How does refugee ethnicity shape locals’ support for integration?
EGAP researchers: Tolga Sinmazdemir
Other researchers: Anna Getmansky, Konstantinos Matakos
Key takeaways: Local populations can hold significant biases against refugees from particular ethnic groups, even when those refugees share a common religion with their host country. While human and social capital can help mitigate these biases, they are unlikely to eliminate them completely for refugees from all ethnic groups and in all domains of life in the host country.
Geographical Region: Middle East
Type of study: Survey experiment
Preparer: Mark Williamson
Executive Summary
This study investigates which factors shape local populations’ support for the integration of refugees into the social, economic and political life of their host countries. The authors use a conjoint survey experiment to assess locals’ willingness to welcome Syrian refugees in Turkey. They show that even though these two countries share the same dominant religion, Turkish respondents remain strongly opposed to having Arab and Kurdish (as opposed to Syrian Turkoman) refugees as neighbours, providing them with work permits, or granting them citizenship. Indicators of human and social capital (e.g. language ability, education) boost support for integration, but not equally for all refugee profiles. Those with an Arab identity benefit less from these factors and generally only in the realm of labor market incorporation. Overall, this study highlights how ethnicity creates a significant barrier to popular support for integration in host societies, a bias which may not be fully overcome by refugees pursuing education and language skills.
Policy Challenge
What factors shape a local population’s willingness to accept refugees into the social, economic and political life of their country? Much of the existing research on this question focuses on attitudes in Western countries, which tend to receive smaller numbers of refugees overall and in which refugees often do not share the same religion or language as the host society. For example, several studies have pointed to anti-Muslim bias in Europe and the United States as a source of opposition to asylum seekers.
By contrast, there is less evidence on how refugees are viewed in societies that are heavily impacted by refugee arrivals and where religious differences pose less of a cultural threat. In these contexts, the authors argue that other immutable and visible cues, like ethnicity, may act as a heuristic, or mental shortcut, for locals to form their opinions about refugees. A refugee’s ethnic identity can simplify complex and uncertain choices for locals about whether to welcome that person into one’s society and lead to a greater preference for those with a shared ethnicity.
But ethnicity may also indirectly affect how other characteristics of refugees, like their social or human capital (e.g. education, language ability, social networks in the host society), affect locals’ support for their integration. The authors argue that these factors are generally second-order concerns, meaning that locals will only consider them after learning about a refugee’s ethnic identity. For refugees who share the same ethnicity as members of the host society, secondary indicators of potential success can further boost support for their integration, but for refugees without a common ethnic identity, those same factors may not be able to surmount locals’ initial ethnic biases. If this theoretical model is true, it has important policy implications about the extent to which refugees themselves are able to overcome exclusionary attitudes through their own efforts to integrate.
Context
Since the beginning of the civil war in 2011, over 5.5 million people have fled Syria. The majority of them (over 3.5 million) settled in Turkey and Syrians now represent the largest foreign-born population in the country. Because of the sheer volume of this population, most Syrian refugees live outside of camps and have frequent contact with the local population. Turkey thus offers a useful context to investigate locals’ attitudes toward refugees, since public discourse centers around the various aspects of integration into the host country, rather than the more basic question of whether they should be hosted in the first place.
This is also an informative setting because, while Turkey and Syria share a common dominant religion, there is significant variation in the ethnic origins of the Syrian refugee population. Around 80 percent are of Sunni Arab origin, with the remainder roughly equally split between Kurds and Turkomans, which are of Turkish ethnic origin. This diversity allows the authors to investigate how shared ethnicity, rather than just co-religiosity, shapes local attitudes toward refugee integration.
While Turkey initially offered an open-door policy to Syrians fleeing the war, that policy has since become more restrictive and locals’ attitudes toward Syrian refugees have become increasingly antagonistic over time. Today, locals express low levels of support for providing refugees with work permits, living alongside them as neighbours and granting them full citizenship.
Research Design
This study is based around an online survey experiment conducted in 2018 with just over 2,300 Turkish respondents. The sample includes regions that make up around 80 percent of the country’s population, with an oversample of provinces that host a larger proportion of refugees.
The authors use a conjoint experiment, which involves randomly varying several factors at once to isolate which are most important in shaping public opinion. In this study, respondents were presented with pairs of hypothetical Syrian refugee profiles in which their gender, age, ethnicity, religion, language skills, education, ties with locals, involvement in the civil war in Syria, and victimization because of the civil war were all randomly assigned. The table below displays an example pair that a respondent might have seen in the survey.
| Attribute | Profile 1 | Profile 2 |
| Gender | Man | Woman |
| Age | 18-30 | 18-30 |
| Ethnicity | Arab | Turkoman |
| Religion | Sunni | Alawite |
| Language skills | Doesn’t speak Turkish | Speaks Turkish |
| Education | High school | Primary school |
| Ties with locals | Has Turkish friends | Doesn’t have Turkish friends |
| Involvement in war | Was not involved in the civil war | Was not involved in the civil war |
| Victimization | Was tortured during the civil war | Was not tortured during the civil war |
After reading each pair of refugee profiles, respondents indicated their willingness to have each profile as their neighbor (social integration), to offer them a work permit (economic integration) and to grant them citizenship (political integration). They provided these responses as both a ranking on a 1 to 7 scale and by making a binary choice between the two profiles in each pair. Overall, each respondent made comparisons between three different pairs of hypothetical refugee profiles. The authors then use regression to estimate the independent effect of each refugee attribute (e.g. ethnicity, education) on respondents’ preferences for each of the profiles.
Results
The authors find that ethnicity has a significant effect on respondents’ willingness to accept a refugee as a neighbor, provide them a work permit, or grant them citizenship. Refugee profiles that indicated a Kurdish or Arab identity saw around 6 to 7 percent lower support for integration compared to Turkoman profiles. This effect appears to be driven by favoritism toward refugees with a shared identity: Kurdish respondents living in Turkey, for example, prefer Kurdish profiles relative to Arab and Turkoman profiles, and a similar pattern exists among Arab respondents. (However, because Kurds and Arabs are both minorities in Turkey, the sample as a whole tends to prefer Turkoman profiles). Interestingly, this ethnic bias is larger in magnitude than biases based on religion, although the authors do find that the integration of Christian refugees is less accepted than it is for Muslims.
On average, refugee profiles indicating higher levels of social and human capital also see more welcoming responses. Those who speak Turkish and that have friends in Turkey are more likely to be accepted as neighbours and granted a work permit and citizenship. These effects vary between 2 and 4 percentage points, with larger effects on respondents’ willingness to accept a given refugee as a neighbor. Having a university education also exerts a strong, positive effect on support for integration, particularly on the question of work permits.
But how do these factors interact? Are better educated and more socially connected Arab and Kurdish refugees able to overcome the ethnic biases against them? Or do only Turkoman profiles benefit from higher social and human capital? The authors find that, at least when it comes to granting a work permit, there is no difference in the effect of education, language skills and local connections across the three ethnic groups. All refugees appear to benefit equally from these factors. That being said, the benefits of education and language ability only boost support for Arab refugees on the issue of work permits; on the question of welcoming them as neighbors and granting citizenship, these factors do not exert the same positive effects. For Turkomans and Kurds, however, these factors boost support for integration across all types of integration. (The authors suggest that the lack of a significant difference in effect sizes between Turkoman and Kurdish profiles could be explained by the relatively larger resident Kurdish population or the history of anti-Arab bias in Turkey). Taken together, these results provide evidence that when it comes to locals’ attitudes, greater human and social capital may not reward refugees from all ethnic groups equally and that the positive effects of these factors on support for integration may be limited to specific domains of life in the host country.
Lessons
This study demonstrates that locals can hold significant ethnic biases against refugees, even when they share the same religion. Having a university education, gaining local language skills, and having friends among the local population can mitigate some of these biases, but may not completely overcome them. Moreover, the benefits of pursuing these strategies may not apply to refugees from all ethnic groups and in all domains of integration. Overall, then, policymakers should recognize that there is unlikely to be a single intervention that can address all types of antirefugee bias that exist in host societies.