Destination choice of asylum applicants in Europe from three conflict-affected countries

ABSTRACT

From the early 2010s onward, the number of asylum seekers in Europe has gradually increased, with 2015, in particular, experiencing a significant rise. These asylum seekers have mainly come from conflict-affected countries. In 2015, for instance, Syrians, Afghans and Iraqis accounted for the top three nationalities of first-time asylum applicants in the European Union. This study uses count data on Syrian, Afghan and Iraqi asylum seekers in 25 European countries over the sample period of 2001 to 2016 and investigates their destination choice within a framework of generalised linear models for modelling count data. The explanatory variables include economic and cultural factors, which the results show significantly, influenced the flows of asylum seekers. For instance, a high level of income per person employed in a European country was a common pull factor for Syrian, Afghan and Iraqi asylum seekers. The three groups of asylum seekers also shared a common preference for cultures in which people were individualistic, long-term orientated and tolerant of unorthodox behaviour and ideas.

KEYWORDS:

• Asylum seekers
• economic factors
• cultural factors
• generalised linear model
• negative binomial distribution

Acknowledgements

The author would like to thank Enago (www.enago.jp) for the English language review.

Disclosure statement

There is no potential conflict of interest.

Availability of Data

The datasets used in this study are available from the author on request.

Code availability

The codes are available from the author on request.

Notes

1. The web address of the Hofstede Insights is https://www.hofstede-insights.com.

2. The web address of the CSP is https://www.systemicpeace.org.

3. The data on $asylu{m}_{ds}$ are ‘inflows of asylum seekers by nationality’ and are available in the International Migration Database on the OECD website: https://stats.oecd.org.

4. Although Iceland belongs to the Schengen area, the country is located far out in the Atlantic Ocean and, hence, is excluded from the sample.

5. The data on $incom{e}_d$ are ‘GDP per person employed, PPP (constant 2011 international $)’ and are available from the World Development Indicators on the website of the World Bank: https://databank.worldbank.org/source/world-development-indicators.

6. The data on the unemployment rates are ‘Unemployment, total (% of total labour force) (modelled ILO estimate)’ and are available from the World Bank website as shown in the footnote 5.

7. The Hofstede’s cultural indices are available on the Hofstede Insights website shown in the footnote 1.

8. There is another Hofstede index. This is an ‘indulgence’ index that is unpublished for Afghanistan and Syria. This study ignores the index due to the data unavailability.

9. The Dynamic Gravity dataset 2005–2016 is available on the website of the United States International Commission: https://www.usitc.gov/data/gravity/dataset.htm.

10. The data are available from the CSP website shown in the footnote 2.

11. The monthly climate data for the world are available on the website of the National Centers for Environmental Information of the National Oceanic and Atmospheric Administration: https://www.ncdc.noaa.gov/IPS/mcdw/mcdw.html. The monthly data on temperature in some of the sample countries are available only for the subsample period.

12. The data are ‘battle-related deaths (number of people) and are available from the World Development Indicators on the World Bank website shown in the footnote 5.

13. Hilbe (2011, 2014)) showed that a NB model and a PIG model can be mathematically derived as a mixture of Poisson and gamma distributions and that of Poisson and inverse Gaussian distributions, respectively.

14. The data on the land areas are available from the World Development Indicators on the World Bank website shown in the footnote 5. For some of the destination countries, the figures slightly changed over the sample period.

15. A NB model needs to be extended to a zero-inflated NB model when there are an excessive number of zero counts. However, Vuong’s and Clarke’s tests are also performed comparing the NB model with its zero-inflated version for each group of Syrian, Afghan and Iraqi asylum seekers and reveal that the NB model is preferable to the zero-inflated NB model.

16. Equation (2)(2) $ln{\mu }_k={\mathbf{x}}_k^{\prime }\mathit{\beta }$(2) implies that multiplying the estimated coefficient of $incom{e}_d$ by the variable’s mean yields the elasticity of $asylu{m}_{ds}$ with respect to $incom{e}_d$ at the means of all the variables. The elasticities in Table 4 are computed using $incom{e}_d$’s mean of 0.825.

17. Given that $unemp{l}_d$is measured in percent, Equation (2)(2) $ln{\mu }_k={\mathbf{x}}_k^{\prime }\mathit{\beta }$(2) implies that multiplying the estimated coefficient of $unemp{l}_d$ by the mean of $asylu{m}_{ds}$ yields the average marginal effect.

18. https://hi.hofstede-insights.com/national-culture

19. https://hi.hofstede-insights.com/national-culture

20. http://www.impaladatabase.org/

21. For instance, using the data on Afghan-born population in each European country decreases the number of observations for estimating the Afghan model from 389 to 155.

Additional information

Funding

This research is financially supported by the Japan Society for the Promotion of Science [JSPS KAKENHI Grant Numbers: JP16K03683 and JP20K01698] and the Murata Science Foundation.

Notes on contributors

Tomoya Suzuki

Tomoya Suzuki is a full professor in the Faculty of Economics at Kansai University in Japan. He received his PhD from the Australian National University. His research focuses on issues in open economy macroeconomics.