Correct. It is an estimate, so one cannot ‘prove’ it.
It is important to note that the CBS has an estimate of 345. Rounded in thousands this is nil. My claim is not that the estimate of 8K is pitch-perfect. My claim is that I don’t think the adjustment of 345 by the CBS is not credible. I also state in my article that instead of 8K one can use 5K or 6K. Then the reasoning and conclusion remains the same. See the table with scenario’s.
The 8K is part of 21K people who died in 2020 from covid that I a distribute over the years in which they would have died without the pandemic. When one changes the estimate in one year, say the 8K from 2021, then on has to change the estimates in the other years in order to get to the total of 21K. This is a waterbed effect.
So when one takes the 345 of the CBS then we get problems in other years when on distributes the 21K over the years. See the table. Which numbers would you put in in the CBS scenario as estimates for the other years?
This is also an important question for the CBS: ‘How would the CBS distribute the 21K over the years with 345 for the year 2021?’
And a question for you: ‘Would you stake your life on the adjustment of 345 by the CBS?’
That is correct.
a. I didn’t claim the calculations by the CBS are wrong.
I claim that the adjustment of 345 by the CBS is not credible. It is also possible that the CBS has another explanation for the difference. Then the calculation are (possibly) right.
b. There is also the possibility that the calculations by the CBS are wrong.
The CBS claims that those calculations and techniques have been used for years and have proven their worth.
However in times of pandemics it is possible that an extra correction is needed.
The CBS calculates mortality probabilities as a function of age and year: qx.t = f(x,t).
But suppose that a ‘pandemic’-factor px.t should be in the formula.
We then get the formula: qx.t = px.t* f(x,t).
Without pandemic : px.t = 1
With pandemic : px.t < 1 for most x,t.
In year without pandemic we have px.t = 1, so then qx.t = f(x,t) = px.t* f(x,t) and the numbers from the CBS are ok.
With a pandemic we can have: px.t = 0.8365. Then the following is false: qx.t = f(x,t) = px.t* f(x,t).
If the CBS, or anyone who calculates with this, is not aware of this effect then they might wrongly assume that the uncorrected numbers are right.
Hence this is a possible pitfall for everyone who calculates with mortality probabilities in times of a pandemic.
c. In my article I look at expected mortality from a different angle. With that I calculate different numbers for expected and excess mortality than the CBS does.
If my reasoning is correct then the CBS has a gap in their numbers. It is up to the CBS to explain this gap.
If one says that I have to give an explanation for the numbers of the CBS then one says that the CBS should post the following on their website.:
—
The CBS has taken note of the gap in its numbers for expected and excess mortality as stated by rm L from A. The CBS takes the view that because Mr L from A has found the gap in our numbers it is up to hin to explain why our numbers are wrong. The CBS will not conduct any research to find an explanation for this difference. Wether is takes 5 or 10 years the CBS will wait patiently until Mr L from A has given an explanation.
—
This not how the CBS works. The numbers are theirs so the CBS shall and will do the research if there is a gap.
The above is relevant for all those who work with expected and mortality. So I advise everyone involved in those calculations and numbers to check wether this effect has a role in their numbers. With everyone I mean all statisticians, actuaries, epidemiologists, demographers and organizations (statistical bureau’s) all over the world (!) and others who are working with this subject.
Alice and Bob are opposite each other. Alice says: ‘I see one person.’ Carol walks around Bob and sees David standing behind Bob. Carol says to Alice: ‘There is another person behind Bob.’ Alice says: ‘That is not true because you haven’t proved that I need new glasses.’
Excess mortality as CBS reports is excess mortality year on year. Hence people who would have died in a later year without the pandemic. There are also people who died earlier than expected from covid but would have died in the same year without the pandemic. Say someone died in April from covid but would have died in October without the pandemic.
In scenario I of the distribution of the 21K people who died from covid in 2020 I estimated this number to be 7.000 or 7K. In 2021 according to the CBS 19K people died from covid. I estimate that form these 5K people would have died in 2021. This is premature mortality from covid. One could also call this excess mortality within the year. People could also have died earlier in the year from other causes like psychological stress. For this article I assume their number to be 1K. Then for 2021 we get the following table.
The CBS reports excess mortality for 2021 at 16K. This table with 30K gives in my opinion a better and more complete picture of expected and excess mortality caused by the pandemic and, possibly, the reactions to the pandemic.
In determining the expected mortality on January 1 of the year, CBS uses the population on January 1 as a base for their prediction. CBS claims that this incorporates everything from the past and that differences are partly due to immigration and changed population. That is too general. According to the CBS the 2021 estimate of 8,000, at least the underlying effect, is included in the new projected estimate calculated by CBS. In 2021 there is a difference of 7,655 from the CBS estimate. If this is explained by immigration, for example, then the question is what CBS expects to cause these additional deaths. Say, the additional immigration is 20,000. Then which 7,655 people will die because of that immigration. Are all immigrants older than 87 so that a third will die within a year? Or do CBS expect younger immigrants to have less resistance to a new disease in the Netherlands so that a third will die within a year? And so on. So it is not enough to say that increased expected mortality is due to immigration and changed population. A further explanation of the increase in expected mortality as calculated by the CBS is needed.