Visualizing Future Migration Scenarios for Europe

Migration has been integral to Europe since its inception. International migration has been a highly-debated topic in recent years and accordingly received high political awareness. A crucial input to effective migration policymaking is information on the likely size and composition of migration flows to Europe under different scenarios in the future. 

Today, approximately 3.4% of the world’s population are international migrants, meaning that they have left the country they were born and settled in a new country. In Europe, the share of the population is much higher, 10.6%. Among countries, this ranges from 1.5% in Roumania to 47.2% in Luxembourg in 2020. This includes persons who have migrated from one European country to another and those who have migrated to Europe from outside. 

Migration to Europe was developed under different scenarios. These mainly include factors that influence migration trends that are more predictable such as natural population change which may create migration pressures in sending countries or demand for labour migrants in receiving countries, differences between wages and income in origin and destination countries as well as the size of migrant communities in destination countries (diasporas). Some drivers of migration are difficult to foresee, such as wars, natural disasters, and economic or political shocks. During the course of this project, the COVID-19 pandemic and the start of the war in Ukraine greatly impacted migration patterns in Europe. 

FUME focused on understanding migration patterns, motivations, and modalities at multiple geographical scales, from international through regional to the local, and on imagining possible futures.  

To enhance this analysis and to fill gaps in existing knowledge, case studies were carried out in immigrants' origin countries addressing different aspects of migration propensity, intentions, and decisions. For the study, four important countries of origin were chosen: Senegal (representing Sub-Saharan origin countries), Iraq ( Middle East origin country), Tunisia (migration transit country with a long tradition of migration to EU countries), and Ukraine (new origin country of immigration to EU from Eastern Europe). 

To better understand the consequences of migration, such as residential segregation, integration, and multiculturalism, case studies were organized in chosen cities of destination countries in the EU Member States, with representatives from different parts of Europe – North (Copenhagen), West (Amsterdam), South (Rome) and East (Krakow). 

Age, sex, and level of education are key characteristics of migrant populations that are shown in the projection scenarios. Migrant populations typically differ from native populations along these characteristics and alter the overall populations in destination countries and regions. Understanding these differences is crucial to effective migration policy-making. 

The following sections visualize migration to Europe under different scenarios at different geographic levels moving from the global to European to national and finally to the city level. 

One of the main objectives of FUME is to understand international migration flows and patterns under several migration scenarios for Europe. To achieve this, we developed a multi-dimensional demographic model with population disaggregated by age, sex, education, and country of birth. Since migration rates tend to differ by level of education, migrants’ level of education was factored into the scenarios. 

The migration model accounted for several important variables and mechanisms that shape global migration patterns: income levels in the destination country and in the origin country, with emigration rates depending non-linearly on the latter; diasporas in the destination country; and the distinction between emigration, transit migration, and return migration. It was extended to account for the effects of education and age, thereby covering multiple dimensions of population heterogeneity. 

Migrants were considered highly skilled if they have at least had an upper-secondary education. Therefore, highly skilled individuals should have at least around 10-12 years of schooling. Since college graduates are included as well, the average years of schooling are raised even further. Low-skilled persons can have between 0-10 years of schooling. 

The set of quantitative migration scenarios is defined as follows: 

• Benchmark scenario: This scenario is identical to SSP2, a “Middle of the Road” scenario which assumes that social, economic, and technological trends will follow historical patterns in the world. Inequalities in development and income between different world regions will persist and the world will only make slow progress towards achieving sustainable development goals. The impacts of the Covid-19 pandemic are considered in this scenario, but not the effects of the war in Ukraine. 

• Ukraine-war scenario: This scenario is like the Benchmark scenario but uses International Monetary Fund growth rate estimates until 2027 (International Monetary Fund 2022). For the years 2027 to 2032, a linear transition back to growth rates assumed under the Benchmark scenario is assumed. 

• Recovery in Europe scenario: This scenario is identical to the Ukraine-war scenario with the exception that European countries transition towards the SSP which they have the highest growth rates after 2027. Developing countries and emerging economies, by contrast, are assumed to transition towards the SSP in which they have the lowest growth rates. These might be different SSPs for different countries. Industrialized countries outside of Europe are assumed to transition towards SSP2. 

• Rise of the East: This scenario assumes the opposite development as defined in the Recovery in Europe scenario. European countries transition towards the SSP which they have the lowest growth rate after 2027, while developing countries and emerging economies transition towards the SSP which they have the highest growth rates. 

• Intensifying global competition: This scenario is like the Ukraine war scenario, but it assumes an economic ‘catching-up’ trend of poorer countries. Concretely, it is assumed that countries with a projected gross domestic product per capita (GDPc) below 15.000 USD in 2040 will linearly transition to 15.000 USD in 2040 and afterward grow with SSP2 growth rates. All other countries will grow in the same way as in the Ukraine-war scenario.

• No migration scenario: In this purely hypothetical scenario, it is assumed that no migration to, from or within Europe will occur in the future. This scenario serves to highlight the contribution of migration to demographic growth. 

In this section, we present a dynamic model of global bilateral migration that accounts for these mechanisms. By simulating flows separately for each combination of origin, destination, and place of birth, individual migrant communities are explicitly represented, and changes in their size feedback on flows. The concept of the diaspora effect is expanded to also account for transit migration flows. Emigration from poor countries is constrained in accordance with observed emigration rates. Return migration is modeled explicitly as a function of migrant stocks. We calibrate the model on a global dataset of bilateral flows; demonstrate its performance for past levels and trends in migration; and project future migration flows under five different Shared Socioeconomic Pathways. 

We now initialize the model with year-2015 stocks to generate future scenarios, or conditional predictions, of bilateral migration flows. Because the model is not perfect and does not account, for instance, for refugee movements, the year-2015 migrant stocks resulting from our 25 years of historical simulation are different than observed. Therefore, our future simulations of flows generally start from a different level than where the historical simulation ended. The smaller the discontinuity between the two sets of simulations for a given region or country, the higher the agreement between the historical simulation and historical evolution of those bilateral migrant stocks that are most relevant for predicting this region’s or country’s flows.  

We run the model until 2100 using projected GDP changes under the five Shared Socioeconomic Pathways and projected natural population change from the UN World Population Prospects 2019 zero-migration variant. Again, an additional simulation with constant year-2015 GDP serves for comparison. This constant-GDP simulation shows continuously rising net migration throughout the 21st century in all world regions and most large countries. The SSP simulations diverge from this. At the level of world regions, all SSPs lead to lower (absolute) net migration flows than the constant-GDP simulation. The projections differ markedly between the different SSPs: Under SSP 1, SSP 5 and, somewhat more slowly, SSP 2, net migration flows approach zero by the end of the century in all world regions and many countries. On the other hand, under SSPs 3 and 4, net migration flows keep rising throughout the century in Africa, Europe, and Oceania. In Southeast Asia, West Asia and, less pronounced, Latin America, SSP 3 leads to a peak around 2060, followed by a decline in net migration flows; the peak is somewhat later in North America and South Asia. In East Asia, SSP 3 and SSP 4 lead to substantial net immigration by the end of the century. This is because of declining net emigration from China and relatively high net immigration to Japan and South Korea. Only the Former Soviet Union region exhibits declining net migration under all SSPs. 

In 2015, the total population of EU28 was 500.5 million. Of those, 90.6 percent were living in the European country in which they were born, 5.8 percent were residing in a European country other than the one in which they were born, and 3.6 percent were born outside of Europe. In 2015, the country with the highest percent foreign born was tiny Luxembourg, with 43 percent born outside the country. This was followed by Cyprus, Austria, Ireland, Sweden, and Belgium, all with more than 15 percent of their populations being foreign born. 

There was a geographic pattern to those countries having the lowest percents foreign born with these countries being in the east and most being among the new EU accession countries. These countries with less than 6 percent foreign born are Lithuania, Finland, Hungary, Czechia, Slovakia, Bulgaria, Poland, and Romania. Many of these countries have been countries of emigration in recent decades. 

The population of Europe is projected to growth rather slowly, reaching 511 million in 2050. This growth will be due entirely to immigration as there is projected to be more deaths than births because of low fertility and an aging population. 

According to the benchmark scenario, the percent foreign-born in the EU28 is projected to slowly increase during the projection period until 2050. The percent foreign-born for all of Europe is projected to be 10.2% in 2020, 11.3% in 2030, 12.5% in 2040, and 14.0% in 2050. 

For the EU28, the percent foreign-born is projected to increase by 4.3 percentage points between 2015 and 2050, to 14.0% from 9.6%. All but a few countries are expected to have increases in their foreign-born populations. Only Hungary and Poland are projected to have slight declines in their percents foreign-born. The largest projected increases in percent foreign-born are projected to be in the Mediterranean island countries of Malta (increasing from 11% to 38% foreign-born) and Cyprus (from 21% to 34%). 

By 2050, with the projected concentrations of immigrants into northern Europe, the geographic pattern of percent foreign-born by country will become distinct. Aside from Malta and Cyrus, the highest percent foreign-born will be in Luxembourg, Sweden, Ireland, Belgium, Austria, Estonia, and Great Britain, all with 19 or more percent of their populations born outside these countries. The smallest percents foreign-born will continue to be in eastern EU countries - Lithuania, Czechia, Hungary, Slovakia, Bulgaria, Romania, and Poland– all with less than 10% foreign-born. 

In the absence of international migration, the population of EU28 is expected to decline and age faster. The population will decrease from 500.5 million in 2015 to 478.2 in 2050, a 5% drop due to a negative natural increase, i.e., more deaths than births. Under the no migration scenario, EU28 will not receive migration from non-EU countries and their children in the future, while in the benchmark scenario, the population of EU28 is expected to remain stable with a slight increase (511.6 million) due to positive net migration flow. 

In the following, we present models at the national level. More than two-thirds of EU28 countries, except in Northern Europe and Cyprus, will experience depopulation under the zero migration scenario. The scenario expects a rapid population decline in Bulgaria (20%), Italy (13%), and Germany (11%) by 2050. However, about ten countries in Northern Europe and Cyprus will continue to grow with a maximum of 20% in Ireland by 2050, followed by Luxembourg (14%), France and the UK (7%), and Sweden (6%). 

The figure shows an example of the results that can be obtained with the new models. The chart represents the estimated stock of intra-EU migrants in the EU member states and the UK in 2019. It is based on census data, official administrative data submitted to Eurostat, data from the Labour Force Survey, and Facebook. An animated version of this figure, showing changes in migration stock between 2011 and 2019, is also available on the FUME website. The width of each chord indicates the size of the migrant stock (in millions), while the arrowhead points to the country of residence. The color represents the region of origin. These new estimates of migrant stocks show that Romania and Poland are the two most important sending countries in the EU, with 2.5 and 3 million migrants from these countries living in another EU member state or the UK. Germany and the United Kingdom are the most important receiving countries of EU migration, each being home to around 2.6 million EU migrants. 

An important contribution of the FUME project was the inclusion of a level of education into the migration scenarios. According to migration theory, more educated people tend to have higher migration propensities than less educated persons. This has important implications for sending and receiving countries. Europe can expect to be the recipient of higher educated populations. The overall level of education among the migrants was higher than natives in 2015. Among 20-49 years old (younger adults), non-EU-CoB migrants had the highest proportion (40%) of post-secondary educational attainment, followed by EU-CoB migrants (34%) and the natives (28%). 

As mentioned before the population and migration are projected by age, sex, education and country of birth. In the figure, we present the population composition of 27 EU member states and the UK by age, sex, education, and broad region of birth in 2015 (base year), and the projected composition according to the Benchmark and No migration scenarios in 2050. The upper row of the figure shows the population composition in 2015 for the native, non-EU born and Non-native but born in another EU country populations. Both migrant groups have a younger age pyramid than the native population. The second row shows the projected composition in 2050 according to the Benchmark scenario, where a significant change in the native population structure (compared to the 2015 population) is visible with a higher proportion of population in the older age groups. There are also differences in the projected migrant population pyramids. Both the number of non-EU migrants and their proportion of population in older age groups are increased. However, the size of non-nativeEU born migrant population is decreased and the population pyramid shows signs of aging for this group. Finally, in the last row, no migration scenario, we see a significant decrease in the size of migrant population with very small proportion of migrants at working age groups. As the children of migrants born in the destination country are recorded as the native-born population, there are no reported migrants in younger age groups. 

From the policy brief based on the migration scenarios, several key messages emerge. 

Immigration to Europe will continue. According to all scenarios that were calculated with the FUME migration and population models (except the hypothetical ‘no migration’ scenario) the EU will remain a region with positive net migration in the future, with more people moving to the EU than emigrating. European societies will therefore become more diverse and policy makers should plan accordingly.  

Migration is key for future demographic developments in Europe. In the absence of future migration, the population in Europe is projected to decline in the future, and processes of population aging would occur at an even faster rate. Future immigration to Europe can buffer trends of population ageing and population decline may be avoided, depending on the size of future immigration flows. 

Migration is highly volatile. Migration flows are influenced by wars, pandemics, economic and other crises that are difficult to predict. Policymakers should expect and plan for volatile migration trends in the future, including temporary spikes in migration in crisis situations, and ensure that the needs of migrants can be met. A focus on creating resilient reception and integration systems is required. European cities need to be able to absorb sudden increases in migration that are driven by global crises in the future. This includes a better mapping of publicly owned housing and properties that could be used to host migrants in case of urgent need.  

Climate change may become a more important factor for migration trends in the future. Research currently suggests that migration triggered by climate change is a rather marginal phenomenon. Nonetheless, if climate change leads to greater inequality between world regions, and affects economic growth potentials and income opportunities in less developed countries, this may influence migration flows in the future, at regional but potentially also at a more global level.   

There is a lack of reliable migration statistics. Despite its importance for demographic development in Europe, we face challenges in measuring and analyzing migration trends. Migration is notoriously hard to capture with statistics, since some migration is undocumented, and definitions of migrants and migration change over time and may differ across institutions and countries. Circular and temporal forms of migration are particularly poorly captured by existing data sources. Traditional migration statistics based on register, survey, or census data are also often published with delay, often have their own deficiencies, and may not always be available for all EU member states. To improve our knowledge of migration flows and migrant populations, new forms of data emerging from social media platforms such as Facebook or Twitter can be used. These data have the advantage that they are instantly available. Nonetheless, these big sources of data have disadvantages as well. For example, they face selection bias since not all migrants actively use social media.  

More efforts and funding are therefore needed to combine different types of migration statistics. A combination of newer and more traditional data sources is required to improve our knowledge of migration trends. A collaboration between researchers, statistical offices, and policymakers is needed to make more progress in this area.

Increased efforts should also be placed on unifying and harmonizing migration registration systems across Europe. Border statistics that document often rapidly changing migration pressures at EU’s borders need to be improved as well. Recent experiences in the context of the Ukraine war have shown that it is possible to move fast, combine data sources, and accelerate coordination between relevant institutions and authorities to gain a better overview of rapidly changing migrant flows. Even though some progress has been made to better capture migration to and from Ukraine, there is a lot of room for improving and combining data sources to also capture migrant stocks and flows from other world regions. 

Nowcasting of migration will become more important. The war in Ukraine has led to one of the largest displacements of people since the Second World War. This experience powerfully shows that migration trends can change rapidly in crisis situations. In such contexts, it is important that information about shifting migration flows and estimations of the number of arrivals can be provided at very short notice, so that public authorities and civil society can react and mobilize support. Methods to ‘nowcast’ migration flows, as they have been developed and tested in the FUME project, can help to provide this type of information. Nowcasting measures use large data sources such as Facebook data to indicate almost in real time where and when people move. Further support and funding should be allocated to develop these methods and make them available and applicable in future economic, environmental, or political crisis situations that may lead to rapid shifts in migration flows.