández et al., 2019: 48). However, the existing evidence is still insufficient to rule out the impact of national origin (Álvarez-Sotomayor & Martínez-Cousinou, 2016: 529–530; CES, 2019: 172), as indicated by the experience of several European countries (Crul & Vermeulen, 2003; Ismail, 2019; Thomson & Crul, 2007; van Niekerk, 2007). As some studies have shown, the socioeconomic background cannot entirely explain disparities in educational achievement between groups (Kristen & Granato, 2007: 344; Khattab, 2018: 459; OECD, 2019: Chapter 7). In turn, there is a significant gap in longitudinal research on second generations in Spain (Medvedeva & Portes, 2018: 23), and very few studies have approached the question of how the children of Latin American immigrants adapt to their new lives (Aparicio et al., 2014).

This article aims to fill part of this gap by identifying the effect of national origin on the educational pathways of children of Latin American immigrants in Spain, based on data from the Longitudinal Study of the Second Generation in Spain (ILSEG in its Spanish acronym). To enable a comparison with the children of native-born Spaniards, this article examines for the very first time the results of the second and third wave of the ILSEG, selecting a sub-sample consisting only of Latin American immigrant children in a context of economic crisis and recovery. The fact that the economic downturn that started in 2008—and affected Spain more than any other European country in terms of job losses—also meant a clear deepening of the existing social and economic gulf separating native Spaniards and immigrants (Pajares, 2010; Colectivo Ioé, 2012; Mahía & de Arce, 2014; Mooi-Reci & Muñoz-Comet, 2016; Moreno et al., 2017).

The study contributes to the existing literature about the second generation of Latin American immigrants in Spain and shows how educational achievement is associated not only with the family socioeconomic background (as previous works point out), but also with national origin. Latin American immigrants’ offspring are more likely to attain lower educational levels than the children of Spanish natives, exhibiting—at the same time—a different pattern of academic success, depending on their national origin. To some extent, these outcomes can be explained by the 2008–2013 economic crisis, which, for certain national origins, fostered more challenges in terms of economic integration, underscoring the influence of some characteristics of their co-ethnic communities (Portes & Rumbaut, 2001, 2005; Portes et al., 2018). Finally, we discuss the influence of national origin on the educational achievement of children of immigrants by pointing out some of its implications for public policy.

Theoretical Background and Hypotheses

The Theory of Segmented Assimilation

The theory of segmented assimilation offers a comprehensive explanatory framework to approach the integration processes of second generations into transnational contexts (Portes & Rumbaut, 2001). Especially relevant is its recognition of the role played by the social and economic receiving context and how the distinctive characteristics of immigrant groups interact with structural features of the destination society (Alarcón et al., 2014: 1617; Portes et al., 2018). The theory of segmented assimilation shows how the integration of immigrants’ descendants does not occur homogeneously in all segments of society (Aparicio & Portes, 2014), arguing that the educational achievement of immigrants’ offspring is a central aspect to consider (Portes et al., 2018).

Educational Achievements According to National Origin

Success in education is probably the most relevant criterion for evaluating the process by which children of immigrants adapt to Spanish society (Aparicio & Portes, 2014: 163). Achieving higher educational levels largely determines the future possibility of accessing a desirable and well-paid job (Ibid. 34, 124). The educational level attained by immigrant children is crucial in the future course of their incorporation into the Spanish labour market. Early dropout or failure to get past initial training courses seriously harms their chances of success (Ibid. 53, 124).

According to the theory of segmented assimilation, the disparity in the educational achievement of different nationalities responds to a process of adaptation that differs depending on the characteristics of immigrant families (Aparicio & Portes, 2014: 128). Nevertheless, if the children of immigrants have higher educational aspirations (ideally) or expectations (realistically) than their native Spanish peers (attributable to immigrants’ optimism or drive), they can achieve similar or even better academic results (Aparicio & Portes, 2014: 38; Haller & Portes, 2019). Such aspirations and expectations are strongly conditioned by their parents’ human capital, although differences due to national origin do not completely disappear, when the effect of their parents’ educational level is controlled (Aparicio & Portes, 2014: 40; Alarcón et al., 2014: 1632).

Expectations are explained—among others—by different ethnic and cultural attitudes regarding the propensity for effort or the value given to education (Alarcón et al., 2014: 1619; CES, 2019: 172) and by a selective acculturation process (Aparicio & Portes, 2014: 45). This selective acculturation process does not involve abandoning the culture and language of the home country but preserves and integrates these resources with those of the receiving society (Portes et al., 2018: 155). In the case of those with Latin American backgrounds, having Spanish as their mother tongue facilitates selective acculturation (Ibid. 155) and, to some extent, academic performance (Alarcón et al., 2014: 1620).

The reception context associated with immigrants’ home country also plays a significant role in explaining the adaptation process of their children (Aparicio & Portes, 2014). The theory of segmented assimilation suggests that immigrant groups’ incorporation varies in accordance with the characteristics of the group and the contexts of reception (Alarcón et al., 2014: 1631). These features may include stereotypes about race or ethnicity, the size and characteristics of pre-existing co-ethnic communities and the economic context of the receiving society (Haller & Portes, 2019; Ogbu, 1991; Portes & Rumbaut, 2001, 2005). In particular, racial or ethnic stereotypes can generate a perception of being discriminated against among certain types of young people (Telles & Ortiz, 2008; Yiu, 2013; Wihtol de Wenden, 2014; Alarcón et al., 2014: 1619). The more intense this perception of discrimination, the poorer the prognosis for successful integration into Spanish society (Aparicio & Portes, 2014: 45), which in turn, influences educational achievement. Conversely, low levels of perception of discrimination can be considered a predictor of successful integration (Ibid. 111).

Nationality differences in educational attainment—from the segmented assimilation theory viewpoint—reflect the combined weight of the abovementioned elements, including the experiences that the child and the family bring from their country of origin. Nonetheless, such factors alone cannot fully explain all differences among national groups, and thus, many significant nationality effects remain, “with this influence being much stronger among some groups than others” (Portes & Rumbaut, 2001: 251). Therefore, segmented assimilation theory predicts significant differences in educational attainment among the different nationalities comprising Spain’s immigrant population (Haller & Portes, 2019: 1845).

Some studies have provided empirical evidence addressing differences in educational achievement associated with national origin. The diagnostic tests carried out in primary schools by the National Institute for Evaluation and Quality of the Educational System, show that—compared with Spanish natives—Latin American immigrants, along with Africans, have the poorest performance rates (Álvarez-Sotomayor & Martínez-Cousinou, 2016: 530). The PISA international tests have also shown that the second generation of Latin American immigrants performs worse than Spanish natives, despite having Spanish as their mother tongue (Zinovyeva et al., 2014: 25–26, 29, 50). Moreover, the results of the Longitudinal Study of the Second Generation in Spain (ILSEG) study point out that, especially in their access to university, the children of native-born Spaniards display a clear advantage over the children of immigrants from most Latin American countries (Haller & Portes, 2019: 1833–1834, 1838, 1842). Additionally, the early dropout rate among the children of Latin American immigrants (between 16 and 19 years old) is almost double the rate for native Spanish children at that age (IOE, 2011), indicating a greater risk of social exclusion (Carrasco et al., 2018). In comparison to natives, several studies have pointed to the importance of national origin in identifying disparities in access to job opportunities and in the type of employment (Dustmann et al., 2013; Carmichael & Woods, 2000; Silberman & Fournier, 2006; Arcarons & Muñoz, 2018). Finally, a recent analysis that revisits the context of reception shows that among the classical indicators considered by the segmented assimilation theory, societal reception and group-level resources of co-ethnic communities predict educational attainment (Luthra et al., 2019).

Summing up, differences in educational attainment concerning national origin reflect various aspects related to educational aspirations and expectations (such as optimism and propensity for effort), the process of selective acculturation (in which language is the key element), the experiences that the child and the family bring from their country and the context of reception (including stereotypes of race and ethnicity). Thus, the following hypothesis is proposed:

  1. H1

    Educational achievement is associated with national origin. The children of Latin American immigrants are more likely to achieve lower educational levels than the children of Spanish natives.

Adaptation Patterns of Latin American Immigrants in Spain

One of the limitations of much research on Latin American immigration in Spain is its global approach. Little attention is paid to the special features of each country (the national, as opposed to continental origins of migration), even though the diversity of origins is precisely what most defines Latin American immigration in Spain (Ayuso and Piñol, 2010). The Latin American case not only comprises a group of immigrants of widely diverse nationalities but, in addition, its composition has changed throughout the different stages of the migratory cycle in Spain. Immigrants from the Dominican Republic and Cuba were prominent numerically between the 1980s and the early 1990s, along with immigrants from Peru. Since the end of the 1990s and during the early years of the current century (coinciding with the start of the economic crisis between 2008 and 2009), groups of immigrants from Andean countries such as Ecuador, Colombia, Bolivia and Argentina took centre stage (Ibid.). It was not until the post-crisis context (2014–2017) that other Latin American nationalities increased notably in migratory flows, especially those from Venezuela, Honduras and El Salvador (Bayona-i-Carrasco & Avila-Tàpies, 2020).

The specific times of arrival among Latin American immigrants allows the groups to be differentiated in terms of what reception contexts they faced once in Spain. Regarding the irregularity rate, nationals from Ecuador, Colombia and Peru were among the top five groups in the latest regularisation process of foreign workers in Spain (February–May 2005). This fact alone shows the high prevalence of irregularity affecting these groups on their arrival in the country. Hence, it also explains the exceptional administrative and legal stability they enjoyed just before the 2008 crisis, when a considerable percentage of them had obtained dual citizenship or possessed a permanent residence permit. In fact, this greater legal stability mitigated to some extent the social and economic impact of the crisis on this group (Martínez de Lizarrondo, 2016). A case in point is the Bolivian community, which experienced strong growth between 2004 and 2007, passing from 7000 to almost 70,000 people in just 3 years. This placed them at the start of the crisis with a high incidence of irregularity (Tapia, 2014). In contrast, the Argentinian community has a much lower irregularity rate, as their European roots (mainly Italian and Spanish) have facilitated the naturalisation of many Argentinian immigrants (Gil Araujo, 2010).

Among other negative effects of the Spanish economic crisis (2008–2013), the deepening in the social and economic gulf between the foreign-born and native-born populations must be highlighted (Pajares, 2010; Colectivo Ioé, 2012; Mahía & de Arce, 2014). The concentration of immigrants with Latin American backgrounds in specific economic sectors (low skilled and low-paid jobs), jointly with the economic crisis, blocked their incorporation processes mainly due to loss of employment. In turn, this also encouraged them to return to their countries of origin, especially in the case of those from the Andean communities (Tedesco, 2010). This negative impact significantly affected the groups with the longest residence in Spain, as in the case of immigrants from Ecuador, Colombia and Peru. Nevertheless, the reverberations of the crisis impacted the Ecuadorian community more sharply. Its effects were noticed less in the number of returnees (Boccagni & Lagomarsino, 2011; Parella & Petroff, 2014), than in a major loss of employment, savings and housing of many of those who had remained in Spain (Iglesias et al., 2015). In terms of unemployment, for example at the height of the crisis in 2009, Ecuadorian immigrants lost a quarter of their employed population (Pajares, 2010: 63).

Differences can also be identified in the socioeconomic composition of migratory flows from Latin American contexts. The Dominican Republic and Argentina are the two with the most singularities in terms of sociodemographic features. In the first case, the Dominican community in Spain stands out for a clear predominance of women with low levels of education, their belonging to certain municipalities in the Southwest region with high levels of poverty and a pattern of labour integration mainly linked to domestic service. At the other extreme, the Argentinian community counts with a higher proportion of educated people, especially among women (Actis & Esteban, 2018; ENI, 2007). The decision to emigrate to Spain is addressed as a strategy to evade poverty or to avoid loss of status or downward social mobility as a consequence of the economic and social collapse that the country experienced in 2001 (Actis & Esteban, 2018).

These empirical findings highlight the extent to which immigrants’ economic attainment does not entirely depend on human capital because it is contingent on the context in which they are incorporated, as suggested by the theory of segmented assimilation (Portes & Rumbaut, 2001: 46). On the other hand, the fact that the economic context of reception affected immigrants’ adaptation patterns differentially according to their national origin may be explained—in part— by the role of co-ethnic communities as they represent the most immediate context of reception. Such communities can help in finding jobs, housing and schools for the children and, at the same time, they can open a whole range of possibilities for immigrants’ potential to translate their education and occupational skills into economic returns. In a similar vein, “communities can be a valuable resource, as their ties support parental control and parents’ aspirations for their young. Among immigrants of limited means, this function of social capital is vital” (Portes & Rumbaut, 2001: 65). Co-ethnic communities can also provide some necessary resources for upward mobility and prevent downward assimilation (Haller & Portes, 2019: 1825). In other words, children of working-class parents can benefit from the solidarity of their co-ethnic communities to access otherwise unavailable resources. Strong co-ethnic communities prevent downward assimilation among second-generation youths, providing ways to avoid, for example dropping out of school, having children prematurely or encounters with police (Ibid. 1828). However, there are significant differences among the ethnic communities due to their limits of information and resources. When no such community exists, newcomers must confront adaptation challenges by themselves (Portes & Rumbaut, 2001: 48). The differences in the available resources of the national groups have a strong intergenerational impact on how educational opportunities are framed (Ismail, 2019; Luthra et al., 2019). In turn, depending on the time of arrival and the socioeconomic characteristics of the groups, the national origin has conditioned the educational pathways of their children and influenced the levels of early school dropout (Fernández et al., 2019: 36; CES, 2019: 170). In this way, the following hypothesis is proposed:

  1. H2

    The pattern of educational pathways can be differentiated according to national origin. The children of Bolivian and Dominican immigrants residing in Spain are likely to attain a lower educational level than those of Colombian, Ecuadorian, Peruvian and Argentinian national origin.

Research Design and Data

The research design is based on the Longitudinal Study of the Second Generation in Spain (ILSEG in its Spanish acronym). The ILSEG’s longitudinal design allows us to trace individual and collective changes over time and to establish causal relationships between variables (Medvedeva & Portes, 2018: 24). The first survey was conducted in Spain in 2007–2008 and collected a statistically representative sample of children of immigrants attending schools in the cities of Madrid and Barcelona (where the largest number of immigrants are concentrated), with an average age of 14 years. Nonetheless, a sample of children of Spanish natives, as a control group, was not considered on that occasion. In the second phase of the study carried out in 2011/2012, a sample of 1965 children of native Spaniards was included as a base of reference. In addition, 3810 young people were re-interviewed and a replacement sample of 1534 children of immigrants was selected to compensate for the loss of observations from the first wave. The total sample included 7300 young people with an average age of 18 years. The third and final wave was conducted in 2016/2017 and was able to recover 2922 cases from the previous measurement (2272 children of immigrants and 700 children of native-born Spaniards), with an average age of 22 years (Aparicio & Portes, 2014; Haller & Portes, 2019). The design of the ILSEG sample allowed us to consider the complex linguistic profile of Spain, which combines the official, co-official and heritage languages (Medvedeva & Portes, 2018: 24).

To enable a comparison with the children of Spanish natives as a control group, this paper analyses the last two waves of ILSEG. For the purposes of this study, we selected a sub-sample consisting only of Latin American immigrant children2. The sample was composed as follows: in the second round, the N = 5216 cases (37.7% children of Spanish parents and 62.3% children of Latin American parents), while in the third round the N = 2079 cases (33.4% children of Spanish parents and 66.6% children of Latin American parents). However, since attrition bias in the last survey was significant, statistical adjustments had to be introduced to make results representative. The multivariate statistical analysis includes a correction for sampling bias using the Heckman method, comparable to the procedure implemented by Portes et al. (2018) and Haller and Portes (2019). In correcting the sampling bias, the variables used were sex, city of residence, education of the parents/guardians3, educational aspiration and the category of national origin “Spain”. Additionally, to minimise potential biases due to missing data, in terms of parameter estimates and the precision of confidence bands, imputations were conducted with SPSS multiple imputation data module using the expectation maximisation algorithm. Values were imputed utilising the information provided by the rest of the variables and then averaged over the imputed dataset before the statistical analyses.

Measurement

Dependent Variable

Educational attainment is measured by the level of schooling attained or current since most of the young people were still part of the school system. For this measurement, the questions and their respective response categories were matched between both applications of the ILSEG survey as follows: 1, obligatory secondary education (ESO)/basic or initial vocational training programme (PCPI); 2, medium-level vocational training (FPGM); 3, baccalaureate or higher level vocational training (FPGS); 4, university (university degree, masters or doctorate).

Independent Variables

In allocating national origin to the children of immigrants, the criterion adopted was their country of birth and, in particular for those who were born in Spain (with at least one foreign-born parent), their parents’ country of birth (Aparicio & Portes, 2014, 143). Individuals from the largest nationalities—over 153 cases belonging to a specific national origin group (ILSEG II)—were classified individually by nationality, whereas those groups with smaller representations in the sample were aggregated into an “other” category.

Control Variables

Although our focus is on the socioeconomic background and national origin, we also consider several other factors that differentiate the outcomes concerning educational achievement. The theory of segmented assimilation predicts that the second generation’s degree of integration into the receiving society, in terms of their educational achievements, depends on factors such as parental resources, ambition, the perception of discrimination and migration status. Other variables mentioned are sex, age (in years), early motherhood/parenthood and marital status (Aparicio & Portes, 2014; Haller & Portes, 2019; Portes et al., 2018).

Parental Resources

Family structure has been operationalised using the indicator “Lives with mother and/or father”: living with the father or mother = 0 and living with the father and mother = 1. For the family’s socioeconomic status, the index prepared by Portes and Haller (2019) was used, based on the standardised sum of the parents’ education, their occupation and whether they own their home, divided by 5. The mean of the index is 0 and the standard deviation is 1.

Educational ambition

The “educational aspiration” is understood as the highest level of education that the respondent would like to achieve (as an ideal), while the “educational expectation” has been conceptualised as the highest level of education that the respondent realistically expects to achieve. In both cases, the original values have been recorded as follows: 1 = university or postgraduate degree; 0 = other.

Migration Status and Perception of Discrimination

For migration status, the values are 0 = without Spanish citizenship, 1 = with Spanish citizenship. Regarding the perception of discrimination, in the second wave the question was: “The following is a list associated with feelings among young people. Please indicate the degree in which they apply to you…: Not how much education I achieve, I will always be discriminated against”. The original values were recoded as 0 = not very true + not true and 1 = somewhat true + totally true. On the other hand, in the third wave, the question included was “Do you think that in general, Spaniards discriminate against foreigners?” 0 = no, 1 = yes.

Data analysis and Findings

Table 1 shows the educational achievement levels of the children of Latin American immigrants in the second ILSEG survey. When compared to those of Spanish natives, the data displays the gaps between both groups. Among these gaps, the most pronounced is between the children of Spanish natives and the children of parents of Bolivian and Dominican national origin. While 58% of the children of Spanish natives had reached or were studying “Baccalaureate/FPGS”, in the case of the children of Bolivians and Dominicans the proportion falls to 40% and 37%, respectively. Although slight differences are observed between the children of parents of Argentinian, Colombian and Peruvian national origin and the children of Spanish natives in the “Baccalaureate/FPGS” category, this situation changes significantly when we note their greater presence in the “ESO/PCPI” category. These results are in line with other studies that reveal significant gaps among Latin American children of immigrants with different national origins. Zufiaurre (2006) points out how, compared to children of Argentinian immigrants, the children of Ecuadorians and Bolivians exhibit lower levels of school engagement and less interest in the social value of education. The disparities with Argentinian immigrants’ children are even higher when access to university is considered (Haller & Portes, 2019: 1833–1834), as we will also see in Tables 2 and 3.

Table 1 Educational achievement by national origin, ILSEG II
Full size table
Table 2 Educational achievement by national origin, ILSEG III
Full size table
Table 3 Determinants of educational achievement, ILSEG II & ILSEG III
Full size table

The gaps in educational achievement separating the children of Latin American immigrants from those of Spanish natives have tended to consolidate over time, as shown by the results of the third wave of the ILSEG survey. With an average age of 22 years in the overall Latin American sample, 51% of the children of Spanish natives had completed or were undertaking university studies. A figure that contrasts sharply with children of Bolivian, Ecuadorian and Dominican immigrants who were in the same situation: 29%, 28% and 25%, respectively. The exception seems to be, again, with children of Argentinian immigrants. The lower educational achievement of the children of immigrants affects their possibilities of integration into Spanish society, including their prospects for incorporation into the labour market, as also pointed out by Carrasco et al. (2018). However, these results should be viewed with caution since the database of the third ILSEG survey is not weighted—the bias introduced due to cases lost between the second and third waves has not been corrected—so they need to be retested in a multivariate analysis that includes a sampling bias correction factor. It is also important to determine whether the gaps observed in educational achievement are satisfactorily explained by national origin rather than by the effect of other relevant variables, such as the family socioeconomic status.

Consistent with the previous findings, from the multivariate analysis presented in Table 3, we can conclude that, ceteris paribus, educational achievement is associated with national origin, in that children of Latin American immigrants are more likely to attain lower levels of education than the children of Spanish natives3. Apart from the Argentinian group, the data in both the ILSEG II and ILSEG III waves confirm hypothesis H1. On the other hand, the fact that the children of Argentinian immigrants are likely to achieve educational goals comparable to their Spanish native counterparts clearly indicates the existence of a different pattern in educational performance during the period under study that depends on national origin. The same finding is supported by the fact that the children of Bolivian and Dominican immigrants are likely to have a poorer educational performance than those of immigrants of Colombian, Ecuadorian and Peruvian national origin. Therefore, the data allows us to also confirm hypothesis H2, in both ILSEG waves. These results are consistent, in turn, with a recent OECD study (2019, Chapter 7), which shows that in most countries the difference between natives and immigrants (in the percentage of students who reach only a basic academic level) is still statistically significant even after controlling for the effect of the family’s socioeconomic status.

Regarding the other (control) variables included in the models, educational ambition—measured in terms of aspirations (ideal desires) and expectations (realistic expectations)5—early parenthood and family socioeconomic status are all factors that significantly influence educational achievement in both waves of the survey. In fact, these findings are consistent with the theory of segmented assimilation (Aparicio & Portes, 2014; Haller & Portes, 2019). Also associated with higher educational achievements, although with much less impact, is age, sex and living with both parents. While marital and immigration status are statistically significant in the second wave, they are not in the third.

Finally, as for the context of reception, perceived discrimination does negatively influence the educational achievements of young people (as expected by the theory of segmented assimilation), yet its impact is moderate in both waves. The reader should bear in mind that the effect of perceived discrimination on educational attainment is measured by controlling for several other factors, including the family’s socioeconomic status and national origin. These, in turn, are closely related to immigrants’ experiences of discrimination. Societies are highly stratified by class and race, which determines the extent to which minorities are accepted, as most immigration studies have reported (see, for example, Telles & Ortiz, 2008; Yiu, 2013; Wihtol de Wenden, 2014). Nevertheless, despite being highly correlated, we have included these factors separately in the statistical models since they are not the same. Our finding is consistent with similar studies working with the ILSEG, which discovered that most children of immigrants in Spain (except for Muslims) report few experiences of discrimination and high levels of self-identification with the country not substantially different from those of children of Spaniards (Aparicio & Portes, 2014: 111–112; 115–117; Portes et al., 2016: 114–116; Portes, Aparicio and Haller 2018: 162–163, 165–166, 178). More importantly, according to Portes et al., (2016: 114): “[t]hese low perceptions of discrimination can also be taken as a prima facie indicator of a successful integration process”. In other words, children of immigrants in Spain face a relatively benign context of reception, “where the absence of entrenched ethno-racial hierarchies among the native population translates into a relative scarcity of racist slurs or other hostile actions addressed to them” (Ibid. 115).

This result allows us to infer that the macroeconomic context of the receiving society may have influenced the educational pathways of the children of Latin American immigrants more than stereotypes about race or ethnicity or even the degree to which the different co-ethnic communities have been accepted. The economic crisis of 2008–2013 had a major impact on the adaptation patterns of immigrants, depending on their national origin, by increasing unemployment, with the consequent loss of savings and housing. Therefore, its impact conditioned the educational pathways of their children, depending on the moment of their arrival and the socioeconomic characteristics of the national group in question. Some were even forced to contribute to family support, so rather than continuing with non-compulsory education cycles, they tried to access the labour market early (Fernández et al., 2019; CES, 2019). Recent figures indicate that while 27.9% of immigrant children leave the school system without attaining the grade of compulsory secondary education, only 6.3% of native-born children are similarly affected (CES, 2019: 169). This issue helps perpetuate a vicious circle of poverty and exclusion, given that a significant percentage of the immigrant population in Spain is located in the lower-middle classes or is in a situation of vulnerability due to family circumstances (Fernández et al., 2019: 37). On the other hand, the economic context of reception affected immigrants differentially according to their national origin, as we saw in the “Adaptation Patterns of Latin American Immigrants in Spain” section. The role of co-ethnic communities may explain this since they represent the most immediate context of reception (Portes & Rumbaut, 2001: 65). When no such community exists, newcomers must confront adaptation challenges by themselves (Ibid. 48). The resources of co-ethnic communities can predict educational attainment among children of immigrants (Luthra et al., 2019). This implies that differences in the available resources of national groups have a strong intergenerational impact on their educational opportunities (Ismail, 2019; Luthra et al., 2019).

In sum, in line with the theory of segmented assimilation, the way in which immigrant groups are incorporated into the host society varies depending on some characteristics of the national groups and the contexts of reception they encounter, an issue that is particularly relevant in the Spanish case. The theory of segmented assimilation allows us to show how the integration of Latin American immigrants’ offspring does not occur homogeneously in the Spanish educational system. Significant disparities in academic attainment exist among different nationalities, which is explained in part by the economic crisis and the role of co-ethnic communities.

Conclusions

Studies tend to agree that family socioeconomic background is the major factor that explains the educational achievement of the children of immigrants. In the case of Spain, some have argued that the main issue is not linked to the ethnic-cultural profile of the students, but to their family’s socioeconomic status. Nevertheless, national origin also plays an important role. We tested this argument based on a statistical analysis of the Longitudinal Study of the Second Generation (ILSEG) databases, in the context of economic crisis and post-crisis. We conclude that—in agreement with the theory of segmented assimilation—educational achievement is significantly associated with national origin. The children of Latin American immigrants are more likely to enjoy lower levels of educational success than their Spanish peers. Moreover, educational pathways differ depending on immigrants’ national origin. Such results are based on a particular host country and specific national groups, but we believe that they may apply in other countries with similar experiences of migratory flows. Further research emphasising the interaction between structural-level and individual-level variables is also needed (Vermeulen, 2010: 1226; Luthra et al., 2019).

In the interpretation of these findings, the context of reception is a crucial element to be considered. The gap in terms of educational achievement does not appear to be driven by cultural or ethnic factors. Perceived discrimination seems to be a less important determinant than other variables included in the statistical models. Chief among these are the impact of certain groups’ specificities when shared structural elements and their context of arrival are considered, such as how the economic crisis impacted their parents’ generation—in terms of unemployment, loss of savings and housing—and the role that co-ethnic communities played in mitigating such adverse effects. These adversities have affected the type of educational strategies adopted for their children. In view of these findings, future research should examine in greater detail: the impact of economic crises on the educational achievement of immigrants’ children, paying attention to (low) academic performance and (high) rates of early school leaving; how co-ethnic communities contribute to mitigating such effects; how reliable is self-perception of discrimination as an indicator of a successful integration process that facilitates educational achievement; and to what extent the effect of national origin on educational achievement can be moderated by generational cohorts of immigrants’ children (distinguishing between second and 1.5 generations).

The main research findings enable us to identify some challenges for public policy, especially among those young people who have a lower probability of success at school. To avoid larger social inequality gaps, the design of public policies must pay attention to the root causes by addressing the specific characteristics of these groups. National origin and family socioeconomic status are relevant variables that explain the differences in educational achievement among children of Latin American immigrants in Spain. Public policies must account for the intersection between social class and the characteristics of ethnic communities, especially in the educational field. The type of socioeconomic disadvantage that immigrant students experience is not the same in all groups and countries. Policies aimed at reducing the disadvantages experienced by immigrant students should address the specific needs shaped not only by the heterogeneity of the students and their families, but also by the characteristics of their ethnic communities at the group-level (OECD, 2019, Chapter 7). Since group-level education strongly predicts second-generation attainment (Luthra et al., 2019), public policy must improve its strategy towards first-generation immigrants by considering the intergenerational influences on the second generation’s well-being (Ismail, 2019: 731). Finally, equity of educational opportunities for the children of immigrants means ensuring that educational achievement does not depend on the social or ethnic/racial group to which an individual belongs. At the same time, it means adopting inclusive educational policies based on a recognition of differences. Unlike socioeconomic inequalities, cultural diversity is not something to be eliminated since doing so negates individuals’ identity and sense of self-worth (OECD, 2019: 137). Developing efficient policy tools for these purposes is an urgent task. The post-pandemic scenario, which has further exacerbated inequity and social injustice for vulnerable ethnic-social groups, makes the task more urgent still.