Germany 2050
Sweden 2050

The projections of future demographics are made using a cohort-component model. The way such a model works is that one divides the population into different groupings (cohorts) and then study the individual components that affect that group over time. We can for instance divide the population into male and female and those two groups further into let’s say 5-year chunks. Then there would be groups of females aged 0-5, 5-10, 10-15 and so on. After that, we ask the question how the number of women in each group changes over time.

Let’s consider the group females aged 5-10 years. There are only two ways someone can enter that group and that is either by growing older (i.e someone that was in the 0-5 year old group becomes 5+ year old) or by immigration. Someone can only exit the group by growing older, dying or emigrating. To track the change over time of that group we thus need data covering immigration, emigration and death rates. We also need to know how many are in the 0-5 year old group, so we can estimate how many age into the group. The same procedure is then repeated for the 10-15 year old group, the 15-20 year old group and so on.

When we reach the final age group then people only exit the group by dying or emigrating since there is no further group to age into. The first group, 0-5 year old, is also special because one enters it by either being born or immigrating. For that group we thus need fertility rates to be able to calculate how many are born into it.

A cohort-component model works better the more finely divided the groups are and the more specific the data is. In the ideal case, we would have cohorts for every year of age and statistics for death rates, immigration, emigration and fertility divided by age. In reality, one almost never has that kind of fine grain data.

Data and assumptions

The assumptions we made and the data we had access to for our model are as follows:

The cohorts we have worked with are in the age groups 0-16, 16-36, 36-65, and 65+ separated into gender and further separated into people with migrant background or people with German background.

We have data for total fertility for German women and women of immigrant background.

We have age-divided birth rates for all women in Germany. We thus had to assume the age distribution is the same between German women and women of immigrant background. In other words, while the total fertility is different we assume the percentage of births by women of German or immigrant background is the same within a certain age group. In other words, let’s say 1000 German women give birth to 1000 children and 1000 immigrant women give birth to 1500 children, in both cases we assume roughly 77% of the children will have been born by women in the age group 16-36.

Here we have probably underestimated the net future migrant population. In the 16-36 age group, likely there are many more migrants in the younger part and many more Germans in the older part of the interval so we underestimate childbirths for migrant women at the same time as we overestimate childbirths for German women. 

We have death rates divided by gender and age but not by German or immigrant background so we assume it is the same between the groups.

We have immigration and emigration rates by age and by German or immigrant background. The age groups are however slightly different (0–14, 14-35, 35-65, 65+) so minor adjustments had to be made to make them fit with our age divisions.

All the data was collected from the Federal Statistical Office (Statistisches Bundesamt) Genesis-Online database except for the population data, which came from a special analysis of the German federal statistical bureau relating to the population with foreign origin, subdivided by sex and age, at the level of regional statistical districts. 

Future scenarios

Based on the above described data and assumptions we have calculated the future demographics of Germany in four different scenarios that all start at the year 2025 and end at 2050 and the details for the scenarios are as follows.

Scenario no change. Immigration continues at the same rate as the average since 2017 (985 887 people per year), emigration continues as the average fraction since 2017 (1.65% of the immigrant population within Germany emigrates each year and 0.33% of the German population). The reason we assume a fixed rate for immigration and a fractional rate for emigration is that immigration is mostly dependent on the population residing outside of Germany and thus not greatly influenced by demographic changes within Germany while emigration is dependent on the total number of people within Germany.

Scenario up. Net immigration is increased by 400 000 people per year from 01 JAN 25 as suggested by the German government. Emigration, as in the above scenario.

Scenario down. Immigration is set to zero and emigration of people with immigrant background is increased by 100 000 per year from 01 JAN 25.

Scenario freeze. Continued immigration/emigration until 01 JAN 25, thereafter full stop immigration and emigration and focus on population change only concerning births and deaths.

In all four cases, the difference in fertility rate between German women and women of immigrant background is assumed constant over time.

No intermingling between Germans and immigrants is taken into account.

Possible flaws of the model

The age division is rather blunt, it would have been better to have the age distributed in at least five year intervals. We know that the immigrant population is younger on average than the German population and some of the impact of that difference is lost in the large groupings.

Especially important is the group of 16-36 year old women since it covers the main part of women’s fertile window. If the intervals had been more finely divided then likely we would have seen an even larger difference in birth rates between immigrants and Germans.

In the same way, we know that immigrant women tend to get children at a younger age but since we do not have any solid data on that we had to assume the age distribution is the same. Finer age divided fertility data would likely have shown a starker difference in childbirths.

In the same vein, death rates are unlikely to be the same between Germans and immigrants since we know that death rates are connected to economic factors. Likely, the immigrant population has a higher death rate in the younger age groups.

Conclusions

As a wise man once said, making predictions is hard, especially about the future. These scenarios do not show what the future will be; they educate us on what the future might bring based on our choices today. At the current trajectory, already by the year 2035 more children will be born of immigrant background than German background.