In the last two decades, the number of applicants for political asylum has increased steeply (Hatton 2020, 2021). With the exception of the crisis years of 2015 and 2016, more than half of asylum claims lodged in Europe have been rejected. While individual countries have faced political pressures to reduce the inflows by imposing strict rules, the EU has been developing a Common European Asylum System (CEAS) through a series of directives, which were subsequently transposed into national legislation.
The Qualification Directive (2004) and its recast version (2011) relate to the criteria for granting refugee status or subsidiary protection. The Procedures Directive (2005) and its recast version (2013) relate to the procedures used in assessing asylum claims. These set out minimum standards to protect the rights of refugees.
The directives also aimed to harmonise policy across European countries and mitigate the so-called asylum lottery. In the presence of other influences on recognition rates, such as the increase in the spread and intensity of persecution in origin countries, it remains unclear how far these goals have been achieved.
Trends and differences in recognition rates
Figure 1 shows the overall recognition rate for 20 countries that account for 90% of all asylum applications in Europe from 2000 to 2018. These are ‘first instance’ recognition rates, which exclude decisions resulting from appeals, administrative reviews, or repeat applications. The recognition rate includes all forms of recognition, from full refugee status (under the 1951 Refugee Convention) to weaker forms of recognition that confer fewer rights.
Figure 1 Asylum recognition rate for 20 European countries, 2000–2018
Note: This is the overall recognition rate (not the average of country rates) for 20 countries: Austria, Belgium, Czech Rep., Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Luxembourg, Netherlands, Norway, Poland, Portugal, Spain, Sweden, Switzerland, and the UK.
Two features stand out. One is that, except for the crisis years of 2015 and 2016, less than half of the decisions resulted in some form of recognition. The average of 30% would be around 7 percentage points higher if successful appeals and reviews were taken into account. But even so, more than half of all applications are ultimately rejected. Rejected applicants are required to leave the country, although many vanish into the informal sector.
The second is that, contrary to popular belief, there is a strong upward trend in the overall recognition rate. Although the crisis years stand out, recognition rates had been rising even before the 2011 Arab Spring, and they remained historically high even after the immediate crisis years. This could be accounted for by the spread of civil wars and the growing intensity of human rights abuse, but policy reforms may also have contributed.
Looking across European countries, Figure 2 shows that there is wide variation in average annual recognition rates for 2003–18. This is the so-called asylum lottery, which contrasts with the EU’s long-standing objective that “no matter where an applicant applies, the outcome will be similar”.
Yet, despite efforts at reform, the standard deviation of recognition rates increased from an average of 13.6 in 2003–07 to 20.2 in 2014–18. Even for individual origin countries, the dispersion across destination countries has not decreased (Hatton 2021: 28, see also Leerkes 2015).
Figure 2 Average recognition rates by country, 2005–2018
Analysing trends and differences
In Hatton (2021), I examine a panel of recognition rates for 20 European destination countries by 65 origin countries from 2003 to 2017. Exploiting the differential timing that these directives were transposed into national law, I find a mixture of positive and negative associations, some of which differ from those implied by the qualitative literature. But consistent with earlier quantitative studies (e.g. Neumayer 2005, Toshkov 2014), variables representing violence, persecution, and human rights abuse in origin countries are important.
The contribution of different groups of variables to the upward trend in recognition rates is illustrated in Figure 3. For each year, the first bar (in blue) is the year effects (compared with 2003 = 0) when no controls are included. The second bar (in orange) is the year effects when origin-country persecution and conflict are included. These increase less steeply, particularly from 2011, indicating that increased persecution accounts for about half of the increase in recognition rates.
The third bar (in grey) shows the year effects when the transposition of EU directives into destination-country legislation are also included. By 2016 the height of the grey bar is reduced by more than a third compared with the case with only origin-country controls. This suggests that the EU directives made a distinct, if modest, contribution to preventing what might otherwise have been a race to the bottom in asylum policies.
Figure 3 Estimated marginal year effects for the recognition rate (2003 = 0)
Notes: The heights of the bars are the estimated marginal effects of year dummies with different sets of controls. The whiskers are 95 % confidence intervals.
The estimated differences between countries are illustrated in Figure 4 as marginal country effects in percentage-point differences from France. The first bar (in blue) is the overall recognition rate, which incorporates both differences in origin-country composition and differences in recognition rates for a given origin country (as deviations from France). The second bar in orange) is the marginal country effects which include only origin-by-year dummy variables. The difference between the height of the blue and grey bars accounts for differences in origin-country composition.
Figure 4 Estimated marginal country effects for the recognition rate (deviations from France)
Notes: The heights of the bars are the estimated marginal effects of destination country dummies with different sets of controls. These are constructed as deviations from France, which is taken as the base because it includes the largest number of origin countries (64) and total observations (949). The whiskers are 95 % confidence intervals.
Some countries with relatively high overall recognition rates (relative to France) look much tougher when adjusted for origin-country composition. These include Belgium, Germany, Luxembourg, the Netherlands, Norway, and Sweden. But adjusting for origin-by-year effects does little to reduce the cross-country dispersion: the standard deviation of the origin-adjusted recognition rates is 14.7 compared with 15.7 for the overall rates. When recognition rates are also adjusted for policy (in grey), they differ only slightly from the origin-adjusted rates and the standard deviation decreases very slightly to 14.6.
What accounts for the persistent differences in recognition rates across European countries? Cross-sectional analysis (for 2013) suggests that institutional structures may be important. The existence of a border agency separate from the status determination procedure is negatively associated with the recognition rate. Similarly, an accelerated procedure for ‘manifestly unfounded’ claims is negatively associated with recognition rates.
While the interpretation of these correlations remains speculative, they are strong enough to suggest that cross-country variation in recognition rates depends not only on procedural rules but also on the structure of the agencies that implement them.
In the aftermath of the migration crisis, the European Commission produced a set of reform proposals, most of which were subsequently incorporated into the New Pact on Migration and Asylum (European Commission 2020). Among these are the transformation of the Qualification and Procedures Directives into Regulations, which would mean shifting from a set of goals which each country implements in its own way to directly binding and precisely specified laws. If implemented, these are likely to leave far less room for discretion at the country level, which might lead to substantial convergence in recognition rates. But in the absence of a fully integrated EU-wide bureaucratic framework, it seems unlikely that differences in recognition rates will be eliminated.
European Commission (2020), New pact on migration and asylum.
Leerkes, A (2015), “How (un)restrictive are we? ‘Adjusted’ and ‘expected’ asylum recognition rates in Europe”, Netherlands Department of Justice, Working Paper 2015-10.
Hatton, T J (2020), “Asylum migration to the developed world: Persecution, incentives, and policy”, Journal of Economic Perspectives 34 (1): 75–93.
Hatton, T J (2021), “Asylum recognition rates in Europe: Persecution, policies and performance”, CEPR Discussion Paper 16709.
Neumayer, A (2005), “Asylum recognition rates in Western Europe: Their determinants, variation, and lack of convergence”, Journal of Conflict Resolution 49(1): 43–66.
Toshkov, D D (2014), “The dynamic relationship between asylum applications and recognition rates in Europe (1987–2010)”, European Union Politics 15(2): 192–214.