Improving refugee integration through data-driven algorithmic assignment

See allHide authors and affiliations

Science  19 Jan 2018:
Vol. 359, Issue 6373, pp. 325-329
DOI: 10.1126/science.aao4408

eLetters is an online forum for ongoing peer review. Submission of eLetters are open to all. eLetters are not edited, proofread, or indexed.  Please read our Terms of Service before submitting your own eLetter.

Compose eLetter

Plain text

  • Plain text
    No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Author Information
First or given name, e.g. 'Peter'.
Your last, or family, name, e.g. 'MacMoody'.
Your email address, e.g.
Your role and/or occupation, e.g. 'Orthopedic Surgeon'.
Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.
Statement of Competing Interests

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Enter the characters shown in the image.

Vertical Tabs

  • RE: Improving refugee integration through data-driven algorithmic assignment
    • Tommy Andersson, Professor in Economics, Lund University
    • Other Contributors:
      • Alessandro Martinello, Assistant Professor in Economics, Lund University

    The article in Science (1) studies algorithmic refugee matching. This is a recent and rapidly growing field inspired by classical market design research on, e.g., school choice, kidney exchange and entry-level job markets. The main contribution of (1) is to provide governments and NGO’s with practical and easy-to-implement data-driven tools. The article specifically considers refugee resettlement problems, i.e., problems of transferring refugees to countries that has agreed to admit refugees and ultimately grant them permanent settlement. There are, at least, two natural extensions to this approach that are left for future research.

    First, refugee resettlement problems are very structured. At the time of resettlement, countries admitting refugees know the exact number and characteristics of incoming refugees (e.g., education, origin, etc.). However, in 2016, only 190,000 out of 65.6 million refugees were resettled according to (2). Most refugees arrive to countries without being resettled, and receiving countries ignore their number and characteristics in advance. The next generation of algorithmic refugee matching models must thus be able to cope with this lack of information, and assign refugees to localities within a country on-the-spot and directly upon arrival.

    Second, the optimality criterion used in (1) maximizes the global average of the probability that at least one refugee in each family gains employment. While this criterion is reasonable, other c...

    Show More
    Competing Interests: None declared.
  • When people get lost in big data

    “Improving refugee integration through data-driven algorithmic assignment” is the title of a recently published article by Kirk Bansak and colleagues. The authors developed an algorithm to assist national administrative agencies in distributing refugees and asylum seekers efficiently. It works as a “matchmaker” between the refugee and the labor market. On the basis of personal traits and characteristics (such as country of origin, language proficiency, sex, and age), an algorithm estimates where in the receiving country the refugee is most likely to find employment.
    Although the authors provide some substantial new and meaningful insights into the possibilities of big data and machine learning for research on migration and integration and its application, I want to raise some general concerns about the theoretical assumptions made in this article. Generally speaking, the manner in which this research is carried out is a reflection of societal perceptions of the target population (in this case, refugees). I would therefore like to discuss some of the theoretical presumptions made in this article that effectively deny refugees’ agency, that is, their ability to act and decide as autonomous individuals.
    The first point made by this article is that labor market integration is the key to successful and sustainable societal integration. Other factors, such as access to psychological support and medical treatment, access to infrastructure in rural areas, and subjectiv...

    Show More
    Competing Interests: None declared.

Navigate This Article