A group of scientists believes that they have found a cheap and easy way to improve refugee placement: an algorithm that picks where people should be placed based on their chances to find employment. But people familiar with the placement process are skeptical.
What is the best way to distribute refugees across a country? Many experts familiar with the distribution processes believe there is significant room for improvement in the European Union and within its member states, including Europe's biggest host country, Germany.
"The reality of current distribution of refugees is in many ways at odds with what an ideal distribution would look like," Bernd Mesovic of Pro Asyl told InfoMigrants.
Now, scientists from Stanford University's Immigration Policy Lab and ETH Zurich believe that they have found a cheap, easy-to-be-implemented mechanism that could help to significantly improve refugee distribution.
"Host countries' current procedures for determining how to allocate refugees across domestic resettlement sites do not fully leverage synergies between refugees and geographic locations," they write in a paper recently published in Science.
The scientists have developed an algorithm that they believe can improve the chances for a refugee to find work significantly, based on sample data from the US and Switzerland.
Dominik Hangartner, one of the authors, explains: "We know that the place a person is assigned to matters greatly for the integration success. So we tried to create an algorithm that helps authorities to figure out to what city or county someone should be sent to."
The scientists collected data on age, gender, native language, education and job experience of refugees and asylum seekers in Switzerland and the US, as well as their date of entry into the country, the location they were assigned to live in and their employment status. Based on the data, they developed an algorithm meant to determine where refugees will have the best employment opportunities.
73 percent more likely to find a job
Using the algorithm, the scientists claim, the employment rate among refugees in the US would be 41 percent higher on the 90th day after their arrival. In Switzerland, third-year employment among people who were granted a temporarily permit to stay in the Swiss would be 73 percent higher, they say.
"We are trying to bring much needed empirical evidence and innovative ideas to a policy field often dominated by heated ideological debates," Hangartner told InfoMigrants. "The goal of our study and the policy we suggest is to address one of the fundamental challenges that many countries currently face: the economic and social integration of a large number of asylum seekers and refugees."
He and his colleagues believe that their algorithm could not just improve procedures in the two countries they studied, but also in other countries, including Germany. While most policies that can help newcomers integrate – such as language classes and integration workshops – are labor-intensive and costly, they believe their algorithm can help government agencies allocate refugees at little to no extra costs.
Difficult to implement in Germany?
However, some experts contacted by InfoMigrants believe that integrating an algorithm into the existing national migration agency structures would be quite difficult, especially in Germany.
"The federal employment agency only collects asylum seekers employment and education information after they have been assigned a state to live in," Birgit Naujoks, the director of the refugee council in Germany's biggest state, North Rhine-Westphalia, told InfoMigrants. "You would have to completely restructure the whole system." Hangartner, however, says that even without knowing someone's work history, his algorithm could already optimize someone's placement, simply based on their country of origin, gender and age.
In Germany, asylum seekers are assigned to one of the 16 states based on a quota – the number of people each state has to take in based two-thirds on their tax revenues, and one-third on the state's population size. The German system is similar to the Swiss one, where refugees are assigned randomly to counties according to a proportional distribution key. Though asylum seekers and refugees in Germany can ask to be placed in specific states or counties, Birgit Naujoks says these preferences are seldomly taken into consideration.
Refugees are required to stay in the state – and at times even the municipality – they were assigned to for up to three years. Sending someone to a place where they can easily integrate can thus be of great importance for both the asylum seeker and the German state.
Refugee placement needs improvement
The current system of mostly randomly assigning refugees a place of residence could be dramatically improved, according to the refugee advocates and experts InfoMigrants spoke with.
"Currently, refugees are treated like they are not mature enough to make their own decision," Bernd Mesovic of Pro Asyl. Birgit Naujoks said that "individual preferences should be taken into consideration much more."
Mesovic praised the idea of optimizing a refugee's placement based on employment opportunities, saying that "of course having a job is very important for people."
Naujok was more skeptical on whether employment should be a key criteria. "Putting the economic value of refugees front and center, which is something that has been done a lot recently, is difficult," she said. "Because it puts the focus not on the refugee's interests, but on how others can benefit from the situation."
She added that whether someone would have access to a support structure – family or just an ethnic community from their home country – should be key criteria when placing an asylum seeker.
Scientist Hangartner said, however, though their algorithm is designed to maximize refugees employment opportunities, it could potentially also take family connections into consideration in the future – and that it already, indirectly, takes other important factors into consideration. "Employment opportunities are often linked to other factors that can improve integration. People are more likely to find a job in a place where there's an ethnic community from their country, so our algorithm would sent them to those places."