The mathematics of immigration

For the sake of this paper, I will use the terms ‘migrants’ and ‘immigration' to encompass asylum seekers, immigrants and refugees. 

 

Immigration has significant implications for the economy, politics, legislation, and culture - not only for the departing and the receiving countries, but also the transit ones. The implications that immigration can cause are often portrayed as ‘doom and gloom’, especially in the media, when we may, in fact, be positively surprised if we analysed the mathematics behind immigration. 

 

At present, immigration is mostly studied and understood by social and behavioural science, and economics, but I would argue that we should make greater use of mathematics to better understand immigration and its impact, and underpin the immigration policies process with sophisticated algorithms, statistical modelling and Artificial Intelligence (AI). Some important elements such as the personal experiences and the emotional implications may only be captured by qualitative methods such as interviews or written reports, but even those may be translated and fed into an algorithm and interpreted by an AI. For example, we can now analyse the language used in some specific online forums and extrapolate themes and feelings, and from there build statistical models. 

 

While statistical and mathematical data are already widely available to policy makers, those represent only one of the many tools that inform the decision process, with the main influences coming from political parties and interest groups. This often leads to the media manipulating the public opinion against immigration by highlighting statistics that reinforce negative stereotypes. The debate then becomes heated and highly politicised, which in turn, legitimises inefficient migration policies. This causes a perverse negative cycle which affects not only the migrants but also the host communities: the more the public opinion is stirred against immigration, the less likely migrants are able to settle successfully in a community, and migrants’ integration is possibly the key to success in the immigration debate.

 

Another facet of the immigration topic that could be tackled more efficiently through a greater use of mathematics is how to geographically resettle migrants in the host country. At present, countries’ policies differ greatly in their approaches: some may decide it at national level while others may leave it to the local authorities and regions. The UK, for example, has introduced a policy called ‘Asylum Dispersal’,  which is intended to help distribute asylum seekers evenly across the country to avoid burdening communities but this allocating process does not take into consideration a key factor: how likely migrants are to be employed in a specific community. While employment rate is not the only success factor when evaluating immigration, employed migrants are more likely to contribute to the economy and create a more positive outcome both for the migrants and for the host community. 

 

The Immigration Policy Lab (IPL) argues that the resettlement policy can be improved by using a mathematical algorithm that matches each specific migrant with the community where they are more likely to find employment. The algorithm takes into consideration a range of characteristics from the two parties, from the migrant’s profile (age, ethnic background, skills levels, training needs and disposition etc.), and from the host communities, for example what type of employment opportunities are offered, schools’ availability, housing, local labour demand, etc). The algorithm is flexible and is continuously readjusted with new data to accommodate changes in circumstances of the migrants or the host community. The system has been a success: a US resettlement agency used the IPL algorithm on 30,000 refugees and employment increased by more than 25%.

 

The new economics of labour migration theory sees the decision to migrate as one taken within families, rather than by a single individual, and not so much as a way of increasing income, as previously thought, but as a way of diversifying family income and being more protected from external risks, such as droughts and natural calamities, or political instability. This is especially true for countries where there is no reliable insurance or credit market, so no temporary shield from such adversities. AI can be used to predict those sudden surges of human migration due to external negative events. For example, AI can forecast natural disasters by combining past weather patterns with current data and other variables. To calculate the intensity of monsoons, for example, AI makes use of trending words in the online activity of nearby communities. AI can collect huge amounts of data before natural disasters actually happen, instead of afterwards. Countries which are likely to be affected by the subsequent migration have then time to prepare, or even intervene beforehand, when possible. Of course, in order to actually make a difference, the AI predictions need to be followed by a willingness from policy makers to act. 

 

Mathematics could be used not only by policy makers, but by most sectors which come into contact with migrants. Schools receiving migrant children can use a statistical method used to show the correlation between two or more variables, to relate migrant children with their peers, taking into account standardised test scores, the academic level in the country of origin, language proficiency, parental education, ethnic background, time of arrival in the host country etc. Maybe surprisingly, migrant children do not lack behind their local peers in English and Maths, and where there are some differences between native students and non-native students, it seems to be caused by other underlying factors, which can be isolated through regression modelling.

 

There are of course many qualitative aspects of the migrants’ experience that cannot be quantified and should not be disregarded. At the same time, there are many facets of the immigration debate that could be greatly improved by weighing more heavily on mathematical and statistical analysis instead of heated and politicised debates. Mathematical-based research could offer valuable insights to study and understand the requirements of migrants and host communities and inform more effective policies, to the advantage of all parties involved. Utilised in tandem, qualitative and quantitative research methodology could comprehensively inform a more accurate and effective immigration policy, programming and media narrative.