Summary
Germany saw a drop in the number of marriages per capita starting with the first lockdowns in early 2020, followed by a surge in live births throughout 2021.
Marriage rates stayed below the lagging 5 year moving average all throughout 2021, which and correlate with a drop in birth rates 43 weeks later.
Live births per 43-week-lagging marriages show a strong seasonal pattern. Calculating an excess time series from this variable reveals that the drop in 2022 excess birth rates is much less pronounced than the preceding drop in excess marriage rates.
While changes in excess birth rates correlate positively with vaccination rates, the regression results are of low significance and considered inconclusive, possibly due to vaccination rates across German federal states falling into a very narrow range and other factors apparently having a large effect on birth rates.
Methods
Data Sources
I used a vaccination rate snapshot from January 11th 2022 that can be found here. All other data were provided by the German Statistics Office.
Downloads
Time Series
Time series representing marriages and births are based on monthly data spanning from 1991 to 2022. Population counts are based on official annual estimates for December 31st of the preceding calendar year (Dec 31st 1990 - Dec 31st 2021).
Because data for December 2022 is missing, these time series include 32*12-1=383 months. These monthly data were stretched to 32*52-4=1660 datapoints and treated as weekly data.
Excess Marriages Per Woman 15-49
Reference period: Preceding 5 years
Lag: 43 weeks
Excess Births Per Woman 15-49
Reference period: Preceding 5 years
Births Per Lagging Marriages
Lag Marriages: 43 weeks
Both time series are replaced by their own simple moving average with a period of 13 weeks (1 quarter)
Births are divided by lagging marriages
Excess Births Per Lagging Marriages
Reference period: Preceding 5 years
Results
I determined the lag of 43 weeks that I am applying to the marriage time series by visual inspection.
Take a look at how lagging marriages and births are lining up in 2022.
While large trends in births and marriages often diverge, short-term changes in marriages seem to be represented very well in births with a delay of 43 weeks.
Germany
Baden-Württemberg
Bavaria
Berlin
Brandenburg
Bremen
Hamburg
Hesse
Mecklenburg Western Pomerania
Lower Saxony
Northrhine-Westphalia
Rhineland Palatinate
Saarland
Saxony
Saxony-Anhalt
Schleswig-Holstein
Thuringia
Regressions Against Vaccination Rates
I will keep this very brief, since the results are inconclusive.
The variables specified in the title serve as y-variables, the rate of fully vaccinated residents on January 11th 2022 serves as the x-variable against which all regressions were run.
Hi Fabian,
My first substack posts in July last year looked at the issue of the dramatic fall in wedding rates in 2020-21 and a possible connection with the dramatic fall in birth rates in 2022:
https://lostintranslations.substack.com/p/why-is-there-a-sudden-drop-in-german
Interesting that you settled on a 43 week lag correlation - I found it very difficult to find data on the average time between first births and marriage. (Data on the average time between subsequent births and first births is more readily available. As is data on the proportion of live births which are a mother's first, second, third etc child.)
There´s a big difference in vaccination rates between West and East Germany. So you would expect less negative excess in East Germany. But that clearly isn´t the case. I wonder whether the confounder might be the proportion of immigrant young women. In West Germany it approaches 40% in some states whereas in the East it is much lower. On the other hand the proportion of vaccinated is much lower among immigrants. So maybe the number of immigrants in the West offsets the lower vaccination rate in the East? Mr. Spieker, has that thought occured to you and do you have any idea how to tease out these numbers?