I am a Real Korean: The Effect of Government Support on Multicultural Youth’s National Identity

ABSTRACT

For many governments, forming a shared national identity and uniting the public are among the major tasks for building a resilient nation. The population of multicultural children is increasing in South Korea due to globalisation. In the face of such social change, the Korean government has implemented policies and programmes to support these children. The purpose of these policies and programmes is to help multicultural children develop a national identity as Koreans and thereby assimilate them into the Korean culture and ensure social integration. This study examined how government policies affect the national identity formation of multicultural youth. It found that government programmes are effective in developing these young people’s national identity as Koreans, and the result is supported by robustness checks. Furthermore, the effects were dependent on the nationalities of the multicultural youth’s mothers, which suggests that the government needs to consider the heterogeneous nature of multicultural families when constructing policies.

KEYWORDS:

• National identity
• multicultural youth
• policy evaluation
• Korea

Acknowledgements

This study uses data from the Multicultural Adolescents Panel Study conducted by the National Youth Policy Institute. The article is a substantial revision of a paper that won first prize at the 2017 Multicultural Adolescents Panel Study academic conference in South Korea.

Notes

1. In this article, multicultural youth refers to adolescents from international families, who are attending elementary, middle or high school. The criteria for inclusion are that at least one parent is a non-Korean; the nationality of the adolescent is South Korean; and they attend school in South Korea. Multicultural youth are also known as “children with an immigration background” or “children of immigrant parents”.

2. A general economic model explains this mechanism using a utility function. To briefly illustrate this, take the utility function $\mathrm{U}\left(\mathrm{x},\mathrm{y}\right)=-{\left|\mathrm{x}-{\mathrm{x}}^{\ast }\right|}^{\alpha }-{\left|\mathrm{y}-{\mathrm{y}}^{\ast }\right|}^{\beta }$, where x and y are components of people’s identity, ${\mathrm{x}}^{\ast }$ and ${\mathrm{y}}^{\ast }$ are the optimum points of identity that they think are the best for them, and α and β are the weight to the identity x and y, respectively. If the choice of x and y is far from ${\mathrm{x}}^{\ast }$ and ${\mathrm{y}}^{\ast }$, then the utility level decreases.

3. In general, they examined how identity influences quantitative recursive outcomes such as educational achievements and labour-market performance (see Akerlof & Kranton, 2002; Battu et al., 2007; Battu & Zenou, 2010; Bisin et al., 2011; Casey & Dustmann, 2010).

4. Several studies have investigated how the labour-market outcomes for multicultural youth are affected by language differences, using examples from the US or Europe (Bleakley & Chin, 2008; Casey & Dustmann, 2008; Clots-Figueras & Masella, 2013; Schüller, 2015). These studies provide a chance to indirectly understand the role of national awareness among young people. However, little consideration has been given directly to national identity. This is due to the different cultural backgrounds in society. The Korean Peninsula had remained as one nation for a long time and rarely interacted with foreign cultures outside East Asia (Shin, 2006; Shin et al., 1999). Hence, the idea of national identity is rather distinctive in South Korea, which contrasts with previous studies using European or American examples.

5. Eriksen (1993) claims that the formation of national identity paves the way to integrating a society and improving the effectiveness of government policy, which is what the Korean government pursues. Many studies assert that social and economic benefits arise when governments succeed in integrating multicultural children into society and when ethnic diversity takes root in a harmonious way (see Ahlerup & Hansson, 2011; Akay et al., 2014; 2017; Dimitrova-Grajzl et al., 2016; Georgiadis & Manning, 2013), while a lack of inclusive and integrated identity among the public can cause social friction and conflict (Desmet et al., 2017).

6. Recently, government support programmes have been changing toward the selective, such as based on income level. Kim et al. (2013) provide detailed information about the means-test principle.

7. The budget for the Multicultural Youth Educational Programme was 22.5 billion KRW in 2014, 14.2 billion KRW in 2015, 17.308 billion KRW in 2016, and 19.12 billion KRW in 2017 (Ministry of Education, 2017). KRW 1,000 is approximately equal to US$1.

8. Korean education is based on a “6-3-3” system (six years of elementary school, three years of middle school, and three years of high school).

9. I aggregated the programmes into one policy variable and identified ${\mathrm{T}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$ as a dummy variable for whether students received at least one of these programmes. A limitation of the analysis was that I could not determine which was the most effective programme for policymakers. However, my method still has several advantages. By including a list of covariates in ${\mathrm{X}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$, I could control for the direct effects of these observable variables on national identity formation. In addition, individual fixed-effect ${\mu }_{\mathrm{i}}$ could capture unobserved confounding characteristics of student $\mathrm{i}$. Furthermore, I was able to reduce the selection problem originating from a student’s choice for each of the government programmes (Ashenfelter, 1978). Policies directed at multicultural youth are commonly offered as a package rather than piecemeal (Ministry of Health and Welfare, 2008). Thus, my analysis can provide useful information about the basic formulation of multicultural youth policies.

10. This is partly due to a ceiling effect since some multicultural youth already felt proud of themselves.

11. Matching is a non-parametric estimator. Thus, it does not require any assumptions about functional forms and error term distributions, which is a major advantage. However, selection of observed variables requires a strong assumption. To apply the propensity score matching, I needed to control for all determinants driving the selection of the treatment (here, the government programmes) and affecting the dependent variable (here, national identity formation). This is called the conditional independence assumption (CIA): $\left({\mathrm{Y}}_{\mathrm{T}=0},\hspace{0.17em}{\mathrm{Y}}_{\mathrm{T}=1}\mathrm{\perp }\mathrm{T}|\mathrm{p}\left(\mathrm{X}\right)\right)$ where ⊥ implies independence; T denotes the treatment; ${\mathrm{Y}}_{\mathrm{T}=0}$ and ${\mathrm{Y}}_{\mathrm{T}=1}$ represent the outcome of the controlled and the treated students, respectively; and $\mathrm{p}\left(\mathrm{X}\right)$ is the propensity score calculated using a number of variables $\mathrm{X}$ which can overcome the curse of dimensionality. The dependent variable Y is independent of the treatment variable $\mathrm{T}$ if I controlled for observed covariates $\mathrm{X}$ (Rosenbaum & Rubin, 1983). Even though the matching estimator had low bias if the dataset contained a number of variables related to the treatment, I needed to provide a stronger argument, since I lacked the knowledge of how national identity is developed (Smith & Todd, 2005). Thus, I used the method as a robustness check. In this study, I used one-to-one nearest neighbour matching.

12. This is based on three important assumptions: ${\mathrm{T}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$ and ${\mathrm{Z}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$ need to be closely correlated (correlation), the instrument ${\mathrm{Z}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$ affects the dependent variable only through the treatment ${\mathrm{T}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$ (excludability), and ${\mathrm{Z}}_{\mathrm{i},\mathrm{j},\mathrm{t}}$ and ${\in }_{\mathrm{i},\mathrm{j},\mathrm{t}}$ are orthogonal (exogeneity). Caution is needed when interpreting this result since I faced limitations regarding whether the assumptions for the instrumental variable method were satisfied. Due to this limitation, I present the estimation result for a robustness check.

Additional information

Notes on contributors

Hoyong Jung

Hoyong Jung is an Assistant Professor in the Department of Economics, Kookmin University, Seoul, Korea (from September 2020). In his research he applies quantitative methods to economic and social issues in Korea. He has published in Asian Survey, Asia Pacific Education Review, Korean Journal of Policy Studies, and OMNES: The Journal of Multicultural Society.