Schooling refugee children: Long-run effects of integration versus separation

Should newly arrived refugee children immediately enter mainstream schools or first attend separated schools? Using the setting of the Netherlands, this column shows that the choice involves a major trade-off in the long run: direct integration reduces educational attainment, but it increases cultural assimilation as measured by hormonal-contraceptive use among teenage women. Separated schools’ greater access to teachers specialised in teaching the native language to non-native speakers offers a possible explanation for their educational advantage.

Integrating refugee children remains one of the most pressing challenges of our time (Boustan et al. 2025, Bahar et al. 2026). Large numbers of displaced children, from Ukraine and from conflict zones elsewhere, have recently entered school systems in many Western countries, reviving the debate on how best to organise their schooling. For these children, schools constitute the first sustained exposure to the host country’s language and cultural norms. Should newly arrived refugee children enter mainstream schools immediately, or first attend separated specialist schools?

Conceptually, both schooling arrangements offer potential benefits and downsides. Separated schools can provide targeted instruction and a focus on language acquisition, but the lack of exposure to native students may reinforce social and linguistic separation. Integrated schools offer immediate exposure to native peers, which can support language acquisition and adaptation to the host culture, but following the native curriculum from the outset may overwhelm newly arrived children. Despite the importance of this decision, existing evidence is surprisingly scant and focuses on within-school placements and short-run outcomes (Schilling and Höckel 2026).

In our new study (Pavese et al. 2026), we examine the long-term consequences of the alternative initial schooling arrangements on refugee children’s education and assimilation. Rich administrative data provide information on refugee children who arrived in the Netherlands between 2003 and 2010. 

We exploit two key features of the unique Dutch setting where both educational arrangements coexist. First, refugees are randomly allocated to municipalities, which decide whether refugee children who arrive at primary-school age receive separated or integrated education. Second, in contrast to the primary-school setting, secondary-school-age arrivals are initially placed in separated schools regardless of the municipality they are assigned to. 

Our empirical design compares children arriving at primary- and secondary-school age across municipalities that do or do not place primary-school-age arrivals in separated schools. This approach allows us to isolate the effect of integrated education on refugee children’s outcomes, employing two-way fixed-effects models that include age-at-arrival fixed effects and municipality by immigration-year fixed effects.

Direct integration lowers educational attainment

Our results indicate that refugee children placed directly into mainstream primary schools are less likely to obtain a high-school diploma by age 18 than those initially attending separated specialist schools. The estimates suggest a 7 percentage-point decline relative to a mean of 36% in our sample.

Figure 1 shows the effects by age at arrival. The negative impacts emerge among children arriving at primary-school age, between 6 and 11. By contrast, no significant differences are observed across secondary-school arrival ages, from 12 to 17. This is consistent with all secondary-school-age arrivals initially receiving the same treatment – separated schooling – regardless of their assigned municipality. Additional analysis reveals that the disadvantage in diploma attainment persists at age 21 among the cohorts we can follow to that age.

Figure 1 The effect of integrated schooling on obtaining a high-school diploma

Notes: Estimated effects of integrated schooling on obtaining a high-school diploma (by age 18), by age at arrival grouped in two-year intervals, relative to the 12–13 arrival-age group. Bars indicate 95% confidence intervals. Source: Pavese et al. (2026).

Why does integrated schooling decrease the probability of having a diploma by age 18? We do not find a significant effect on early school dropout, suggesting that the diploma disadvantage reflects differences in educational pathways. In particular, direct integration increases the probability of attending the very lowest secondary-school track by over 8 percentage points. This track is designed to prepare students who have low IQs and severe learning deficits for participation in society as decently as possible, mostly conveying basic practical skills such as filling out forms or cooking. During the period studied, it did not award a nationally recognised secondary-school diploma.

Attending this track has enormous implications for outcomes later in life. The track is associated with substantially lower earnings and a much higher incidence of benefit receipt at age 30 compared to the six other tracks. While less than 3% of native students attend this track, 17% of refugee students do.

The negative impacts of directly integrated schooling on obtaining a diploma and attending the lowest ability track are concentrated among refugee students with lower-educated parents and those from Muslim-majority countries. By contrast, students with tertiary-educated parents and those from non-Muslim-majority countries are more likely to experience grade retention when directly integrated. This pattern reflects a basic trade-off in ability-based tracking systems: when students fall behind academically, they face a choice between progressing on schedule into a lower track or repeating a grade.

Integrated schooling increases cultural assimilation

In addition to educational outcomes, there is extensive interest in cultural assimilation (Boustan and Abramitzky 2016, Mazumder et al. 2020). Intriguingly, we find that the effects on cultural assimilation go in the opposite direction to those on educational attainment. Following Achard (2026), we measure cultural assimilation as hormonal-contraceptive use among teenage girls, recorded in health-insurance data.

Direct integration raises the probability of hormonal-contraceptive use by age 18 by around 5 percentage points, consistent with the interpretation that early exposure to native peers facilitates adaptation to host-country norms. This effect is driven by fully integrated schools, where refugee children attend all classes alongside native peers, and does not emerge in semi-integrated schools, where some or all instruction takes place in separated classrooms. This pattern suggests that sustained interaction with native peers may play an important role in cultural assimilation.

The effects differ strongly by country of origin. Figure 2 shows that the increase is concentrated among girls from non-Muslim-majority countries, with no detectable effect among those from Muslim-majority countries. Relatedly, the literature has found that migrants with Muslim backgrounds tend to have more traditional cultural norms and show slower cultural assimilation (Zenou et al. 2007, Rainer et al. 2020).

Figure 2 Heterogeneous effects of integrated schooling on hormonal-contraceptive usage

Notes: Estimated effects of integrated schooling on females’ probability of using hormonal contraceptives in at least two years by age 18, by parental education and origin country. Bars indicate 95% confidence intervals. Source: Pavese et al. (2026).

We do not find effects on the use of antidepressants or other frequently used medications. This suggests that the contraceptive result reflects a culturally salient behaviour rather than a general increase in healthcare access.

The role of specialised teachers

The employment of specialised teachers offers a possible explanation for the educational advantage of separated schools. Specialisation allows separated schools to employ teachers qualified to teach Dutch as a second language. As shown in Figure 3, municipalities with separated primary schools employ on average about four such teachers, against fewer than one in municipalities with integrated schooling.

Figure 3 Specialised teachers in municipalities with separated and integrated primary schools

Notes: Number of teachers with a Dutch-as-a-second-language qualification in municipalities with separated and integrated primary schooling, by level of education.
Source: Pavese et al. (2026).

Records of school inspections also suggest that separated schools offer more targeted support. They receive higher ratings by the inspectorate than integrated schools for tailoring instruction to students’ needs and supporting those with special learning needs. This advantage does not extend across other inspection measures.

The overall quality of the school environment does not differ between the two school arrangements: refugee children attend highly disadvantaged schools in both arrangements, with similar funding and similar student bodies. These similarities suggest that differences in general school advantage are unlikely to explain the educational results.

Policy implications

Taken together, our findings point to a fundamental trade-off between integration and specialisation in refugee education. Direct integration promotes cultural assimilation, particularly among girls from non-Muslim-majority countries, but reduces educational attainment, especially among children with lower-educated parents and those from Muslim-majority countries. Temporary placement in specialist schools can support children’s longer-term educational prospects, while limiting early interaction with native peers. The policy choice therefore depends on the relative weight given to educational and cultural outcomes. 

However, our results also suggest potential ways to reduce the trade-off under either arrangement. In integrated systems, employing specialist language teachers and providing more targeted instruction could help address the educational difficulties faced by newly arrived children. In systems with temporary separated schooling, encouraging contact with native peers through after-school activities and ensuring subsequent access to mainstream schooling could support cultural assimilation.

Source : VOXeu