Inferring the Prevalence and Age Dependence of Infection Fatality Rate from Excess Mortalities for COVID-19 in Germany

13 Pages Posted: 19 Feb 2021

Date Written: February 15, 2021


Background and Aims: The reported case and death numbers of the coronavirus disease (COVID-19) are often used to estimate the prevalence and infection fatality rate (IFR). We use the excess mortality instead in order to have an alternative and independent measure. In calculating age dependent IFR, we distinguish between healthy and weak individuals, where the latter are statistically defined as individuals with a remaining life expectancy of 20 weeks.

Methods: We estimate the cumulated prevalence of SARS-COV-2 from the excess mortality curve of the 90+ age population, since this part of the population is most likely to die from the infection. We then assume this estimated cumulated prevalence to be the actual cumulated prevalence of SARS-COV-2 across all age bands in Germany during the first COVID-19 wave. From there we calculate the age dependent infection fatality rates. We distinguish between the individuals with life expectancy below 20 weeks and those above 20 weeks.

Results: For the first COVID-19 wave, we find a cumulated prevalence of about 3% for Germany across all age bands.

The steep age gradient of the IFR is significantly smaller by a factor of around 10 when considering healthy individuals only, as compared to the IFR calculated for all individuals in the respective age bands.

For the 90+ age band we find an overall IFR of 7.0% while for healthy 90+ individuals, the IFR reduces to 0.50%. For the 80-90 age band the IFRs are 2.2% and 0.74% respectively, for the 60-80 age band 0.40% and 0.025%, for the 40-60 age band 0.071% and 0.023%, and finally for the 0-40 age band 0.0014% and <0.00085%. The overall IFR for the total population is 0.35% and for healthy individuals 0.074%. The IFR for individuals in the 90+ age band with life expectancy smaller than 2 weeks is assumed to be close to 1.

The results for overall IFR are in agreement with previous seroprevalence studies, supporting the here presented approach.

Conclusion: We conclude that around 2.5 million individuals in Germany were infected during the first COVID-19 wave, whereas less than 10% of infected individuals were detected via tests.

In addition to the known fact that the infection fatality rate depends significantly on the age of the infected individual, we furthermore show that a second major factor is the life expectancy of the individual under mild (COVID-19-free) conditions. This second factor has not been paid much quantitative attention to so far. We find that a large fraction of deaths due to COVID-19 actually stem from such weak individuals, i.e. people who pre-died in the first COVID-19 wave and would have died without COVID-19 only a few weeks later (harvesting effect, well-known for COVID-19 from other studies).

Note: Funding: No funding by any third party, since independent researcher.

Declaration of Interests: No competing interest. The corresponding author works as a Lead Data Scientist in the finance industry in Germany, the employer is not involved in this research, which is done outside the job.

Keywords: COVID-19, age, excess mortality, prevalence, infection fatality rate, harvesting effect

Suggested Citation

Weber, Alexej, Inferring the Prevalence and Age Dependence of Infection Fatality Rate from Excess Mortalities for COVID-19 in Germany (February 15, 2021). Available at SSRN: or

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics